Film vs Digital - Dynamic Range

Discussion in 'Film and Processing' started by mauro_franic, Nov 24, 2008.

  1. This is continuation of an up-to-date comparison of digital vs film as of November 2008. In the previous issue I
    compared how even the small 35mm format film outresolves all digital SLRs in the market (all low speed film:
    color slide, negative B&W and negative color).

    In this issue I compare how negative film has a wider dynamic range than digital SLRs.

    For clarification I included the information of the setup directly on the pictures:

    http://shutterclick.smugmug.com/gallery/6616619_YJEwK#424020444_n2LsD-O-LB
     
  2. I'm not an expert on dynamic range testing but the logic behind your test is about the best i've ever seen....very interesting. I began with Digital with no professional film experience and, if this is true, there is in fact quite a difference i must say. What i did notice on-screen when viewing your samples is that the digital versions seem to be much more contrasty, which wouild impair the Dynamic Range somewhat verses your scan set at a low contrast setting. Cool comparison.
     
  3. Michael, I noticed the same and played adding contrast to the film shots and they look awesome. I still posted direct scans with defaults settings without PP to stay with the standard results.
     
  4. Mauro....was the 40D set on JPEG preferences or RAW? If JPEG, then RAW is where you really need to evaluate the DR of digital capabilities V/S film.
     
  5. I don't shoot jpeg. Set on raw 16 bit export to photoshop.
     
  6. I have no room to talk but...I recently saw some portraits taken with a 40D printed on Endura by a good lab and they were not very good. The skin tones were off. I know that that could be because of a few things but that is only what I saw. The photogrpher basically considers herself as a professional...
     
  7. Thanks for the test. Anything besides constructive criticism or questions asked indicates that one is afraid of not knowing/understanding something. Those who are confident in themselves either believe that they know the truth (but admit that they could be wrong in face of contradictory evidence) or they want to learn. You never use one source for you information (my God Almighty, imagine the poor souls that would listen only to Ken Rockwell and no one else?), and the more tests the better...even if there are some flaws. Flaws found with testing can be fixed prior to the next round of testing. I appreciate other people's work because then I do not have to do it. : )
     
  8. Thank you Alex. I hope this test adds to the contributions many others have done.
     
  9. Isn't it more about your digital workflow or film scans? I've seen some pretty poor examples of film work because of poor scan quality. And I've seen some pretty blocked up digital work, too. But I've also seen amazing quality from both. These tests usually go where people push them.
     
  10. Howard, can you push your digital camera to give the 14 stops film gave in this test without blowing highlights? And post the results?
     
  11. Also can you push Ektar to blow the highlights 4 stops over the exposure metering? Also be interested to see how you do it.
     
  12. thanks for the test
    there are so many factors to weigh in though
    though i'm sure your aware of that
     
  13. Yes Anthony. This is just my best attempt - I'm sure it is not close to being perfect. Good enough for my to draw new conclusions and share with other. Similarly to Les, the film's ability to exceed in the highlights was expected/known already but the shadow results blew me away.
     
  14. How did the Ektar grain hold up in the shadows?
     
  15. Then why is it that so many digital users claim they can record up to 10 to 12 stops of dynamic scene range? Can someone shed light on this? (pun intended)
     
  16. I say thank you for this test. But I too have understood this for some time. I feel that consumer buys digital because of not understanding film. Film has been cloaked in mysticism and ignorance for too long. Professional buys digital for different reasons than consumer. There costs can be amortised over time and needs for turn around with to client are higher.

    We did not have good access to film scanners in the past. I think we still do not. It is true that Nikon make most excellent scanner but still use is not simple and much learning is required. This could be made to be easier and more automated, but there is not as much profit for Nikon when compared to selling many D40, D60, D80 ...

    The film companies do not make enough money by this but camera company do and so we move to digital.

    so the good question is: is digital better?

    I have both for different occasion
     
  17. geez...how do you overexpose Ektar 100 anyway? I still haven't done that. My biggest problem is underexposure. That film seems to want all the light it can get. It will give beautiful colors with a different look that I've never seen before...but it wants a lot of light and definitely seems to prefer overexposure. I feel like I couldn't really overexpose that film unless I opened up the camera! :)
     
  18. Couple of thoughts:
    What iso did you use on the 40D? ISO will affect the DR.
    For whatever reason, DPP doesn't make the full use of the sensor's linear data. For a better comparison you could use lightroom/acr, but for the best, use DCRAW and output linearly.
    I'm not sure of your methodolgy, but I presume you determined the lower limit of the film/sensor based on being able to see something in the shadows? A better test would be something with detail (say some text or whatever). This is likely to give you a better judge of usable dynamic range.
     
  19. Bernie, of course in the 40D the iso was set to 100; ektar is a 100 iso film, and he used the same exposure. Why would he have used a different iso on the 40D with the same exposure?
     
  20. Look you guys, with your film VS dig. there is obviously things that film can do better than dig and things dig does better than film like less cost and I can see what I get when I take the picture and of coarse photoshop. What about dig HDR? I have a 40D and I am not a pro photog, although I would like to be one day. I really think film is not worth it for me right now or anyone starting in Photography as the cost can get expensive. Now that I have been doing dig for a min I want to get into film cause of the look.
    I don't Know Just thought I would throw that out ... Now you can rip me apart
    Josh
     
  21. Very nice...about what I would expect on the highlight end of things, but the shadow end was interesting - I'd never really thought of film picking up more in the shadows like that, I guess.

    Any chance of doing something like this with a middle of the road slide film, such as E100G or Provia (not too saturated / contrasty and not too dull)? I'm curious hows much "usable" latitude you could pull out, and how it compares to the digital and Ektar100.

    Thanks,
    Jed
     
  22. Then why is it that so many digital users claim they can record up to 10 to 12 stops of dynamic scene range? Can someone shed light on this? (pun intended)
    Benny, most dslrs get around 8 to 9 stops. It has been said the D700 gets something in the 10-12 stop region.
     
  23. Just messing around with a curves adjustment layer and I found I do get some more detail out of the digital image. This is the before image.
    00RbGs-91865584.jpg
     
  24. This is after with the adjustments applied to the D40 image.
    00RbGt-91865684.jpg
     
  25. To be honest I would not want to use either film or digital with this amount of under exposure. Just much easier to expose correctly and get good image from either medium than to jump through hoops to rescue bad digital or film images.
     
  26. Why did you only adjust the digital image? What is the point? Adjust both.
     
  27. Lets not for the DR of the Fujifilm S5/S3 Pro cameras.
     
  28. Heres both adjusted.This is one with both adjusted. Now they really look bad and there is nothing between them in terms of shadow detail.
    00RbHZ-91873584.jpg
     
  29. One thing we can learn from these is that it is better to expose correctly. I don't think anyone would welcome images of this quality. Heres is one more adjusted to the extreme. Not much use for anything but here it is.
    00RbHp-91875584.jpg
     
  30. Full Frame

    You should redo this tests with a full frame digital. If you are comparing 35mm film you should compare it to a 35mm sensor. Because the sensor on this full frames is bigger, the DR is wider. New comers to this market like the sony A900 are pushing the DR on digital cameras to new limits .
     
  31. isn't this the old slide/vs negative issue in an alternate form?

    Negative always allowed a more leeway than slide film for errors in exposure. Digital is now forcing folk to be a little more disciplined than they might have been on negative film.

    I'd be more interested to see slide film vs digital at comparable ISO/settings. You'd have to test on a full frame DSLR so that lens ratios are not a consideration in the test.
     
  32. SHADOWS:

    I experimented with DPP exposure adjustments to the raw and curves last night on the digital shot, I also scanned the film with different analog gains. I'd say they are very comparable - with film probably having one stop advantage in the shadows for similar amounts of noise.

    HIGHLIGHTS:

    Here the film shot has at least a four to five stop advantage, before highlight detail is affected. I actually agree with the previous poster, it almost impossible to blow the highlights with properly exposed negative film.
     
  33. That would be an interesting test: slide vs. digital.I liked this thread. I learned quite a bit about both Ektar and Digital. These types of threads are why I come to PN. I don't think it's a Digital vs Film argument. I look at it as a comparison of the medium; like comparing water colors to oils. Folks prefer one to the other and they each have their own limitations and strengths. (That's the best analogy I can come up with.)I also like to see what the limitations are of digital in a "film language" that I can see. The numbers and whatnot that the marketers throw out there don't mean much to me and I haven't gotten around to learning all the technical terminology and what they actually mean when it comes to the image rendition. Anyone can throw out dynamic range claims and numbers to "back them up", but show to me how exactly it affects the image with a comparison of something I'm familiar with. I want to see what the numbers are and I don't mean a graph or chart. Which is exactly what has been done here.
     
  34. Thanks Kevin. It was challenging to arrange the strobes to get exactly one stop difference with figurines similarly spaced. As far as know, it has never been done and the results were better worth it for me than looking at a strip with shades of gray.
     
  35. Digital cheap, but bad.

    Film expensive, but good.

    If I can be of further help, just let me know.
     
  36. LOL, Steve! I don't need to come here anymore! :)<p/>On another note: I think when the megapixel race fizzles out, the next sensor race will be dynamic range.
     
  37. Using Full Frame to do the test will be a hard one being that higher MP can lead to highlight clipping.
    I think the best thing is to use a sensor that was built for DR on the physical level i.e. the SuperCCD which was made by a
    film producing company.
     
  38. As usual, the limitations are more with the photographer, not the gear or materials. There are plenty of fancy tests that come to the opposite conclusion as the OP here.

    http://www.clarkvision.com/imagedetail/dynamicrange2/index.html

    http://www.normankoren.com/digital_tonality.html
     
  39. Thanks for this test Mauro. When a dynamic range issue came up at DPReview, I did a test somewhat like this. You'd think I gored their sacred cow or something. I used Fuji NPS and Ilford HP5. I was able to maintain around 12 stops with NPS.....trying the same with my old 1Ds Mk2 was an exercise in futility.

    Where it got interesting is when the B&W results showed around 14 stops. Instead of trying themselves, the personal attacks and insults started flying by a number of them.....the people who have never used film it appears.

    Mauro, those of us who figured out how to expose film and digital properly found early on that film maintains highlights better and digital can hold shadows better.....but the range captured by film is larger than ANY, and I mean ANY DSLR available on the market today.
     
  40. After many, many years of not using film, I bought and old T50 to experiment. I figured for $50, two lens a canon 50mm f/1.8 and a zoom. I could compare so your test is pertinent to me. Thanks.
     
  41. Mauro: I shot all of the Ektars since the eighties, and they didn't have that much dynamic range, really. They were almost like shooting slide film - especially the Ektar 25. Well, you asked for an example from my digital with fourteen stops of range. I'm sending a shot with a very broad range, not sure if it's fourteen. But it was taken on a bright, cloudy day using RAW. After the RAW was processed in Aperture 2 it was necessary to copy it as a reduced jpeg for this upload. But if you look carefully, there is detail in the brightest cloud cover overhead, and in the studs under the bridge in the deepest shadows. That is good enough for me.
    00RbNb-91933584.jpg
     
  42. I guess the clouds blocked up a bit in the upload here. But in the original jpeg conversion from RAW it's all there.
     
  43. So, how do the prints compare?
     
  44. This is an exceedingly interesting approach to the challenge. Sure, no test is perfect, but I really like what you did. Good
    job Mauro on giving us a different way of looking at the question.

    If Kevin is correct then Fujifilm has already won with the S5 Pro. Being an S5 user myself, I am of the belief that the S5
    exceeds any film in dynamic range. But I have no evidence. I only hope there will be an S6.

    James
     
  45. James - I believe the S5 was rated at about 12 stops. That's a good match for films like Reala and Pro 160S. However, it is still 2 to 3 stops below what some B&W films can deliver. That said, I find 10-12 stops sufficient for virtually any work.

    Mauro - Did you use multisampling on the 9000 to reduce noise. When I used to scan Astia with a 9000, I used 16X multisampling and it did wonders to reduce noise further in the shadows. In my opinion, it was like gaining an extra stop in noise.
     
  46. Mauro, thanks a lot for your test. A simple test that shows, at just one view, how film is still a better medium for recording images than digital. Forget about curves and noisy words from others. The truth is the truth. I use a dslr from time to ime, but when I plan to take pictures seriously, I always use my film cameras. By the way, a film scanner is not usually able to extract all the info stored on film, so just imagine. Cheers.
     
  47. The middle grey is not reproduced same between the digital and film images. This could be scanning or exposure - either way there is no control point in the "data." For photography and what you're attempting to do, that would be a 10-step grey scale with the middle grey reproduced to the same level for each system being tested.

    The means you cannot draw conclusions from what you've done as you have no control point that can be compared as equal in each test.
     
  48. If you know well film processing and master on this field, than do it!
    If you well know digital post processing, than use it.
    If you master in dark room and in digital processing, than enjoy your best work!
     
  49. What I don't get is why you'd use analog gain to dig out shadows in negative film. Are you *reducing* hardware exposure to preserve detail in the clear parts of film?

    Matt, the clarkvision site isn't maximizing the dynamic range of negative film as it under rather than overexposes it. I wouldn't rely on it for the purposes of comparison.
     
  50. All this stuff is really old stuff; like debating a cordless drill versus a AC powered one; or whether folks will accept color TV versus B&W; whether power steering is better than manual; whether slrs will catch on; whether color film will be accepted; whether film will displace glass plates.
     
  51. Anyone who has uses a DSLR and shoots film knows how much more lattitude film has. You shoot it, you got it. With a DSLR you have to resort to all sorts of work arounds such as underexposing to save the highlights, then lightening the shadows in photoshop. Or, shooting multpile images at differing exposures in PS and combining the images later in photoshop, using HDR. Either way is a pain in the *ss for me. I wish they would start automating some of this in the DSLR's processors, like they do in the Point and Shoots now. Hell, while I'm on this rant, why not add face detection for autofocus and exposure, really good AF tracking, etc. like in my FX150??? After all, a DSLR is MUCH more expensive, it should have these features.
     
  52. I agree; nice job Mauro.

    But I also agree with Pedro, that these tests should be done with a full-frame camera. The pixels on a 5D are 8.2 microns in size, whereas Canon's 20MP (including the 5D II) have pixels 6.4 microns in size. I'd be interested to see some tests comparing DR between the 5D & 5D II... but, regardless, I'm sure a 5D would fare better in DR than a 40D.

    Also, Mauro -- did you try, for the 40D image, to set 'contrast' to -4 in DPP? This typically brings out shadow detail.

    Of course negative film has an advantage in highlights as exposure of film leads to clusters of reduced metallic (black) silver which makes penetration of light more and more difficult through the layers of film as exposure increases. It's a beautiful system, really, chemically. Too bad no one's thought of how to do it on silicon chips... yet. But I do think that digital is quite capable of shadow detail, and I'm wondering if lowering the contrast or using 'fill light' will bring that out in your digital image better than it would in the film image (using 'fill light' on film scans typically exaggerates grain very quickly).

    Of course, digital underexposure is really bad, given the # of bits available for shadow detail. Also, you don't want to end up with banding like that which is typical of certain Canon CMOS sensors (http://www.photo.net/canon-eos-digital-camera-forum/00RBTe). But I do think that the shadow performance could be made better by different post-processing in your 40D image above.

    Cheers,
    Rishi
     
  53. "Anyone who has uses a DSLR and shoots film knows how much more lattitude film has.."

    You mean like Kodachrome, Velvia, or Provia....? Those films have about 6.5 stops of dynamic range. A professional digital camera will have at least 8, and some of the newer ones 10-12. If you mean negative film - then, yes, negative film it has about 12 stops of dynamic range.

    You shoot a digital camera like transparency film - I don't see the problem with that ...you have to know how to use the equipment for its inherent performance ability whether the workflow is film or digital.
     
  54. I don't see how someone can say that Film camera gear is more expensive than Digital camera gear. Don't factor in the
    lenses, because they are the same. Brand new Nikon F6 - $1600, Used, but perfect condition Nikon F100 - $350, Used but
    perfect Nikon F80 - $90, Brand new Nikon D700, $2500 - 3000, D3 $4500. Cost of buying new digital camera because
    current one is now obsolete or sensor is worn out and has too many hot pixels, $2500 - $5000. Film purchasing and
    development costs during that same timeframe, considerably less. I love film, and I love digital. I use both for different
    reasons. I would never venture to say that digital is less expensive than film, because pro quality top-of-the line film
    camera gear (including medium and large format) can be had for much less cost than mid level digital gear (such as D300
    or EOS 50D)
     
  55. Stuart:
    "To be honest I would not want to use either film or digital with this amount of under exposure. Just much easier to expose correctly and get good image from either medium than to jump through hoops to rescue bad digital or film images"

    You can decide how to expose the picture but you can't control how many stops the scene has.
     
  56. Pedro, I didn't have a full frame digital handy last night, it would be interesting if someone else can run a test and add to the comparison.
     
  57. Matt,
    "As usual, the limitations are more with the photographer, not the gear or materials. There are plenty of fancy tests that come to the opposite conclusion as the OP here"

    You are looking at the ACTUAL pictures, if there is a conclusion out there that states Ektar100 shouldn't have that much latitude, or that the 40D should have more.....then that conclusion is wrong.
     
  58. Howard:

    "Mauro: I shot all of the Ektars since the eighties, and they didn't have that much dynamic range, really."

    Then the new Ektar is much improved.
     
  59. Dave:

    "Mauro - Did you use multisampling on the 9000 to reduce noise. When I used to scan Astia with a 9000, I used 16X multisampling and it did wonders to reduce noise further in the shadows. In my opinion, it was like gaining an extra stop in noise."

    No I didn't, I will try tonight and post for you.
     
  60. Steve,

    "The middle grey is not reproduced same between the digital and film images. This could be scanning or exposure - either way there is no control point in the "data." For photography and what you're attempting to do, that would be a 10-step grey scale with the middle grey reproduced to the same level for each system being tested.

    The means you cannot draw conclusions from what you've done as you have no control point that can be compared as equal in each test."

    Both cameras were shot in the same setup, shutter speed, fstop, tripod, and lighting, one burnt the highlights and the other one didn't. No white/gray balance will change that but feel free to experiment and post results.
     
  61. Apart from slide film, the negative or digital image is only a step towards the final stage which is a print, either traditional or inkjet or an image on a screen. Should we not be more concerned about the dynamic range of these than the film or CCD/CMOS device?
     
  62. Roger;

    "What I don't get is why you'd use analog gain to dig out shadows in negative film. Are you *reducing* hardware exposure to preserve detail in the clear parts of film? "

    This is just the terminology of the Nikon Scan.
     
  63. Rishi,

    "Also, Mauro -- did you try, for the 40D image, to set 'contrast' to -4 in DPP? This typically brings out shadow detail. "

    Good point, I will try this.
     
  64. Steve,

    "Apart from slide film, the negative or digital image is only a step towards the final stage which is a print, either traditional or inkjet or an image on a screen. Should we not be more concerned about the dynamic range of these than the film or CCD/CMOS device?"

    When detail is gone because you burnt the highlights, IT IS GONE - you can't print it. If it is there, it is always possible to print it with adjustments to match your printer and goals.
     
  65. Thanks Mauro,

    I look forward to seeing the multisampling scan to see if it assists as I think it will in the shadows. Ektar seems so good that I can see a number of people putting their heavy, expensive DSLRs in the cabinet only to be used for snap shots.
     
  66. I'm seeing some really basic confusion in this thread about scanning and film.

    Howard- how do you know Ektar in the '80s had poor dynamic range? From looking at the negative? Scanning it? Printing it yourself? You're not judging based on commercial lab prints, are you?

    Dave and Mauro, multisampling is irrelevant for color negative film- are you seeing scanner noise in the film highlights? Remember, the shadows of negative film are clear and CCD noise is not in highlights but shadows.

    Ditto for adjusting film analog gain. Set it right so as not to blow out the shadows and lock exposure for the roll. If you want more recommendations on maximizing color neg. film scan quality look up Erik Krause's Nikon-based "super advanced workflow" where you use hardware analog gain, rather than software clipping, to subtract the film base. That can help you with color accuracy but probably doesn't matter for this thread, apart from the fact that you need to set exposure properly for the type of film.

    I strongly agree with Steve that you haven't calibrated your scan or digital file in any meaningful way. Just leaving software settings on the defaults make no sense at all.

    What you want at the minimum is to do something like use middle gray on your QP card to set middle gray on both images to the same point. Then you can at least more clearly see the difference in shadow vs highlight rendition. This is basically what dpreview does in their dynamic range tests using a blacklit stepwedge.

    Now, how you map the negative film to your final image is an open question, I think. Negative film compresses highlights, so is the goal to preserve overall tonal relationships, accurately match a backlit step wedge, show how much highlight detail it can preserve before clipping, etc? How you uncompress them is really up to you. What you *don't* want to do is scan with the defaults on and let the scanner software do this for you, and heaven's forbid let it *clip* highlight data in an attempt to increase scene contrast and make it look more attractive.

    That said- I don't disagree with the basic conclusion that any negative film has a gracefull roll-off of data in the highlights and digital doesn't. That's not news.
     
  67. Also, has no one read this statement from Canon's own 'Full Frame CMOS' white paper:

    "Canon’s full-frame sensors have reached another image quality milestone as well. Their gradations and dynamic range are now the equal of the best positive films, and their resolution and lack of grain are superior. No smaller sensor has achieved this level of performance."

    That was as of 2006. I don't know if their 'highlight tone priority' or whatever they call it has extended this performance. But I seriously doubt Canon was *understating* their cameras' dynamic range.

    In real-life shooting, in my experience, shadows in digital captures are easier to work with than those from Velvia 50 scanned on an Imacon 848 even. But I can bracket +/- 2 stops with negative film and get the same image from all 3 with enough post-processing. It's kinda ridiculous. But on the other hand, it's the same logarithmic response of negative film that typically makes the images not-so-appealing because of lack of contrast; Velvia is so pleasing because of the enhanced local contrast, which is how our eye-brain system sees things (even though overall we have a logarithmic response to light due to multi-sampling the visual scene that we see).

    Rishi
     
  68. Mauro,
    Again, thank you very much for very entertaining and selfexplanatory test.

    Crop vs FF should not make any difference, for the dinamic range of 5D and Canons with smaller sensors is about the same (according to DPPreview)
     
  69. Error above- should be "backlit stepwedge" and a qualification- multisampling for color neg. film is generally useless but if you do see scanner noise in the lightest areas, give it a try.

    Here's how DPreview does Dynamic range testing:
    http://www.dpreview.com/reviews/sonydslra900/page23.asp

    Here's how Dxomark.com tests (Rishi- this also addresses your questions of a 40D vs 5D, etc)
    http://www.dxomark.com/index.php/eng/DxOMark-Sensor/DxOMark-metrics/Landscape-photography-maximum-Dynamic-Range

    Both of these sites test digital cameras, but the concepts carry over to film. I'm not sure what the point is of testing film vs digital for dynamic range, but if you think it's useful...
     
  70. Thanks for the comments Roger.

    "Dave and Mauro, multisampling is irrelevant for color negative film- are you seeing scanner noise in the film highlights? Remember, the shadows of negative film are clear and CCD noise is not in highlights but shadows."

    Dave and I are referring to use it for the shadows.

    I will check Erik Krause's article.
     
  71. Thomas Wilson , Nov 24, 2008; 10:22 p.m wrote:
    Enough already!
    +1
     
  72. so, when do the digital cameras come out with extended dynamic range...what would it take...
     
  73. One thing regarding testing of DR of digital cameras I can not get. All graphs demonstrate brightness range from 0 to 256, which is 8 stops, therefore 8 stops should be the maximal DR for any digital camera. How can one say that the DR of some cameras is 9.1 or 11 stops?
     
  74. Pedro Esteves , Nov 25, 2008; 06:22 a.m.
    Because the sensor on this full frames is bigger, the DR is wider.
    James Nakia Johnson , Nov 25, 2008; 08:33 a.m.
    Using Full Frame to do the test will be a hard one being that higher MP can lead to highlight clipping

    Where are some of you people getting your information?!?
     
  75. One thing regarding testing of DR of digital cameras I can not get. All graphs demonstrate brightness range from 0 to 256, which is 8 stops, therefore 8 stops should be the maximal DR for any digital camera. How can one say that the DR of some cameras is 9.1 or 11 stops?
    Because the analogue to digital converter is using 14 bits, or 12 bits on older cameras. These 14 bit dslrs could theoretically be able to manage up to 14 stops, but the reality is that noise kills the lower few stops.
     
  76. "Dave and I are referring to use it for the shadows."

    Multisampling does nothing for the shadows of negative film. It can help with noise in the dark areas captured by a CCD which are the highlights of negative film. Is that really what you're talking about- that you see scanner noise in the lightest frame?
     
  77. And the DR is dependent upon the maximal vs. minimal voltage that the amplifier can read, and the tonality determined by how small the steps between voltages that can be distinguished.

    It's true that the smaller the pixel, typically the less maximal charge it can build-up before it 'clips' or 'spills over' so to speak. If you just up the megapixels and shrink the pixels, you get crappy results like the new high MP Canon digital ELPHs.
     
  78. If you just up the megapixels and shrink the pixels, you get crappy results like the new high MP Canon digital ELPHs.
    Agreed Rishi. But this is independent of 'full frame' that some people above are sprouting.
     
  79. Roger, correct. There is some noise in the darker parts of the image (thinner negative) when analogue gain is increased.
     
  80. Roger: In 1984 we didn't have the technology to scan the Ektar 25 negs. I examined them through a loupe, and saw about the same dynamic range as the Ektachrome 64 I was shooting then. But that was a subjective evaluation. The prints came out with somewhat blank deepest shadows and somewhat washed highlights in brightest sun, on Kodak commercial print papar. I can't remember which series the paper was. If I went down into the storage boxes and found those now almost 25 year old Ektar 25 negs it'd be very interesting to see how they'd scan. I remember the Ektar 25 being a short run. Amateur photographers had a hard time with it. Perhaps I should get out my old 35 camera and try the new, but it isn't available here.
     
  81. Hi Howard,
    I was asking as what's possible with automated optical enlargements is very different from scanning.

    I've just done a few scans from 1980s era negatives (B&W and color) and have made prints far superior to what we got at the time, for both dynamic range and color. As far as I can tell all color negative film has the ability to record a large range of brightnesses.

    Good luck finding the negatives and hopefully they'll be in good condition.

    Rishi, you might find this interesting regarding DR and noise:
    http://www.dxomark.com/index.php/eng/Insights
     
  82. Rishi,

    "Also, Mauro -- did you try, for the 40D image, to set 'contrast' to -4 in DPP? This typically brings out shadow detail. "

    Here is the result of DPP Processing Contrast -4 on the highlight test for the 40D:

    http://shutterclick.smugmug.com/gallery/6616619_YJEwK#424643294_NZu3f-O-LB
     
  83. Dave:

    "Mauro - Did you use multisampling on the 9000 to reduce noise. When I used to scan Astia with a 9000, I used 16X multisampling and it did wonders to reduce noise further in the shadows. In my opinion, it was like gaining an extra stop in noise."

    Here is the result of 16x sampling on the Ektar shadow test (I skipped ICE):

    http://shutterclick.smugmug.com/gallery/6616619_YJEwK#424643305_CbA8c-O-LB

    (I see reduced noise)
     
  84. I dunno, but I get the feeling you (OP) are setting everyone up for a sell job of some sort. I notice that if you click on the word 'Close' at the top of the window you are taken to Alex Franic's page where he just happens to be selling his film scanning services. The OP is also named Franic. Marketing???

    Mel
     
  85. Mauro, any chance of getting our hands on the 40D highlight raw? I'd like to develop it linearly and see how much of the blown highlights were captured by the sensor. Not expecting it to be a lot more, but DPP is pretty bad with handling highlights.
     
  86. Interesting. Setting contrast to -4 brought back some highlight data. I bet, now, in Lightroom, if you were to add some 'Highlight Recovery' and some 'Fill Light', you'd get some formidable results with that digital image that would put slide film to shame (in terms of DR).

    Indeed, it must be so, given that CCDs respond to light in a linear fashion (some argue that CMOS respond logarithmically, making it more like negative film, but I haven't heard a strong or sensible argument for that yet nor have I looked enough into the potential logarithmic buildup of charge on said sensors), while positive film responds in an exponential fashion if you look at Fuji's own response curves.

    Mauro -- would you mind posting the RAW of the 40D image? I'd love to try processing it in LR using the new ACR profiles, and attempt some highlight recovery & fill light & whatnot.

    Thanks,
    Rishi
     
  87. I don't know why you chose the example where you fiddled with analog gain to do the 16x comparison with. That shot looks really underexposed. F8 seems a lot more appropriate for this scene. I put it into photoshop and compared new against old- both look terrible but I don't see a noise difference (maybe there's more of a difference in the full resolution original- this one is a rather small file). If you email me the original raw file and scan (analog gain done so as not to blow out shadow detail and without any white or black point clipping) and I'll show you how I would compare the two using Photoshop and Adobe Camera Raw.
    00RbpH-92147684.jpg
     
  88. Ellis said:

    "You shot Jpegs didn't you? -- proving nothing more than your ignorance."

    Then Mauro responded:

    "Ellis, for the second time, they were shot in RAW. You can take your derogatory tone elsewhere. Are
    your results any different than this? Would you post them?"

    Then Ellis goes catatonic.

    Ellis, grow up; or at least be a man and respond to Mauro's response.

    Ridiculous.

    And, yes, everyone -- I know Ellis lives on photo.net. I know he's taught everyone here (including myself)
    tons of stuff. But he shouldn't get a pass for being a drive by jerk and then going silent when he's proven
    wrong.
     
  89. If you guys tell me how/where to post the RAW file I can do it now.
     
  90. It is 11.7MBs.
     
  91. Eric, I believe Ellis took personal offense. Not sure why because when I looked at his portfolio, the main picture has the highlights blown on the forehead. He should understand this issue with digital first hand.
     
  92. Do you have any server space with your internet service provider? If so, upload it there, and then post a link to it in here. Or maybe you can even upload it straight to here?
     
  93. I don't have server space that I know of. If you send me instructions I'll upload it anywhere you tell me.
     
  94. Mauro, thanks for doing these tests. I don't understand why people are so angry at you for trying to share information.

    I'd really like to see this test repeated, but with a few differences...

    Film vs. either a 1DS or D700/D3 so that we are comparing versus a full frame DSLR.
    Using the SAME EXACT Lens on Both the Film camera and Digital.
    Having the Film and Digital Cameras on seperate tripods and taking the exposures at the same time.

    I'm sure there is someone here generous enough to donate all the equipment =)

    My only point is that because you are using different lenses, not making the exposures at the same time, and comparing 35MM film to a smaller sensor, that the results might be skewed a bit. Of course, this was a great experiment, and you did a great job with what you had available.

    I'd just be curious to the results of the test comparing the film to a full-sized sensor... that's all.
     
  95. That is a good point Keith. Some lenses may differ in a fraction of a stop in their speed (although not five stops).

    I live in the Atlanta area so if anyone is interested this weekend we can repeat as a joint effort.
     
  96. I meant someone with a 1DsIII or D700/D3.
     
  97. Mauro.... try uploading it directly into a forum post. I think I have seen this before. After you make a post, there will be an option to attach an image to your post. Try uploading the raw there, and see if it accepts it. Otherwise, try this site: www.yousendit.com. From memory, I think it allows you to upload it there, and then sends an email with a link in it to you and whoever else you want. Cheers.
     
  98. Cannon vs. Nikon
    Film vs. Digital
    Negatives vs. Slides
    Medium vs. 35mm

    Do some damn research right here on Pnet before posting such stupid posts. These comparisons have been discussed to death. Not only are you beating a dead horse, you've dug the corpse up just to beat it some more. Get over yourself.

    Move on to some other discussion that has real relevance to your ability as a photographer.

    And Mauro Franic you should apologies to Ellis. He's been a contributor to Pnet a whole lot longer than a year and has forgotten more about photography than you'll probably ever know.
     
  99. Nice!

    I attempted yousendit.com... let me know if it works.

    http://www.yousendit.com/download/TTZtZEV3YTJ0d0ZFQlE9PQ
     
  100. What's the point of posting anything other than mundane questions/comments on these forums if the best response you can expect is a string of insults, personal assaults and rude behavior?!

    I read these to gather information. If you have nothing to contribute, please stand aside (Eric A and others). I intended to do so myself, but I am tired of unsubscribing to threads that interest me but have degraded to the point they're not worth reading anymore.

    You don't find this behavior in archived threads. Is this a recent phenomenon? Are there any other sources of up-to-date, relevant information out there?
     
  101. Eric A, I don't think you can vindicate a person that uses derogatory words by using some yourself.

    You should be glad you know how many stops a 40D and Ektar can record. Some of us don't know, and have to find out with our own experience.

    I meant no insult to Ellis. I don't find constructive to use words like ignorant or stupid though. It is a fact his portrait has blown highlights and I meant no disrespect with pointing that out.
     
  102. Mark, please don't be discouraged. This thread is overall very constructive. There is a lot of people working on it together.

    I just posted the RAW files and some people will use better tools (than the ones I have) to find out if the burnt highlights can be recovered from the brightest stops. Stay tuned.
     
  103. Yeah right, you two (Mauro and Mark), who aren't even paid members, are the gate keepers of decorum on Pnet. Here's some decorum for you. DO SOME GOD DAMNED RESEARCH BEFORE MAKING YET ANOTHER STUPID VERSUS POST. Those who have paid their dues, and have been around for awhile have seen them a thousand times. You have proved nothing with your post except you haven't a clue about what the hell dynamic range is. You've read it in a book somewhere so now you think you're an expert. Well guess what? Folks like Ellis and others wrote the damn book you read on dynamic range. And I'd sooner listen to his BS than your so-called comparisons.
    Your first and most deadly error in your so called comparison is that you are comparing apples to oranges, pigs to cows, airplanes to helicopters. Since your premise is based upon a false comparison nothing that follows makes a lick of sense.
    So in summary your post is indeed ignorant since it was done with little real understanding of the technology involved or the terms used.

    Hows that for contributing, Mark?
     
  104. Forum moderator please help.
     
  105. Mauro,
    You have done excellent work here and on other threads with your comparisons.
    Don't let a few "subscribers" who have more money than brains critique your work.
    Your tests have shown me and others, things that translate into real world shooting.
    Keep up the great work and I look forward to your future posts.
     
  106. Some people like publicly embarse themselves. It is their problem. We should not take them seriously.
     
  107. Thank you.
     
  108. "Folks like Ellis and others wrote the damn book you read on dynamic range"

    good thing, the photography isn't very interesting, hey Eric, how is the view from Ellis's backside?
     
  109. Let's all stay focus on the discussion. I emailed the moderator for help. He will take care of it. We should ignore derogatory remarks and continue on.
     
  110. "Eric, I believe Ellis took personal offense. Not sure why because when I looked at his portfolio, the main picture has the highlights blown on the forehead. He should understand this issue with digital first hand."

    Mauro - As evidenced by his many contributions to these fora, Ellis understands digital far more than you seem willing to give him credit for. Please note - the portrait you critique is not Ellis'. It's a portrait OF Ellis, shot by, and appropriately credited to, someone else.

    While I don't support the way Ellis chose to respond, I certainly understand his level of frustration, as I share it.

    Predictably, this discussion has gone the way of most such discussions. Why digital "vs" film, with all of the chest-thumping that goes along with such a "smackdown" approach? I'm all for a rational discussion of advantages and disadvantages of both formats (both of which I use, enjoy and appreciate for their respective strengths), but these never ending wars between the two, with excruciatingly limited comparisons...? Which "digital" vs which "film"? With what assumptions and with what levels of expertise and experience with each? And with what intent? And yet it always boils down in too many people's minds to "film rulz" or "digital rulz", and we get ignorant statements like: "Digital cheap, but bad. Film expensive, but good.", as if there weren't outstanding work being done by talented shooters with both formats.

    All you're showing here is a comparison between two very specific and limited examples and treatments of each format, given your equipment, understanding and skills with each. There's nothing wrong with you doing these kinds of tests to prove which format might be "best" for you in a given situation and given your skills and available gear. However, as with most such tests, there is little general relevance other than for you in these very specific situations. Summarizing these tests as being generically about "film vs digital" is, plainly stated, incredibly simplistic. And the unfortunate, but predictable, result is that those with axes to grind are out in force grinding them. Post this in a film group, and the discussion will go one way - post in a digital group, and it'll go the other.

    I long ago proved to myself that, given my skills and equipment, and given low ISO shooting situations and no time pressures, *I* can produce "better" results from scanned film (that is, results which please *me* more) than I can from any dSLR format camera that I've tried. But I've also proven to myself that all of those qualifiers are very, VERY important. Change a few of those qualifiers, and I've also proven to myself that I do FAR better work with digital.

    Scott
     
  111. Scott, I agree with everything you said, with the exception that I don't give Ellis the credit for his digital understanding. He may be the best photographer and I don't put that in question since I don't know him or his work. My first and only interaction with him was his contribution to this post.

    I understand your point and I agree that this exercise is more relevant to me that I use both the 40D and Ektar than to other people who use different systems.
     
  112. I was not referring to the picture of Ellis but the portrait on his portfolio. It is not a personal criticism either, just a technical observation.
     
  113. Can you upload the full resolution flat scan (no clipping) as well? Jpeg is fine (save at level 10/12 or 90% compression).
     
  114. Perhaps she doth protest too much?

    Eric A,

    All Mauro did was a pretty damn good job at conducting a fairly basic but relevant experiment and showed his results
    in a fairly informative manner (which I also frankly thought were a bit surprising). Your protests about being "the
    expert here" ring completely hollow and the shrillness in your tone only denigrates whatever point you were going to
    make -- either his experiment was flawed (with no explanation as to why) or perhaps more tellingly, that amateurs
    can't conduct experiments? If such attitudes had always persisted, we would never have had the
    Scientfic Revolution of the Middle Ages. Many well meaning amatuers have contributed immensely to our
    understanding of the world in all sorts of fields. If you are so knowledgeable (which frankly most people who try to
    stop such inquiries usually aren't), then perhaps you could just address the evidence that Mauro provides? Claiming
    some higher authority (which only religion not science recognizes) is wholly unpersuasive. Mauro has gone out of
    his way to truly do some basic experiments and open his results to the community. You've just tried to shout him
    down (with nothing but diatribe). Then you try to claim you somehow own the site? Go back and read the ordeal of
    Galileo Galilei and ask where your post above fit ins.

    Good luck.

    Cheers,

    Steve
     
  115. Here is the scan - uncropped - no PP. I had to convert to 8 bits to generate the JPEG:

    http://www.yousendit.com/download/TTZtZEV6SEJrUmxFQlE9PQ
     
  116. Excellent post. Thank you Mauro! Keep it up! The Ektar looks great buy the way. I have to try it. Petr
     
  117. When you compare 35mm film to digital,compare it with a full frame CCD/COMS sensor.Not with a APS-C or APS-H sensor.Full frame sensors have a grater D-range to cropped sensors.The bigger the
    sensor,the higher the dynamic range.
     
  118. Rashed Ahmed "When you compare 35mm film to digital,compare it with a full frame CCD/COMS sensor.Not with a APS-C or APS-H sensor.Full frame sensors have a grater D-range to cropped sensors.The bigger the sensor,the higher the dynamic range."

    I still think the S5 can and should be considered when doing these tests. I have no issues with S5 DR.
     
  119. I suggest you all have a look at this:

    http://www.clarkvision.com/imagedetail/film.vs.digital.summary1.html
     
  120. Actually, scratch that- This is what you want to read:

    http://www.clarkvision.com/imagedetail/dynamicrange2/index.html

    "This shows that the Canon 1D Mark II has a much higher dynamic range than either Fujichrome Velvia slide film
    and Kodak Gold 200 print film. Kodak Gold 200, in this test, showed 7 stops of information, Fujichrome Velvia 5
    stops, and the Canon 1D Mark II, over 10 stops of information! Further image analysis shows at least 10.6 stops
    are recorded by the canon 1D Mark II camera (the full range of of detail in this image, Other testing of the
    noise level versus intensity shows the Canon 1D Mark II has 11.7 stops of dynamic range."
     
  121. Do the Clarkvision test really reflects the Full Frames being mention here ? the canon 1D Mark II is a 8mp Full Frame so the sensors are a lot larger then the newer Full Frame cameras. From we see each new camera with more mp have a shorter DR and produce higher ISO noise. Now camera companies started to come out with in-camera software solutions to fix the physical limitations (Even Fujifilm does this with there digi-cams).

    I personally will like to see Clarkvision throw the Fujifilm S5 Pro in the mix, yeah it is an older camera with a smaller user base but it is the only camera along with the S3 Pro that have a sensor that was design for the wide Dynamic Range and to come close to the results of negative film. Fujifilm was talking DR way before any other of the camera companies were but it is rarely is brought up in any of the discussions.
     
  122. david_henderson

    david_henderson www.photography001.com

    Whatever. After years of shooting contrasty slide film, I find exposure on my 5D a whole lot easier to the point that the spotmeter doesn't see light of day and the grads get a small fraction of their previous use. So it would be pretty pointless trying to persuade me that "film" has a higher DR than digital because for me it just doesn't.

    But then I understand that the dynamic range of digital cameras varies and so does that of film types and emulsions. Trying to compare one to the other in general is simply unproductive. Any analysis, no matter how interestingly carried out, will have no general applicability and its usefulness is confined to the comparisons actually made.
     
  123. Stop talking about it and go out and shoot.
     
  124. With such solid and undeniable proof, why does anybody shoot digital? Not being facetious or sarcastic, would just really like to know.
     
  125. That Clarkvision test really should be looked at with a high degree of skepticism. It gets posted every single time there is a film vs. digital debate for some reason.... maybe because it's got so much talky talky. The only thing that test proves is that the person who did it does not know how to operate a scanner properly. No offense to that person, but I've never seen Gold 200 behave that way except on that page. The test spends an aweful lot of time talking about how the digital photos were processed and says almost nothing about how the film was scanned, from a print or from the negative, what scanner was used, what settings where used, nothing. If it's in there somewhere, it's buried in the charts. And the funniest part is that one of Canon's most expensive cameras is being directly compared to one of Kodak's cheapest films. Yep, I've said it before and everytime I see that site getting pawned off as some sort of Final Word, I'll say it all over again.

    I think Mauro's test is a little more fair in that we get to see all of the exposures, unlike the Clarkvision test in which we have to take his word for it about his methodology. Lots of charts and a long page doesn't mean more accurate or more fair, it just means more. And of course, the real fatal flaw with all of these tests is that you are comparing one digital image to another. A scanner is just another digital camera, and I can guarantee you that I've tried my best to scan negatives that were dramatically over or under exposed to little avail despite the fact that I can clearly see the image on the negative with my own two eyes. If someone had a microscope camera rig with some sort of controllable flash below the negative, it might be possible to take correct exposure photographs of severely under or over exposed negatives to prove how much data is actually in there, but it really doesn't do much good since pretty much the only way to get color negatives printed is via a digital scanner and printer. As mentioned by others, some of the "grain" showing up in many film scans is really digital noise created by the fact that the negative is too dense for the scanner to read through, since the scanner cannot adjust it's exposure to pump more light through the negative, it produces noise instead.
     
  126. James Johnson -- : the canon 1D Mark II is a 8mp Full Frame so the sensors are a lot larger then the newer Full Frame
    cameras."

    The 1D Mark II is not full frame. Just a clarification.
     
  127. Mauro,
    thank you for an interesting article. Ignore the insults, articles like yours are much of the reason I like photonet.
     
  128. While not exactly brand new to photo.net, what has not been brought about before. I have using digital for years and only now have started playing with film, first a T50 with a 50mm FD mount, liked the results and bought a used Canon EOS 1V to use with my lens (after reading the article and the comments). So Maruo ignore the insults and thanks again.
     
  129. What about if you bracket your shots and merge to a HDR image? Oh, but I can scan my bracketed film exposures and do the same thing in Photoshop. Yeah, yeah but I can stitch enough 6.3 mega pixel images together to rival 4x5. Oh, yeah I can also whip out my 4x5 camera, shoot multiple angles, scan and stitch them to rival Jesus. But what about Phase One or Better Light backs? Think of the turn around time as a pro. Yeah but I ENJOY developing 4x5 ortho under safelight. It's well, fun. Oh, nooooooooo! What will I do! From all of this B-I-T-C-you-know-whating I'm convinced I have to choose! I guess I will smash my Crown Graphics with a sledge then go tuck my DSLR in at night. Please. You are all limiting yourselves. Shoot what you have to for a job. Shoot what you enjoy for fun and end this already.
     
  130. VERY SIMPLE:

    There is an easy way for people with better DSLRs to find out whether theirs could have provided better results.

    Just set a subject with lighting at measuring f64 with the sphere of the meter pointed at the camera, shoot it at f8 and see. It takes two seconds, much less to write what it could be.
     
  131. Also can shoot (always at f8) with lighting measuring f8 and f1 to get comparable results to the center and boundaries of the test posted.
     
  132. DON'T FORGET to shoot the same shots with Ektar to compare.
     
  133. Mauro,

    Thanks for the updated multisampled scan. Does indeed look better. Ignore the insults from some of the posting trolls. The most laughable of which were the comments of one on how Ellis had "piad his dues." Maybe.....but he was wrong.....and didn't have the courtesy of coming back and admitting it.

    All in all, the thread has been pretty civil. Like I said, if you posted results showing film is better at DPReview, you'd see the knives come out pretty quick......normally from those who have never tested ;-)
     
  134. Thanks Dave.

    Here is a picture "Shot by me" where some of the highlights were blown. It was shot with TMX100. The point is that there are some scenes with more dynamic range than even negative film can capture. The number of stops in a scene are beyond the photographer's control.

    http://shutterclick.smugmug.com/gallery/3639504_X4XUj#382039618_mcLcC-X3-LB
     
  135. Patrick,

    Keep on saying it. Its much appreciated because while some of the conclusions seem pretty accurate, the Kodak 200 gold comparison is silly. As a whole, I have always thought I liked the way digital performed in low light (its not "murky" like film) but yet Mauro has done a fantastic job at challenging that assumption. Kudos to him and you.

    Cheers,

    Steve
     
  136. I ran some Fuji Acros 100 when I was photographing at Spider Rock in Arizona. I had just run through a roll of Tech Pan exposed at iso 6. I then loaded the Acros and forgot to change the iso....effectively over-exposing the first couple of shots by 4+ stops. No blown highlights at all. Try that with a DSLR and see what you get.

    Regards,
     
  137. The objective way to measure dynamic range is to plot the output of the medium vs the input (light intensity).
    DPReview routinely measures the dynamic range of the cameras under test using a calibrated step wedge with 41
    stages 1/3rd stop apart and plotting the results. A similar test is reported in most technical data sheets for film,
    plotting exposure (absissa) vs density (ordinate), as the "characteristic curve". The absolute level of exposure is not
    important, just the range or ratio of the steps.

    There are no objective claims of a 12 stop range for a small-format digital camera. The results for the Canon 40d are
    reported in (http://www.dpreview.com/reviews/canoneos40d/page20.asp), citing a dynamic range of 9.6 stops. The
    same test for a Nikon D3 concludes a dynamic range of 8.6 stops. DPReview states their assumptions and criteria
    for adjuging the dynamic range (starting from the 2% level or noise floor, whichever is greater). In my view, this is over-
    generous. It is more realistic (IMO) to measure the range from the intersection of the slope at the inflection point with
    the noise floor, which gives a dynamic range of about 7.5 (comparable to Reala).

    I have not seen any objective tests performed on medium format digital sensors, although all of the manufacturers
    claim a 12 stop range. My own experience comparing a D2x vs.an Hasselblad CFV vs. Reala is that the CFV back
    has significantly better range than either. I have no way at present to put numbers on that observation.

    The fallacy of Mauro's test is that the figurine has a wide range of reflectivity. There is no documentation of that
    range, nor any certification of the linearity of the Sekonic 358 meter over such a wide range of exposures.
    Interpretation of this demonstration is skewed because there is tendency to rate the overall "exposure" of the figurine
    rather than the true range of reflectivities represented. One can examine the histogram of each figurine, but without
    knowing what the values should be, no conclusion can be made.
     
  138. I should expand a bit. The figurine has a wide range of reflectivity. Furthermore there is a wide (and unknown) range of light intensity striking it (at least 1/3rd is in shadow). Therefore you can't know what the reflected intensity should be.
     
  139. Edward brings up some valid points.

    "There are no objective claims of a 12 stop range for a small-format digital camera. The results for the Canon 40d are reported in (http://www.dpreview.com/reviews/canoneos40d/page20.asp), citing a dynamic range of 9.6 stops"
    - THIS IS CONSISTENT WITH MY RESULTS.

    "The fallacy of Mauro's test is that the figurine has a wide range of reflectivity"
    - AT WORST 100% REFLECTIVITY (LIKE A MIRROR) SHOULD GIVE YOU PURE WHITE AT THE SENSORS LIMIT. I SEE PURE WHITE SPOTS IN MULTIPLE OF THE STOPS. THE TEST IS VALID.

    "linearity of the Sekonic 358 meter"
    - VALID POINT. YET EKTAR AND THE D40 SHOT THE SAME SCENE (LINEAR OR NOT) AND EKTAR DID NOT HAVE BLOWN HIGHLIGHTS IN A SINGLE FRAME.

    "The objective way to measure dynamic range is to plot the output of the medium vs the input (light intensity)"
    - WE WILL HAVE TO EXPLAIN THAT TO THE FIGURINES.
     
  140. Edward, without being sarcastic, your points are valid pertaining to the ABSOLUTE precision of amount of light reflected. They do not affect the RELATIVE observation that Ektar could cope with it and the 40D couldn't.
     
  141. Can we discuss this from a technical point of view? There are fundamental differences in the mechanisms of film and digital technologies that exist which alter the characteristics in favor of film at present. Specifically, the highly linear nature of the initial step in the flow - the CCD or CMOS element converts photons to signal that is then digitized. At some point, when the gold rush for number of pixels is behind us, and perhaps once noise levels at insane ISOs is fantastic, this will be addressed. The techniques to do so are not difficult - attenuation proportional to signal level is well known, but doing so on a per pixel basis prior to initial digital conversion will be a challenge. I expect that the solution will be arrived at in an unexpected way - linearity is not really a requirement, repeatability is, so there is no fundamental law that says the response to light needs to look anything like it does today.

    A couple of ideas of theoretical approaches to this:

    1. Per pixel active ND filtering. This approach would involve an electrically active layer that was in the light path directly above the sensor. A real time feedback system would detect a threshold value in the pixel well below saturation and trigger a darkening of the layer above that pixel for the remainder of the exposure. In theory, the threshold could be set very low and the darkening proportional to the amount above the threshold. This would provide a highly non-linear response which could be converted in whatever form was programmed into the camera (the response curve).

    2. Rather than inserting a layer above the sensor, the sensor could be actively drained during the exposure. Again, this could start at an arbitrary threshold and in theory could be done proportionally to the amount above the threshold or inversely proportional to the remaining headroom. This is probably a bigger electronic circuit challenge, but a smaller manufacturing challenge.

    I would note that similar (similar as in analogous) things take place and explain what happens in film. The rolloff in response as I understand it is an artifact of non uniform particle size in the emulsion. An element may have a different depth in the film, a projection in the light path and different absolute sensitivity. Those things create the non linear response that we seek to replicate in digital photography. From my perspective, aesthetically, digital is different than film. Not better or worse, but different, and as a result requires different technique. For those so inclined, it is not technically impossible for the sensor manufacturers to create sensors that behave identically to film. In fact, the fact that there are differences in films tell me that manufacturers could hawk cameras that behave like one film OR another and thus expand the market again as people want a Velvia camera, Ektar, etc. This is Moore's law put to good use in my view.
     
  142. In this thread, there also seems to be a confusion between exposure "latitude" and dynamic range. The OP and
    most of the responders appear to refer to "latitude".

    Dynamic range is the range of exposure over which the film or medium can produce a useful range of output (e.g.,
    density). Latitude is a convenient "fiction" expressing the deviations from the "correct" exposure which can produce a
    useable image or print. Indeed, the "correct" exposure is usually a subjective decision (q.v., the Zone System).

    "Latitude" is a composite of the dynamic range and the range of light intensity reflected by the subject. Furthermore,
    the range of the subject is subjective to the extent of what is important in the subject. It's generally accepted that
    some shadows can be featureless black and some highlights (e.g., specular reflections of the sun) can be
    featureless white. The range of the subject expands or contracts depending on where you want to see some level of
    detail.
     
  143. Mauro,

    I was careful not to indicate an "absolute" exposure, rather the range or ratio of high to low and (perhaps) steps in
    between.

    The tendency of digital capture to "blow" highlights is well documented, and is comparable to that of reversal film -
    about 2 stops middle grey. Negative film is much more forgiving toward highlights than either. In RAW images, it is
    rare that all three channels are overexposed. If any one channel is in range, a considerable amount of overexposure
    can be recovered through interpolation. This amounts to an as much as additional two stops for a typical scene.

    I refer to my followup post regarding the confusion between range and latitude.
     
  144. Agree it seems to be a known fact (although some photographers are not convinced). It is still good to run up-to-date comparisons as sensors evolve and new films appear, as in the future, digital sensors may match or surpass the ability of negative film to retain detail in the highlights.
     
  145. Eric A,

    I don't see anything wrong with Mauro wanting to do a test of his own, even if it has been done 100 times before. This is all part of the learning process.

    As far as the subscriber vs non-subscriber thing goes.... if photo.net didn't want non-subscribers to be able to post, then non-subscribers wouldn't have the option. Honestly, I see Mauro as a potential-subscriber, rather than a non-subscriber. He is testing this community out before he decides to pay for a subscription. And unfortunately, responses such as yours, Eric, can deter people from subscribing. Thank god the people who are interested in participating in meaningful, helpful discussion outweigh the ones who want to bite someone's head off for doing something on their own.

    I would also like to note that I am impressed with Muaro's ability to keep his cool and not return insults after receiving personal attacks.
     
  146. Tim Holte,

    From what I have read most Professionals use Digital over film because that is what their clients want.

    I have a feeling that many amateurs are using digital because of the cost savings over film.

    Personally, I use digital for both cost savings and convenience. Film+Processing can get expensive. I also like being able to run home and pop my flash card into the PC and immediately upload my images.

    I've honestly thought about picking up a film camera and taking the Intro to Photo class at Montana State University because they teach you how to develop your film in the dark room. Why? Because I think it is something I would enjoy learning.
     
  147. One more thing...

    About the Clarkvision test....

    The 1D Mark II is NOT a full frame sensor camera. The 1D sensor measures 28.7 x 19.1mm.

    Its is the 1Ds that has the full frame sensor. I know a couple of Pro photogs who are still using the 1Ds Mark I because of its full frame sensor and lower megapixels. Plus the fact that it will sync with a flash to 1/500th. Check out Al Berger's website at http://www.prorodeopix.com/. He uses a 1Ds Mark I with a 300 f/2.8, and I'm pretty sure all of his photos are available light. I had asked him if was going to upgrade to a Mark III when it came out, and he replied "Why would I downgrade my camera?"

    Just thought I would toss that in, because I thought I had saw someone say the 1D was a full frame camera.
     
  148. "I would also like to note that I am impressed with Muaro's ability to keep his cool and not return insults
    after receiving personal attacks."

    That has been very impressive, indeed. And it shows he's got more character than the know-it-alls who
    hurl unfounded insults and then disappear into the night.
     
  149. Hello all, I shoot digital and still develop my own b&w film. Thank you for the information in this thread. I
    don't understand why some individuals would sound as if they are personally offended when others discuss a topic
    which does not interest them. I bet they do not try to act like bullies when meeting someone face to face.
     
  150. Mauro,
    Thanks for all the effort. Information like this is always useful. I primarily use color film these days, with a little P&S digital ready at all times to catch the kids doing the kids' stuff.

    I look forward to the time I can ditch the film equipment. I'm hoping that digital full-frame sensors will achieve the dynamic range of film (or better) at a reasonable cost (under $1000) within the next three years. Possible? What do you all think?
    Jamie
     
  151. The fact that negative film has better range than digital sensors or slide film is common knowledge.
    Another fact however is that a tonal range does not a picture make.
     
  152. Thank you all. Roger and a couple other people have downloaded the RAW files I posted to see if they could rescue the detail in the brightest stop of the 40D using software. Stay tuned for the results.
     
  153. Mauro the TMAX 100 shot with the blown highlights is that a scan or a traditional B&W print.
     
  154. Just a scan. I didn't tinker with it much since the Coolscan does a good job at it. The scene was so bright I couldn't look at it directly and the surroundings were almost pitch black. I think it was one of those rare exceptions were TMX got out-ranged.
     
  155. I wonder how much would burn in if you made a traditional darkroom print.
     
  156. I had a brief play with your 40d highlight raw. I only had a brief play as my wife is monopolising our computer for work, and I didn't have much time. I will try and post the results tonight when I get home. Briefly, some interesting things I found: Doing 'linear' processing, I found that DPP doesn't make full use of the sensor's data (this is something I knew already anyway). The Linear dcraw and lightroom processing shows a little bit more detail (though not anything like a stop's worth) than the linear DPP. I was fairly impressed with DPP's ability to regain some highlight detail using a contrast of -4. The main problem I have with DPP is not knowing what's going on under the hood. The exposure adjustment seems to be some sort of combined exposure and highlight recovery tool. Also when I developed linearly and boosted the shadows, it was clear there was some shoddy noise reduction going on, even though I had noise reduction set to 0. The real interesting (and unexpected) thing for me was the seeming quality advantage of dcraw over lightroom in the shadows. Although, I will try to confirm this with a bit better processing tonight (last night I was rushed, and our computer is acting like it has the weight of the world on it's shoulders i.e. it's chugging along). Mauro, I wonder if you could post the 40D shadow raw, as I would like to see the difference between the three processors, and compare to the film scan. I feel somewhat positive that dcraw could dig a fair bit of detail out of the shadows.
    Below is a quick preview of some of the results. These are a bit hapazard, as I was rushed and the computer kept crashing. From memory, the DPP image was converted as srgb (coz I couldn't figure out in a short time-frame how the hell to output the damn thing in prophoto). And the dcraw image from memory is probably still in prophoto. Top image is DPP no exposure adjustment, white balance as shot, faithful or neutral picture style (can't remember which one), -4 contrast; bottom image is linear dcraw initially no white balance, imported to lightroom as 16bit prophoto tiff, fill light 100, white point set using positive exposure (nb. dcraw default white clipping point is higher than the 40d pixel saturation point as far as I can tell), then exported to photoshop and a quick-and-dirty two-click curves white balance adjustment (as some of you may or may not know, the linear image off the sensor is green). Tonight I will try and process the dcraw image in dcraw itself as much as possible. This should make a bit of a difference as white-balance and exposure adjustments post-demosaic is not an optimal approach. In summary: notice the little bit of extra detail in the shadows and the highlights of the dcraw vs dpp image.
    00RcP3-92457584.jpg
     
  157. Probably the best way to have saved it would have been with a little pulling in development and using TMAX DEV instead of XTOL.
     
  158. Thank you for the work Bernie.

    It still looks like 3-4 stops have areas with clipping beyond recovery although you dug up important detail. I'd say Ektar has at least a 4 stop advantage on the bright side over the 40D since even the f64 frame was not enough to compromise the highlights.

    Here is the raw file with the shadows:

    http://www.yousendit.com/download/TTZuS3dqMGMwMEZjR0E9PQ
     
  159. Thanks Mauro. I'll have a play over the next few nights. You're right about the improvement; the 40D will never match negative film.

    One query I had - your initial results you show, what settings did you have for the 40D dpp image? When I open it in dpp, it opens with +2 contrast and +2 or so saturation. What settings did your initial posted images have (including what picture style)? Cheers.
     
  160. Those are the settings. I did not touch the raw at all to either post or send to you.
     
  161. Please check against the post on your monitor to double check.
     
  162. j_a

    j_a

    Mauro,

    You have my apologies for my earlier posts. They were off base and frankly I came off sounding like a bit of a crank.

    My criticism of your comparison still stands however. You're trying to compare two radically different technologies. You can compare trains and trucks, they do the same thing (move things) but they do it in different ways. But the comparisons will always fall short. Digital cameras and film cameras both capture light, but they do it in different ways. Both benefit from inherent strengths as well as inherent weakness. And the most notable of both technologies weaknesses is the first steps after exposure.

    In the case of digital (raw or not) the software used to view the image has a direct effect on the image itself. Raw just leaves all the info there, it still needs to be interpreted by a software in order to be viewed. Some software's do this better than others. You need to list what software you used and the setting used by that software.

    Let me take a step backwards with film and begin by stating the first inherent weakness' is in the actual material. Film, relying on chemistry and not physics is subject to a number of factors that could alter the exposure: age, temperature and humidity in storage, manufacturing quality control and handling. Then there's developing process itself. Every strip of film run thru a developer is going to be minutely different than the preceding strip.

    So the first part of my critique of your comparison is the potential for +/- errors in the comparison samples.

    My next criticism of your comparison was touched on by another poster. Limited scope and simplistic labeling. Your comparison specified one digital camera and one type of film. You simply can not make a sound claim that, "negative film has a wider dynamic range than digital SLRs." If every digital SLR had been tested and compared you might be closer to making that claim, but not all the way.

    The next problem with your comparison is in making the assumption that all things are equal. They aren't. I'll explain by way of example. I own a Nikon D200 and a Nikon D300. You might think that, all things being equal, that ISO100 on the newer and greatly improved D300 would be better than ISO100 on the D200. You'd be wrong. In fact, or at least in observance, the D200 out performs the D300 at ISO100 and ISO200. However at ISO400 and above its no contest the D300 wins hands down. The same problem with not all things being equal is in effect in your comparison. You compared film ISO versus digital ISO. The way a digital camera defines ISO is just too different than the way film handles ISO (ISO being originally introduced as a measuring tape for the sensitivity of light sensitive chemistry has been shoehorned into use for light sensitive physics). As a stark example of the difference I'm referring to simply compare an images shot on 3200 speed film to a image shot on a D3 at ISO3200. No contest, the D3 will produce far superior images. The same contest at ISO100 will may produce different results.

    My last (thankfully eh?) criticism of you comparison is you have a built in bias as evidence by your previous posts and website. That bias may have colored your research and skewed your conclusions. Its a lot like asking someone who sells Ford's for a living to make a scientific comparison between Chevy's and Ford's.

    In summary, your comparison suffers from too many faults.

    I have no doubt you tried to make a sound argument for film. Its just that it is a flawed argument, such as I've seen hundreds of times on Pnet (hence my crankiness). But I do encourage you to keep trying. For if nothing else we've learned that at ISO100, Ektar100 beats a D40 in dynamic range.
     
  163. All good J A.

    "So the first part of my critique of your comparison is the potential for +/- errors in the comparison samples"
    HERE IT IS NOT ABOUT MAKING THE IMAGES COMPARABLE BUT TO UNDERSTAND THE POINT WHERE INFORMATION IS NO LONGER RECORDED. WHEN YOU RECORD PURE, THERE IS NOTHING TO RECOVER IN RAW PROCESSING.

    "You simply can not make a sound claim that, "negative film has a wider dynamic range than digital SLRs." If every digital SLR had been tested and compared you might be closer to making that claim, but not all the way."

    TRUE.

    "My last (thankfully eh?) criticism of you comparison is you have a built in bias as evidence by your previous posts and website. That bias may have colored your research and skewed your conclusions. Its a lot like asking someone who sells Ford's for a living to make a scientific comparison between Chevy's and Ford's."

    GOOD POINT. THOUGH I TRY FOR MY TESTS TO BE AS OBJECTIVE AS I CAN (SO I CAN PICK THE BEST MEDIUM LATER AND BECOME EDUCATEDLY BIAS FOR IT).

    DIGITAL IS SUPERIOR THAN FILM IN MANY ASPECTS. IN RESOLUTION AND DYNAMIC RANGE LOW SPEED FILM STILL HAS AN EDGE. DEPENDING WHAT IS IMPORTANT TO YOU, YOU MAY SHOOT ONE OR THE OTHER, OR BOTH (LIKE ME).

    "For if nothing else we've learned that at ISO100, Ektar100 beats a D40 in dynamic range."

    CONCLUSION WELL PUT.
     
  164. j_a

    j_a

    Mauro,
    I should explain that somehow I logged in under a very old account and I'm the a-hole Eric that bit your head off for no particularly good reason. Again, my apologies. Sometimes my bad day shows up in my posts.
     
  165. No problem. I like your J A side better.
     
  166. Hi Mauro, I'm afraid that so many people took your test as the ultimate match between digital VS film!
    I understood the reason of your test: I shoot film and scan it with a CoolScan 5000 ED and I'm amazed by the tone richness I achieve from some negative or B&W film. But I stop here. Film and digital are two different technologies to record the light of a scene with their strong and weak points... Then all is filtered by our eyes and... our brain... and the result is always different from reality. And it differs from person to person. The comparison may be useful but I prefer to understand how taking advantage of these technologies to express my vision, rather to worry about something that is always a limited medium.
    Kind regards, Alberto.
     
  167. Alberto, you are right, the goal is not to beat one over the other but to share information. Even if you shoot digital exclusively, understanding its characteristics can help you produce better results, make educated compromises and ultimately be more effective in realizing your vision.

    Digital has many advantages of its own and in a short time has dominated the mainstream. It's future couldn't be brighter (fun intended).
     
  168. Lowering contrast in both DPP and/or LR, along with highlight recovery, does bring back some considerable highlight detail, but always with the potential for hue shifts & saturation losses:
    Here's the original LR conversion:
    [​IMG]
    Link to Full-Size Image
    And here's what I did to try to recover some highlight detail in Lightroom:
    [​IMG]
    Link to Full-Size Image
    Still, nowhere near as good as the Ektar100...
    Rishi
     
  169. "You shot Jpegs didn't you? -- proving nothing more than your ignorance."

    DSLR PMS :) Stop hating. Self discovery is a great journey. And the fact the OP is sharing is a privilege. The OP will probably be taking better pictures as he finds insights in his tests. And if all this does not interest you, just hit the back button and try not to bite off heads :)

    Thanks OP!
     
  170. Thank you Rishi, that is a very good job.

    Are the white areas in the second image void of information in the RAW file or is there more detail to be recovered?
     
  171. Thanks Juan.
     
  172. What I don't understand, though, Mauro is this (and as a disclaimer, I'm not saying the following in a harsh tone):
    For all your talk & examples of how 35mm clearly outresolves digital (a la http://www.photo.net/film-and-processing-forum/00RV7N), how come in this particular test, the 40D image, upsampled (using bicubic algorithm in Photoshop) to the resolution of the Ektar 100 scan (which I take it was 4000 dpi on a LS-9000?), clearly looks sharper?
    Ektar 100:
    [​IMG]
    Link to Full-Size Image
    Canon 40D:
    [​IMG]
    Link to Full-Size Image
    Please view images at 100% to make a fair comparison.
    I understand that local contrast has an effect on perception of sharpness; negative images inherently have less local contrast, though you do try to tease out local contrast when expanding the tonal range of negative film from the 'flat' scan (generally I do this myself, as I don't trust NikonScan or Vuescan to make my vision).
    But, I mean, given your last post that basically said a 26MP theoretical digital SLR couldn't even match 35mm Provia, how come the 10MP real digital SLR here, set up so that the same area is covered as the full frame film shot, seems to resolve at least what the film resolved?
    Please explain.
    Thanks,
    Rishi
     
  173. Mauro,

    Those white areas seem to be just completely void of any information, according to Lightroom. Of course, I haven't read the raw data off the RAW file myself... but I trust LR :)

    Really looking forward to your response to my last post!

    Cheers,
    Rishi
     
  174. No need to apologize Rishi. It is good you question these things. That is why I open the discussions.

    First since I was not intending to test for detail, the figurines where not perfectly aligned on a plane parallel to the film plane. This, combined from the narrower depth of field, definitelly would compromise detail are fine contrast.

    Nonetheless, even in these conditions, with some quick sharpening and contrast this is what the Ektar crop you posted looks like.

    http://shutterclick.smugmug.com/gallery/6616619_YJEwK#425581323_3ygce-O-LB
     
  175. I don't even have LR but I trust you. It appear the 40D starts clipping highlights at least 3-4 stops before Ektar.
     
  176. Mauro -- narrower depth of field... I understand that full-frames have narrower depth of field at equivalent apertures (I think), but at f/8 -- do you really think it'd make that much difference?

    Reason I'm particularly skeptical is because in my own subjective tests, I just don't see my film images appearing sharper than a 12MP Canon 5D. Which is why I find the results from the images above much more on par with my experience, but the results from your '26MP' thread unbelievable. Although I'd love to believe that '26MP' thread because I still shoot & love film :)

    Rishi
     
  177. Just a quick word about sharpness, Rishi - there is a difference (and I am sure you know this) between sharpness and resolution. Imagine a square wave - sharpness has to do with the shape of the edges - how quickly the wave goes from one value to the other. A quick transition translates to higher sharpness. Resolution is about how many low-high transitions you can fit per unit length. Digital almost always appears to have higher sharpness because pixels already have hard edges, allowing sharp transitions.

    That said, the Ektar shots actually seem too soft - maybe they were not exactly in focus as Mauro says. Also, smaller sensors have greater depth of field and Mauro used strobes measured with a hand held light meter. This probably means that he set both cameras to the same aperture resulting in greater DoF with the digital.
     
  178. Rishi, yes at f8 and probably a just couple yards from the subject, DoF to evaluate very small detail (at the resolution limit) is too narrow. Especially when all the subjects are that close and not in an exact line.

    I don't know how to pull the exact focusing distance out of the RAW file but maybe you can.

    This is shot with the exact same equipment, lenses, film, etc as the dR test but paying attention to the focusing and DoF:

    Noticed TMX outresolves the scanner. TMX outresolves Ektar. Ektar outresolves the 40D. And Velvia is pending:

    \http://shutterclick.smugmug.com/gallery/6616619_YJEwK#421902416_ibL4V-O-LB

    TMAX next to 40D at 200%:
    http://shutterclick.smugmug.com/gallery/6560295_hraSq#417849621_RBnJg-O-LB

    --------

    I'D RATHER NOT TAKE THE DR DISCUSSION TO DETAIL. FIRST RESOLUTION IS A LOT EASIER TO TEST THAN THIS DR EXERCISE (getting the dummies similarly spaced and one stop apart is not easy), SECOND RESOLUTION OF ALL 100 ISO FILMS I USED IN TESTS OUTRESOLVED THE 40D CLEARLY AND BY A LARGE MARGIN.
     
  179. Also if you want to use this test for resolution I'd have to rescan without ICE, etc, etc.
     
  180. Vijay! Good to see you again :)

    Yes, good point, I understand that about digital -- that point ties in with what I mentioned about the inherent 'lack of local contrast' in film, yes? Although, arguably that's better with positive film due to its exponential response to light.

    Anyway, Vijay, Mauro, or anyone else -- really? Even at f/8 you'd expect that much softness from lack of depth of field?

    Rishi
     
  181. Oh ok Mauro just posted before your response.

    And ICE does not affect sharpness at 'Normal' (NikonScan) or 'Light' (Vuescan)!!!

    :)

    Rishi
     
  182. Oh yes I did see that post. Bicubic upscaling would help your 40D but wouldn't get any more resolution off the charts.

    Can't wait for the Velvia 50 results, as that is what I shoot :)
     
  183. Should get them tomorrow. I'm sure Velvia 50 will outresolve the Coolscan (since I saw that already in my 6x7 test) but it'd be nice how it compares side by side to Ektar.
     
  184. Rishi, I hate keep correcting. I didn't think ICE normal made a difference until not long ago but as usual I test everything myself:

    http://shutterclick.smugmug.com/gallery/6616619_YJEwK#425687124_i7c3H-O-LB
     
  185. Rishi... couple of points:

    As Vijay said the 40D will have greater DOF for the same f-stop and field of view than the 35mm.

    Did you have sharpening turned off in lightroom?

    And for both Vijay and Rishi.... remember this isn't fully a film vs digital comparison, as the film has been scanned.
    It is really figital vs digital....;) ie. the sharpness due to square pixels should be the same between the two.

    Another general point to make (this includes you Mauro), is the raw conversion factor. As you can see from my
    couple of examples, the choice of raw converter and the settings can affect both dynamic range and detail. In your
    original images Mauro it seems you have used dpp settings of +2 contrast and saturation, and a picture style that
    may be not optimal (I can't remember what the setting actually was). As far as I am concerned, DPP is basically
    useless for these types of comparisons, as it can't present you the more-or-less untouched raw data. It is doing it's
    own things behind the scenes, which some marketer at Canon deems is good for you. To improve on these tests in
    the future, you should use dcraw (or Lightroom/ACR at the very least), and shoot something that will give you a
    better determination of USABLE dynamic range.

    By the way, you won't get anymore highlight detail than what you see on my dcraw conversion. This conversion
    includes ALL the data coming off the sensor (minus of course the noise below the black point clipping point).
     
  186. I had no clue how good/bad DPP was at this - that is why I posted the raw for people to take a stab at it.

    In DPP: Minimizing contrast and plus -2 exposure adjustment gave me a good idea that at least 3-4 stops had blown areas.

    WITH THIS TEST - and everyone's collaboration; we all got a very quantifiable idea on the D-range and clipping thresholds of the 40D vs Ektar.

    I'd give the 40D 8-9 stops vs at 14 stops for Ektar. I was surprised how well Ektar did considering how contrasty and saturated it is. I was also positively impressed on how well Ektar did in the shadows.
     
  187. If only I could get my hands on some Ektar. I've only worked on scans from other people. Shops in the US wanted to charge $44 in shipping for 5 rolls of film....yes, I call that gouging! For some reason, Kodak hasn't rolled it out here in Canada yet.
     
  188. I order it from Amazon. I can send you a few rolls. Email me your postal address.
     
  189. Wait, & I forgot: Bernie!! Good to see you again :) I feel like we've all bonded on that last thread.

    Mauro -- OK, visually, without resolution test charts, no normal human being could see that difference brought on by ICE Normal :) But, yes, I do see a difference... a bit less is resolved with ICE on, but, on the other hand, surprisingly it seems that there's a tiny increase in contrast or sharpness. Funny enough, I thought I saw this on my own comparisons of ICE on/off with regular images... but I put it out of my mind b/c I couldn't rationalize ICE actually increasing sharpness.

    Also, as I mentioned earlier (I think), the extra stops on Ektar is really not surprising. Not only do the layers get blacker as exposure increases, making more exposure harder -- but also the sensitivity specks on the surface of the grain start accumulating electrons, thereby repelling more electrons being knocked off of bromide ions by photon absorption, making the process of collecting electrons to reduce silver ions less and less efficient. Which amounts to harder and harder exposure as exposure continues.

    Anyway, I still find ICE Normal quite acceptable.

    Cheers,
    Rishi
     
  190. But we are not normal. we are perfectionists... ha ha

    I usually prefer ICE fine because it has a smoothing effect without sacrificing much detail.

    Please post your results of your ICE test. I am very interested in contrasting results.

    This ICE exercise should answer your question on detail since we are moving in-between 3600-3900 lines per picture height for 35mm with the Coolscan and the 40D captures barely over 2000.


    It is the Holiday spirit rubbing off and making us bond...
     
  191. Rishi: Wait, & I forgot: Bernie!! Good to see you again :) I feel like we've all bonded on that last thread.

    Stockholm syndrome?

    Kidding, buddy, kidding :)))
     
  192. How about this for dynamic range? Cheap Canon Rebel G with 50mm 1.4.HP5 No manipulation in PS. Back using film and loving every minute of it !!!! " While my 20D gently weeps "
    00Rd2G-92765784.jpg
     
  193. That sounds good to me (that used to be my favorite combo too). Also Greg, order yourself a brand new Canon Elan 7ne before they go out of production - only $350 at BH with metal body, MIRROR LOCKUP, eye control, etc. You'll regret it if you don't.
     
  194. Rishi, I'd benn tinkering with Ektar and the different levels of ICE. I', almost tempted to weite an article with the findings but people will get tired of me.

    In short, NO ICE is my clear preference with Ektar. I can have a masked layer with ICE FINE to get rid of any spots.
     
  195. Man I had a Elan 7 and sold it a while back, regret it now, then again it wasn't a 7NE. Will check it out on the Bay. Cheers.
     
  196. Re Roger: In 1984 we didn't have the technology to scan the Ektar 25 negs

    There were early drum scanners in 1984; but no Ektar 25; since it came out in 1988 to 1989. Thus this might prove some existance of a time traveler here? ie scfi or x-files stuff!
     
  197. "Stockholm syndrome?"
    Wait, Vijay, I don't get it -- who's my victimizer? Or was the thread that traumatic for everyone involved? :)
    Mauro -- if you get a chance, you should post your ICE results. Certainly the difference between 'off' and 'ICE Fine' are compelling (for the use of 'off' :)
    Also, if you get a chance to take that Velvia 50 shot of the resolution chart, and do the same 'off' vs. 'Normal' vs. 'Fine' -- that'd be awesome. I wonder if it'd have as much of an effect on positive film given the extended tonal range of a positive compared to a negative (tonal range of data actually on the film)... pretty much shooting in the dark here, as I have no real reason to believe it'd affect one more than the other, other than noticing how film defects (holes, scratches, dust, etc.) affect negative film much more than they affect positive film because when there's a, say, scratch or hole, a larger final tonal range (after inverting & processing the negative) is affected since that scratch/hole existed in a medium with a very 'flat' tonal range (and since that 'flat' tonal range is later extended, any defects are amplified). I don't know if that was a good explanation. Hopefully someone gets it.
    Rishi
     
  198. Rishi: "Wait, Vijay, I don't get it -- who's my victimizer? Or was the thread that traumatic for everyone involved? :) "
    The latter - everybody felt traumatized by everybody else and now everybody is bonding... where's Petrana?
     
  199. PHOTO.NET has several main threads that have been rehashed over the last decade many thousands of times. <BR><BR>(1) Folks buy a gooberflex digital and ask how BIG can I enlarge the image.<BR><BR>(2) Folks learn that slides have less dynamic range than a negative.<BR><BR>(3) Folks learn that many digital sensors have a less dynamic range than a negative.<BR><BR>(4)Folks learn that a raw digital file can have more range than a jpeg<BR><BR>(5) Folks who post giant images on the web and learn that they got borrowed and stolen<BR><BR>(6)Folks who learn that having more ram makes their Photoshop computer work better<BR><BR>(7)Folks who learn that a high contrast lens test shot with camera bolted to a granite block give great resolution numbers to brag about.<BR><BR>(8) Folks who learn that disc drive can carsh and "learn" that they should havce had a duplicate file<BR><BR>(9) ZILLIONS OF FOLKs who are are lost souls; who crave an exact megapixel equalvalent for 35mm film; often posting results. The same folks probably want and can prove the number of MP3's per 13 year old girl; shoes per woman, lenses per photo.neter, or beers per football game. <BR><BR>(10) Folks who buy a lens with a dinky scratch and ask others if it a matters; but do not have the ability to shoot few frames as their own test<BR><BR>(11) Folks who crave ratings on posted photos; but complain when they get negative feedback.<BR><BR>(12) More threads about film versus digital; posted like its a new subject; when its several decades old now.<BR><BR>(13) Folks who discover that scanning takes time; when its a 2 decade old known fact.<BR><BR>(14) Folks who think that cleaning off fungus from a lens "magically" removes the potholes that are etched in advanced cases of fungus.<BR><BR>(15) Folks who worry about Macs versus PC's instead of worrying about results delivered to clients.
     
  200. Velvia seems to like ICE better in my opinion too. I will post the three shots on Velvia 50 on Saturday when I get them back.

    ICE TEST ON EKTAR: A good way to observe the differences in ICE is to have the 3 layers lined up and play turning them on and off (I already prepared it in photoshop). So I posted a photoshop file - 22MB. All layers have equal sharpening.


    https://www.yousendit.com/download/TTZtQmtRdWNqV0R2Wmc9PQ
     
  201. Truly insightful, Kelly. Just a few comments.
    (8) Folks who learn that disc drive can carsh and "learn" that they should havce had a duplicate file
    Dang. Add the cost of a RAID NAS server to the cost of that digital camera. Yeah, you can backup, but I'm lazy.
    (9) The same folks probably want and can prove the number of MP3's per 13 year old girl; shoes per woman.
    Well, if you have a daughter or a wife, or planning to get either - those numbers are very meaningful.
    (15) Folks who worry about Macs versus PC's instead of worrying about results delivered to clients.
    Those who value their own sanity generally prefer Macs (or Unixes in general).
     
  202. You forgot:

    (16) Folks who relentlessly deny results put in front of their faces by posting a link to Luminous Landscape's opinion on the subject.
     
  203. Film has log color but it's still scanned by a linear scanner. There's no such thing as film. Everything is
    scanned & viewed on a computer eventually, using the exact same linear photodiode technology. It's should be no
    problem
    to make digital camera sensors capture exactly the same image as a film scanner but it hasn't happened. Scanned
    film still has better color than digital sensors. Maybe its because they can average data points over a
    longer time than digital sensors. Maybe projecting artificial light through a negative allows more information
    to be captured than directly sensing natural light. In the mean time, U might be better off shooting film &
    scanning it if U really want the best color.
     
  204. Kelly,

    Great list. Number (10) takes many different forms and apparitions and is frankly the one that annoys me the most.
    Anyhow, from someone who has read the posts here off and on since 2001, I do find Mauro's post here extremely
    well illustrated and well put (maybe now the count is a zillion plus one -- i.e. including me?). I would even argue that
    it is also of a high degree of utility.

    Tonight, I was shooting a school stage show with my M8 and was, of course, having to be so careful with the
    exposure (it was generally set at ISO 320 with -1 exposure compensation with the 28f2 wide open) because of the
    dramatic contrasts. Once the highlight is blown, thats it. And I remembered that I used to used Kodachrome 200 for
    such purposes (no such issues). Anyhow, I was later thinking about this precise article at the time and its relevance
    to what I was trying to do. Should I have shot film? Well, no film at an effective speed of 640 is near that good.
    Second, I still prefer the immediacy of the digital feedback in such circumstances (and funny enough, when the head
    teacher came over afterwards to see some of the pictures, she then me asked to email the pictures to the school
    because they were better than the professional ones that had been shot earlier at one of the rehearsals). However, if
    you were to do some very professional and controlled shooting, Mauro's experiment demonstrably points out the
    advantage of film for many such circumstances.

    And yes, Mauro, people would rather cite authority than use their own brains to understand the actual evidence
    presented to judge the case. It some cases, its understandable (we can't always analyze everything; experts can
    be "efficient".) Thats why your post is so effective -- it actually uses photography to clearly and effectively illustrate
    the comparison. Masterfully presented. I find it remarkable though that someone can look at it, ignore the prima
    facia evidence, and then cite the same Luminous Landscape report (which later Michael admits is somewhat in error,
    but brushes it off since he so prefers the direct digital chain). Even worse, the type of responses like Eric A's above
    (how
    dare you post any contravening evidence to what the elders have decided?). Anyway, we're all human and
    photography (at least for some) is supposed to be fun. You've really helped out here in that pursuit.

    Cheers,

    Steve
     
  205. Re Mauro's request for ICE examples. Here is a comparison of a Kodachrome 64 slide scanned with a Nikon 5000 using Nikonscan with ICE on and off: link. The full-size image is 884 pixels wide. Look at the ice axe wrist strap in particular.
    Best,
    Helen
     
  206. I got interrupted: I meant to add that the effect of ICE on unblemished film seems to be related to density, not surprisingly. The dyes are not totally transparent to IR.

    Best,
    Helen
     
  207. Wonderful thread, and more like preaching to the choir for me. As the company photographer, I shoot with a 40D,
    a 1Ds, and an FTBn, a T70 (I like the FD len's), and Mamiya 645. For quick and dirty or same day turn around,
    I'll shoot with the digitals (and prefer the 1Ds for portraits), but only if I'm fulfilling a single or double
    page spead on a product brochure.

    However, when it comes to landscapes, large studio sets, or something in the 30" print range it gotta be film.
    Yes, it's extra work, and yes it can be a little more in cost over digital. However, after a full day of setup,
    or trudging through the woods, I would feel more confident having a negative over a digital RAW file any day.

    So, I agree, more DR is where I'd like to see digital achieve. Although I recognize it's merits, I am willing to
    put up with film if I know I can get a better image in the end, and that's what it's all about it.
     
  208. Thank you for the post Helen. The ICE on the Coolscan 9000 (what I have) is pretty good on Kodachrome.

    There is not a single answer on what Coolscan ICE levels do to detail. The answers are very specific to each film. It honors probably a proper article so people can use as a reference. I will try to post next week.
     
  209. Thanks Helen. Nice example.
     
  210. Arrived somewhat late here...a lot of points were already made...so let a picture speak.

    The attached pic has been taken by some Fuji device...but do you think it was slide, negative or digital?
    (The only processing that I applied to the photo was some de-saturation of the right side). No HDR stuff.
    Framing, taking one shot, and period.

    Notice the reflection of the sun at the right - while taking the photo it appeared as being blown out to my eyes.
    Shooting cond's: sun in the back, no clouds. Nothing more, just guess. In the end everything boils down to
    delivering pleasing pictures.

    Finally, www.dxomark.com says that your Canon 40D has 11.3 stops of DR, www.diwa-labs.com says it has 10.22 -
    10.90 (RGBY),
    and dpreview rates it as having 11 stops, "although with no guarantee of color accuracy".

    Cheers,
    S.
     
  211. Sorry, my browser went nuts...here the picture.
    00RdNr-92971584.jpg
     
  212. Here with printing we have many examples of prints large and small; with inputs from barbie cam to drum scanned 4x5. <BR><BR>Once the *source* of the print or giant poser is revealed; folks often are more critcal; ie to fit their agenda. <BR><BR>Its like revealing the race, creed, country or origin; religion, hair color, height of a person place or thing and folks jump on the dogma bandwagon to preach their biased agendas.<BR><BR><BR><BR>Thus with Svenja's great example above folks typically are more vocal once the source is revealed; this based on us here having samples going back for about 1.5 decades. One can have a poster shot with a cellphone and folks will like it if you say its film; and hate it if yoiu say it was shot with a cellphone; "since a cellphone is never a camera; and never will be" according to many folks. Its actually an interesting exercise in human bias.
     
  213. This is an interesting thread and a good experiment indeed.

    That said, I'm really not too fussed one way or another - I mean I like shooting black and white film and the odd slide but that doesn't mean everyone should. I'm not more right than the next person. There again I love the processing + printing bit in the darkroom but again it's not for everybody and the thing I applaud most is a good photo whatever it was taken on. The point is if you have an eye for it, you'll get it no matter what you use, including camera-phone; which I too use from time to time.

    All I want to see is people enjoying their photography and the results, and two hoots to how they got there :) There again I use Macs and PCs in equal measure and like both so you know, I'm quite weird. I don't know why I don't just go and get two religions whilst I'm at it to follow :))
     
  214. zomg.

    Shouldn't you people be out shooting? I mean, why does anyone still care about this? It's so totally ridiculous- great photographers shoot fancy cameras, they shoot crappy cameras. They shoot film, they shoot digital. They print palladium, they print inkjet.

    Maybe some film has more DR than some digital, maybe some other digital has more DR than some other film...


    ...and maybe DR isn't the most important part of photography and so it totally doesn't matter at all. Stop dicking around and go make some pictures already. This debate is the posterchild for all of the stupidest aspects of photography.
     
  215. I think it's great that you did the test, if that's what you like to do. But isn't this old news? And the differences.. how many good photos will they make or break?

    The best part of the film thing IMO, is that you can stick that stuff in something as small as an Olympus Stylus Epic, and other compacts, and still have access to all the good stuff, including "full frame", that 35mm films have to offer. Digital just isn't able to offer that in the compact small sensor cameras.

    That said, I'm going outside with my XA right now and do some photography.
     
  216. Svenja, the picture you posted seems to have a small area of pure white. Since it is a flat object and there is no detail to be had there anyway (it'd be different if it were a person's eye), plus it may have been your creative intention, I see nothing wrong with it.

    Regarding whether it is film or digital, both give you the ability to blow highlights, with negative film it is harder but you can always blow them in Photoshop if that is your intention.
     
  217. "This debate is the posterchild for all of the stupidest aspects of photography..."

    Mr. Farwell, it's a discussion, take it easy. Most folks here use both formats, and that's the just it, a format, a look, or another tool in the old box. If you don't like reading such threads, then go take some pictures, (like you said... :)
     
  218. Sorry- got a bit carried away... Teaching photo and trying to get people to realize that good art doesn't come from equipment makes me a little batty at times... Going back to photoshopping my MF scans and waiting for my new digital camera to come in (at which point I will gladly trade the 7000x7000 pixels of my hassy scan for not having to make the scan in the first place. ;)
    00RdYf-93053584.jpg
     
  219. Digital definitely has more dynamic range than Film. Expose the bg, expose the fg, expose the water, expose the
    sky, expose whatever....blend them all in photoshop and you have HDR :) hahaha film blows and digital rules. Ok
    now go ahead and crucify me.
     
  220. Sinh,

    Easy.....blend a bunch of film shots for HDR......ha ha ha.....digital blows again ;-)
     
  221. Agree, but digital is easier done in photoshop, film you need a darkroom...not a lot of people have a darkroom. Digital 2, film 1. :) hahaha. I just messing with ya'll. I just feel like we needed some laugh, the thread is way too long, it took me over an hour to read.
     
  222. Svenja, what is your point? I shoot towards the high sun a lot of times and because of that I want to carry the medium that serves me unfailingly.
     
  223. Brad, It's that time of year. You're hearing the Hallelujah Chorus and the audience is invited to sing. Pick a part - Tri-X, Ektar, Kodachrome or Velvia.
     
  224. All, this is a follow up to ICE levels, I tested Provia at with and without glass holder, at all ICE levels:

    (Shot with RZ67, you are looking at aprox 5 square milliliters of film out of 3900 total available in the frame)

    Quick observations:
    1- The effect on detail of ICE on Provia seems to be less than on Ektar
    2- Each ICE level affects detail progressively but minimally
    3- The texture of the glass holder becomes evident when sharpening is applied (ultra fine frequency so it should not show on prints. It can also be cleaned up with software easily.)

    Straight from the scan:
    http://shutterclick.smugmug.com/gallery/6616619_YJEwK#426487710_gwKGu-O-LB

    With fine sharpening:
    http://shutterclick.smugmug.com/gallery/6616619_YJEwK#426487722_6mhBs-O-LB


    Let me know what your observations are.
     
  225. Notice the glass texture observable when sharpening the scans (glass holder only) changes frequency with the different levels of ICE.
     
  226. At second examination, it does not appear to be the glass texture. It appears to be noise introduce with the GLASS+ICE combination. Still this may be too high frequency to have a direct impact on the print.
     
  227. @HC Lim

    You already said it in your answer. Given you have a stellar optics, shooting into the sun is perhaps easier with film...unless you do HDR (but hey, carrying around this tripod all the time).

    S.
     
  228. I also shoot towards the sun with lots of people moving about. How is that done in HDR?

    Come to think of it, nowadays people don't talk much about lens flare. It must be pretty obvious why.
     
  229. I verified indeed the fine noise introduced with ICE+Glass Holder+Sharpening does not show on the prints. Perceptually, it may even have a smoothing effect on very large prints.

    I tested on my Epson 3800 at: 500dpi (18x22 print), 360dpi (25x31 print), and 200dpi (45x55 print).
     
  230. Hi HCLIM,

    With film, it is easy. With HDR, you can merge manually the layers with different exposure. By now, the only way to get the job done. Automatic mergers usually produce ghosts. But perhaps at some time in the future this issue becomes resolved.

    Lens flare? Maybe this is no issue with the kit lenses? I'm not sure, though...

    Regards - S.
     
  231. You don't need the darkroom. I just have my negs processed by a local shop that I trust. I get a good whiff of
    the developer while there to keep me fixed... :)

    Then I just scan MF negs on an epson flatbed, and get 30" prints.

    For 35mm, hate to admit, but I secure those in the neg holders that came with my scanner, put them on a light
    table and shoot them as raw with a 40D or 1Ds and a 100mm F/2.8 macro (sharp sharp lens that one). Reverse and
    recover contrast in Lightroom using a preset on batch capture and edit right along with my digital shots. 24
    exposures takes only twenty minutes.

    Now whenever I encounter a scene that requires the DR of film. I bag the digital and shoot with my T70 or FTBn.
    In fact, I'll experiment with the digital first (pixels are free and at least I get the shot), and follow up
    with film. I use a lot less film that way, and have more keepers.
     
  232. Double Wow.

    How can anyone follow a thread this long?
     
  233. Jack:"Film has log color but it's still scanned by a linear scanner. There's no such thing as film. Everything is scanned & viewed on a computer eventually, using the exact same linear photodiode technology."
    Right, but I hope you're not trying to say then that the final product of film + scan gives you a linear recording of the original scene. Because it doesn't. Film (negative) compresses the tonal range of the original scene; the scan then makes a linear representation of that compressed tonal range (which it shouldn't have any trouble doing; that is, it should be able to recover the entire recorded range on the film itself with 14 bit A/D conversion saved to 16 bit files), and so you end up with a file with a compressed tonal range of the original scene. You then expand the tonal range carefully using whatever adjustments you want to bring about your final vision of the scene.
    The digital camera just gives you a linear representation of the original scene. OK it's not that simple b/c the RAW converter does a lot, and there's a lot more processing involved (including, these days, 'highlight tone priority', but the effects of that are debatable), but you get the picture.
    Important point: film + scanning does NOT give you a linear representation of the original scene, which a digital camera otherwise does.
    Rishi
     
  234. Kelly:"One can have a poster shot with a cellphone and folks will like it if you say its film; and hate it if yoiu say it was shot with a cellphone; "since a cellphone is never a camera; and never will be" according to many folks. Its actually an interesting exercise in human bias.
    Great point. It's called 'synthesizing happiness'. Check out the incredible TED talk given by Dan Gilbert entitled "Why are we happy?":
    Dan Gilbert asks: Why Are We Happy?
    Rishi
     
  235. Mauro,

    I don't really see the texture of the (presumably anti-newton) glass in the sharpened image -- could you point it out?

    Unless you're talking about the fine grain color noise across the entire image... which I assumed was just due to the sharpening.

    I have seen texture of AN glass on a LS-4000 scan... but it's quite big and not fine grained. This is AN glass from focal point.

    I have been able to subdue it, however, with an optical diffuser.

    I need to write a thread on it :)

    Rishi
     
  236. Umm, I don't get the whole log vs linear thing - it is dynamic range that matters, so lets say a dynamic range of 80 db (10,000:1) is a dynamic range of 80 dB = 13 plus stops = 10,000:1 = 4.0 and so on. Shapes of curves don't really matter, you can always convert between curve shapes. If it is log, apply an antilog function. If it is linear and you want log, just take the logarithm etc.
    Film may well have 80 dB dynamic range, but digital sensors will eventually get there if they aren't already there. Beyond that, someone needs to explain why digital giving a "linear" representation matters at all.
    And please explain what you mean by film having a non-linear response. It is linear - a look at any film datasheet will confirm this. For the new Ektar 100 for example, you'll see curves that show density vs log-exposure. Density is itself a logarithm, so the response is exactly linear. What am I missing here?
    If you mean the S-shaped curve with a heel and a toe - that is just how real linear systems behave: they have clipping points; that is not non-linearity. Between the heel and toe there is a linear region; this is how even electronic sensors work.
     
  237. Rishi, yes I am referring to the noise introduces with the Glass+ICE+Sharpen combination. It so happens this noise is ultra high frequency (very fine) and does not show on the prints.

    Here is with even more exaggerated sharpening:

    http://shutterclick.smugmug.com/gallery/6616619_YJEwK#427295802_gSvPJ-O-LB

    Compare the noise on frames 2 and 3 (Glass+ICE+Sharpen ) with 5 and 6 (No Glass+ICE+Sharpen ).
     
  238. Vijay, film gets more resistant to light the more exposed it gets. I an over simplified example, it is like walking into a wall,

    With film the closer you are to the wall the shorter the steps you take (yes, you can always linearize later).

    With digital (today) you take equally spaced steps and you hit the wall sooner (no, you can't remove the bump on your head by taking smaller steps later when you walk to bed).
     
  239. Rishi, to clarify, you are correct it is not glass texture it is fine noise. This actually benefits the prints (per sizes I listed above).
     
  240. And this is my (subjective) reading of the provia test chart resolution captured by the Coolscan 9000 (along the transversal higher resolving axle) :

    Glass + No ICE 3900

    Glass + ICE Normal 3750

    Glass + ICE Fine 3400

    No Glass + No ICE 3900

    No Glass + ICE Normal 3850

    No Glass + ICE Fine 3400
     
  241. dpi lines per picture height
    Glass + No ICE 3900 8598
    Glass + ICE Normal 3750 8268
    Glass + ICE Fine 3400 7496
    No Glass + No ICE 3900 8598
    No Glass + ICE N 3850 8488
    No Glass + ICE Fine 3400 7496
     
  242. Not sure what happened there:

    In DPI:

    Glass + No ICE: 3900.

    Glass + ICE Normal: 3750.

    Glass + ICE Fine: 3400.

    No Glass + No ICE: 3900.

    No Glass + ICE N: 3850.

    No Glass + ICE Fine: 3400.


    _.

    In lines per picture height:

    Glass + No ICE: 8598.

    Glass + ICE Normal: 8268.

    Glass + ICE Fine: 7496.

    No Glass + No ICE: 8598.

    No Glass + ICE N: 8488.

    No Glass + ICE Fine: 7496.
     
  243. Glass + ICE + Whiskey = 8432

    Glass + No ICE + Whiskey = 8200
     
  244. (No Glass + ICE + Whiskey = Mess)
     
  245. Very nice!
     
  246. Vijay, sure there's a region of linear response (sometimes where the slope is higher than 1) -- this is the most 'usable' region, hence you want to place the main subject exposure somewhere in that region. But on the upper end, it becomes harder & harder to expose the film due to the build up of metallic silver at sensitivity sites as well as the building negative charge at sensitivity specks (electron traps).

    Digital sensors don't do this, AFAIK, because I'm not all-knowing. No one is. Except for Feynman.

    But thanks for bringing it up... Because I've always wondered if it's right to call it a 'log' response. Also, this all makes sense for negative film, but positive film also works with silver halide emulsions, correct? The difference being that color dye does NOT form where sufficient exposure takes place. So if exposure is capped off in positive film, as it is in negative film, doesn't that mean that highlights would be saved in slide film, because it becomes harder and harder to expose, so there'll still be some chances of color dye FORMING in highlights?

    I don't get it... please someone explain if you know. Although, sorry Mauro, I should post a separate topic for this question :)

    Rishi
     
  247. Mauro -- weird! Why would glass introduce fine color noise?!
    Here's what I saw with AN glass between the LED source and the film in my LS-4000 scanner (w/ no optical diffuser, which the LS-9000 has, keep in mind):
    [​IMG]
    Link to full-size image
    It's not a fine texture at all. It's the etching in the glass. What you see is rather strange I have to say... but much easier to get rid of than what I see with my AN glass (AFAIK, impossible to get rid of)... your noise I'm pretty sure we could get rid of using Neat Image.
    And while we're on the subject: isn't Neat Image amazing?? I've tried Noise Ninja, but for film scans, have been very unhappy. What sets Neat Image apart is the fact that you can control how much of high frequency vs mid frequency vs low frequency noise reduction you want... which allows you to single out film grain from the rest of the image details.
    Rishi
     
  248. Neat image is indeed very nice.

    The glass surface you posted looks pretty bad/horrible. The Nikon glass holder does not have that problem.

    The noise that you see on the Glass+ICE example (shows also without the glass holder but not as much) actually has a smoothing effect on the print. I believe if it is an educated intended result from the ICE algorithm.

    Give it a try and print my test strips on your printer at 300, 360 or 600 or 720 dpi (or any dpi you use on medium format). PLEASE LET ME KNOW YOUR OBSERVATION FROM YOUR PRINTS. MY PRINTS DO NOT SHOW THAT NOISE AT ALL.

    Use the one with sharpening: "With fine sharpening: http://shutterclick.smugmug.com/gallery/6616619_YJEwK#426487722_6mhBs-O-LB"
     
  249. Rishi,

    Which side of the film do you have against the AN glass for those scans?

    Thanks, Helen
     
  250. Mauro:"The glass surface you posted looks pretty bad/horrible. The Nikon glass holder does not have that problem."
    Right but you're using a LS-9000 right? So if your holder had that surface, you wouldn't know b/c the diffuser eliminates the pattern. Or rather, the lack of a collimated light source illuminating the AN glass de-emphasizes the pattern.
    Indeed, when I place an optical diffuser in the path between the light source and the AN glass in my LS-4000 -- the pattern disappears! Results pending a thread on this topic :)
    Helen: the light source on the LS-4000 is on the top. The CCD is below the film. I have the emulsion side facing the CCD with no glass in between the film and the CCD. The base side (shiny side) of the film is facing up, with the AN glass placed above it such that the etched side is touching the base side, and the non-etched side of the AN glass is facing up toward the light source. Hence why I don't get any Newton rings (yet).
    Cheers,
    Rishi
     
  251. That is a good point, still that pattern seems large an intrusive though. Yet it could be the difference in the light type.

    If you want to, I can mail you the Provia frame I used for the ICE test so you can run thorugh your 4000 for comparison.
     
  252. I'm baffled by the controversies aired here as this is my first exposure to this chat group. I feel like I should get a
    Ph.D. in photographic process in order to feel confortable discussing some of these issues.

    As an architectural/interiors photographer using a Sinar F2 4x5 camera with a roll film back, I and many others in my
    professional still use primarily film because either the clients request it (refersing their earlier preference for digital) or
    because if it ends up being cheaper to use than a digital front end camera process - given the much greater PS time
    and cost (if you charge for it) from using a prof'l level digital camera and related equipment, process. A few more
    sophisticated architects can be quite fickle about the use of specific types of cameras (depending on the age or
    generation of their art director or the latest buzz about new cameras) but most don't even care which front end
    process the photographer uses since all they want is a CD at the end of the process and they couldn't care less how
    you acheived it as long as the client's perceived required level of quality is met and evident in the final printed image
    in a website, proposal or publication. I know major arch firms that remain 85% film based but even they don't discuss
    the issue (as we do) whether film or digital front end equipment and their respective results are "better". It, in my
    humble opinion, is very personal, situation specific, and hopefully based on the evidence we photographers have or
    don't have regarding what will give the client the best results. Bottom line: what is important is which process gives
    you the best level of quality for what the client says he or she needs and later applauds, and thereby also gives you
    a competitive advantage in your marketplace. From a purely business point of view, what good is bulking up and
    using an expensive digital camera if it doesn't ultimately give you a competitive advantage in the eyes of the
    customer?

    As a result of my limited rentals of digital cameras, my personal observation is that film is still superior to and less
    expensive vs. the digital front end process. As an aside, many at B&H Photo continue their argument that "why go
    digital when good professional film gives you, right out of the box, the roughly 26mp equivalent - which they add most
    digital cameras can't even approximate.... Same for dynamic range, color saturation, three dimensionality". Keep in
    mind that my market is to architects and construction companies, publications, designers who have little need
    for 'quick turnaround' vs. art directors who want instant feedbacks as in a fashion shoot or daily sports news
    newspaper environment. I might also add that the annual NPR NY Times roundtable broadcast "seminar" (by a host
    of acknowledged experts) on this topic, still leans in the direction of film in terms of its technical end-results achieved
    for a variety of end uses. The debate rages.

    Having said all that, I hope to make my first purchase of a new Nikon 26mp full-frame DSLR digital camera when it
    comes out in a year or so! - so I too can enter the grand experiment you folks are inmeshed in. Keep up the good
    discussions. Eventually, I'll figure out a fresh way of identifying what is "right" for my clients' needs as time
    progresses, as I start full scale technical experiments and evaluations. All the best. Tom Kieren
     
  253. Mauro: "Vijay, film gets more resistant to light the more exposed it gets. I an over simplified example, it is like walking into a wall,
    With film the closer you are to the wall the shorter the steps you take (yes, you can always linearize later)."

    No, yes - I actually understand what you were saying before, I was pointing out that film is actually linear, not log. What you are describing now is "soft clipping"; digital has "hard clipping", I suppose, but nothing prevents engineers from designing in soft clipping either. However, that does not increase the dynamic range. In fact soft clipping is obtained at the expense of the linear region, and can actually be considered detrimental in linear systems. I can't for a moment imagine the clipping characteristics to matter much (or at all) for the purpose of discussing dynamic range.
    Rishi: "Digital sensors don't do this, AFAIK, because I'm not all-knowing. No one is. Except for Feynman."
    Of course digital sensors do this, and so do all physical systems. As you wind a spring, it gets harder towards the end (soft clipping), and as you fill up a glass, the last drop causes it to overflow (hard clipping). With digital sensors, it is a bit of both, depending on the actual circuits, but generally if you consider charge in a reverse biased photodiode (or photogate) it will clip "soft". Think of how a flash charges - initially it makes a loud, low pitch sound, and as it gets more and more charged, it makes a higher pitch sound. This is actually the current charging the capacitors; initially it is high, and as time goes on, it becomes lesser and lesser; this kind of behavior is called "asymptotic". The charging current asymptotically approaches zero.
    All nice and fun, but doesn't affect dynamic range one bit. Dynamic range is a ratio: the max to min value of whatever. In our specific case, it is the response of a medium to light. This has little to do with how many bits you use to represent the values, or how the response varies (log curve, sine curve, who cares). The only thing that matters is the final max/min ratio. The only exception is if you can actually convert "soft clipping" to useful dynamic range - unfortunately this barely results in a third to half stop extra with film, and this is perhaps unavailable with digital. Nevertheless, it is a tiny fraction of the actual dynamic range.
    It would appear that for Ektar 100, Mauro found that this ratio is upwards of 8000:1; 2^13 is 8192, so this is a dynamic range of upwards of 13 stops. In dB terms, a ratio of 10,000:1 is (since log_10(10,000) = 4) 4 * 20 dB or 80 dB. Digital, with about 9 stops is around 512:1, which is around 55 dB.
    55 dB is quite low, and I am quite sure that the current generation of imagers being designed should have at least 80 dB or more (I've seen some papers that claim 115 dB) and we will start seeing these sensors come in fairly soon - 2-3 years, I think.
    And finally if you want to test dynamic range, you don't have to plot response versus exposure or whatever; just find two points, one for underexposure and one for overexposure at which detail either disappears into all black or all white. The ratio of these is the dynamic range. Simple.
     
  254. Vijay:"The only exception is if you can actually convert "soft clipping" to useful dynamic range - unfortunately this barely results in a third to half stop extra with film, and this is perhaps unavailable with digital. Nevertheless, it is a tiny fraction of the actual dynamic range."
    Reference??
    We had the same problem in the binary vs analog thread -- if you're gonna claim something that goes against observation, you gotta back it up with some explanation for the observation... in this case the observation being that there's clearly more stops recorded by negative film, yet here you are saying that they all respond linearly to light. That doesn't explain the observations. Furthermore, if layers of film start blackening when exposed, how does that NOT lead to a logarithmic response, or a more 'spread out' soft clipping (in your terminology)? If the film layers get blacker, it becomes harder for photons to penetrate (b/c they can be non-productively absorbed), which might explain the large latitude of exposure we see... yes?
    Thanks for the example with the capacitor, but please translate that to how it is relevant with a photocell. If you could explain or point me to a reference for how exactly the photon encounter dislodges electrons, etc., that'd be appreciated. Most Google searches lead to non-technical references, which I'm tired of reading.
    Finally, explain to me how, if negative, positive and digital sensors are all linear, why it's almost impossible to overexpose negatives ('logarithmic' response to light, as I've read, I'm not just making this up, so blame all the carriers of misinformation out there if I'm wrong), too easy to overexpose positives (exponential response to light), and digital somewhere in between...??
    Thanks,
    Rishi
     
  255. Nice, Rishi - attack before you research.
    First, a google search of "soft clipping" turns up 38,500 hits if you put the phrase in quotes, otherwise 870,000 hits. Go read some of them, including the wikipedia articles at the top. My terminology? You don't know what you are talking about at all, do you?
    Second, don't spread FUD, please. Linearity has nothing to do with dynamic range. The first is the shape of a transfer curve, the second is a ratio. This thread is about dynamic range, not about linearity. Look at the thread title.
    Third, a photogate is a capacitor. So is a CCD. Go look up how they work.
    Fourth, you should know better, after interminable arguments in the other thread, that silver halides mostly don't get converted to silver until after developing. How is it that silver that doesn't exist block light?
    Fifth, the response of film (black and white, slide, color) is all linear (or closely linear). If you have any doubts about this, pull up film data sheets and look at the characteristic curves. Then go look up exponential or log curves.
    Sixth, you asked for references. Here's one. The definition of dynamic range. Look it up. http://en.wikipedia.org/wiki/Dynamic_range
    Seventh, don't go off on a tirade of "how" before the "what".
    Lastly Mauro, I really don't want to hijack your thread; but this is a discussion about dynamic range (at least what I am doing). So if you want me to stop, I shall; and if you are interested in letting me put forward the reality, I can do that as well.
     
  256. References:
    Clipping, hard and soft: Clipping (signal processing)
    Sensor technology: Active Pixel Sensor
    Pixels: Pixel Circuit - talks about the transistor M_sf - "an amplifier which allows the pixel voltage to be observed without removing the accumulated charge". A device that stores charge is called a capacitor. A zero biased photodiode is essentially a capacitor whose charge is proportional to the light striking the junction.
     
  257. Here is another example I shot today of some dead roses on the ground with TMX (35mm) where the dynamic range of the scene even exceeded the film! (This was developed with Xtol 60/40. If I had intended to expand the highlights and the shadows pass 15 stops I would have used TMAX developer but I consciously opted to go for the midtones -Xtol- instead).

    This picture would have been literally impossible with my 40D.

    http://shutterclick.smugmug.com/gallery/4811519_F9MBv#427771770_c45Yh-X3-LB
     
  258. Vijay, "What you are describing now is "soft clipping"; digital has "hard clipping", I suppose, but nothing prevents engineers from designing in soft clipping either. However, that does not increase the dynamic range. In fact soft clipping is obtained at the expense of the linear region, and can actually be considered detrimental in linear systems."

    No offense intended at all, but you are confused. You are thinking like "Photoshop" when you think that changing curves does not affect dynamic range. We are talking about the "effect" a non-linear capture has on the number of stops recorded.

    In my "walking into the wall example" if you take steps half the distance to reach the wall then how many "steps" can you take? Obviously infinite.

    In the same example if you take equally spaced steps say 9 (a la 40D) and you hit the wall, yes you can plot the number of steps against the distance using a logarithmic scale if you want to, but pass the last step and before you hit the wall there is nothing.


    And you are probably right that a digital system could be created to have a non linear capture to expand dynamic range. But here we are comparing current existing technologies.
     
  259. [Quotes in italics by Kelly-not picking on this specifically but it seemed to have much of the neg content that some of the other posts have I felt like addressing] hmm seems italics dont work

    Quote PHOTO.NET has several main threads that have been rehashed over the last decade many thousands of times.

    No doubt they have been rehashed Kelly and if you are spending a great deal of time on the forums then you would be sick of it. But as strange as it may seem, digital is new to me (I just shoot film, I have an old 2Mp digital camera that is the size of a house (well maybe a caravan ;) , but its horrible so I never use it, the one on my phone is better). Considering the technology of digital is always advancing, old discussions aren’t as relevant as new discussions. On occasion in the past and again now I am considering buying into digital. But before I consider spending the huge amount of money buying into it and all that goes with it, computers, hard drive storage, scanners, printer, camera, lens, surge protectors and software ect. I would like to know some of these obvious things that are discussed. So I can weigh up whether it’s right for me given the latest improvements and whether I really want to spend the huge money.

    I can’t say I enjoy these threads that get taken over with all the techno babble at all, but I don’t mind as much so long as someone eventually comes to a straight forward easy to understand conclusion, that I can relate to in everyday use and the finished print. I don’t have the benefit to hire equipment in my location.

    Quote(1) Folks buy a gooberflex digital and ask how BIG can I enlarge the image.

    As lame as this type question is, it does make me wonder sometimes as well. When I go to the camera store they have some picture of a scantly clad girl enlarged to 6x4 foot poster and say ,see this is digital. I don’t get much of an explanation out of them how they achieved it (and they usually come from Canon or Nikon reps anyway) so I eventually brush it aside as sales pitch. Another shop told me I am wasting my time with 8x10 large format negs as digital is just as good as that too…I find that impossible to believe and think it is just them wanting to sell me digital equipment. I do get lost and frustrated with biased advice.

    Quote(2) Folks learn that slides have less dynamic range than a negative.

    Well I guess if they don’t know they find out by people saying so, even tho some can understand just by looking at the prints others may appreciate the explanation.

    Quote(3) Folks learn that many digital sensors have a less dynamic range than a negative.

    That seems to be what this thread is about and just a few posts have been straight forward enough to understand why. I find it useful.

    Quote(4)Folks learn that a raw digital file can have more range than a jpeg

    seems obvious

    Quote(5) Folks who post giant images on the web and learn that they got borrowed and stolen

    yeah well duh, but basic good advice given inst any skin of your nose

    Quote(6)Folks who learn that having more ram makes their Photoshop computer work better

    may seem obvious, but considering I don’t have photoshop its still a reasonable question if i would like to know just how much ram others find that works with the latest release computer, PS and large (files) negs scanned .

    Quote(7)Folks who learn that a high contrast lens test shot with camera bolted to a granite block give great resolution numbers to brag about.

    Quote(8) Folks who learn that disc drive can carsh and "learn" that they should havce had a duplicate file

    its still useful information; it seems to relate to just how huge a cost I will need to spend on raid storage to hold medium and large format files/scans

    Quote(9) ZILLIONS OF FOLKs who are are lost souls; who crave an exact megapixel equalvalent for 35mm film; often posting results. The same folks probably want and can prove the number of MP3's per 13 year old girl; shoes per woman, lenses per photo.neter, or beers per football game.

    Like I say I don’t much like all the techno babble and fine nit picking but I do like to know how 35mm, 120, 4x5 and 8x10 film compare in terms of digital. What cameras or formats can be replaced if I go digital and is it cost effective to me ect.

    Quote(10) Folks who buy a lens with a dinky scratch and ask others if it a matters; but do not have the ability to shoot few frames as their own test

    yeah well you get that, seems obvious to me as well but who knows what goes on in the minds of others. Maybe they just want to know if it didn’t have any marks would their images look better

    Quote(11) Folks who crave ratings on posted photos; but complain when they get negative feedback.

    Doesn’t concern me, I guess its bothers people whom spend more time on the forums.

    Quote(12) More threads about film versus digital; posted like its a new subject; when its several decades old now.

    It is new to me, old posts a near irrelevant to me as the comparisons are to older digital cameras. I cant say I appreciate the bickering that goes on , I just want simple straight forward answers on the differences, advantages and disadvantages

    Quote(13) Folks who discover that scanning takes time; when its a 2 decade old known fact.

    seems obvious

    Quote(14) Folks who think that cleaning off fungus from a lens "magically" removes the potholes that are etched in advanced cases of fungus.

    some people just don’t know, it doesn’t help to dismiss

    Quote(15) Folks who worry about Macs versus PC's instead of worrying about results delivered to clients.

    This is actually something I was interested in lately so i read up on it. I am due for a new computer and if I have to spend up to $4000 ish then I would be stupid not to gain information and ask a simple question or read posts on the latest equipment available.



     
  260. Vijay also, "I actually understand what you were saying before, I was pointing out that film is actually linear, not log"


    Please read Ektar's publication. Look at the gain chart... it looks linear right? well now look at the scale it says log.

    http://www.kodak.com/global/en/professional/support/techPubs/e4046/e4046.pdf
     
  261. No offense intended at all, but you are confused
    Oh Mauro.... now you've gone and done it....
     
  262. Mauro - the chart you pointed out to me is that of density vs log-exposure. Density is itself a logarithm. So it
    is a log-log chart, meaning that the shape of the curve is exactly as it appears - i.e., linear. Take the inverse
    logarithm on both axes - the curve shape won't change one bit. No offense taken, Mauro - please don't take any
    when I point out that you are in error here.

    As for the walking into a wall example, it doesn't apply. You have maybe 13 stops of completely linear transfer
    function, and maybe a quarter stop on the top and bottom parts of the S-curve. This half stop is "curved" (just
    the knee parts of the curve), akin to your walking into a wall thing. If you were to "straighten" it out thereby
    extending the linear range, you could get back that half stop or so. Beyond that, it's all lost, because the film
    response has become "flat" - change in exposure results in no change in density.

    And finally, dynamic range is a simple ratio. Nothing to do with curve shape at all. If you can linearize and
    extract an extra half stop, then fine - the dynamic range is now 13.5 stops; that is the final max/min ratio.
    Thats all.
     
  263. Thanks for the references Vijay. Re-read my post -- I was asking for explanations of the observations & references for your broad claim that clipping only saves 1/3 to 1/2 a stop for film. Do you expect us to just take your word for it?

    And why does it always come down to a game of semantics with you? Previously I had to spend hours explaining to you why perceived resolution and tonal range of the image forming element *were* related, while you simply touted that tonality & resolution are unrelated -- which they are, in the broader sense, but not in the sense we were discussing.

    Same thing here. Yes I know the linearity, or slope, or whatever of the actual curve *does not have to be related* to the overall range of intensities recorded; HOWEVER, if the response of your light-responding element, which may otherwise be linear, encounters a phenomenon such as darkening of the layers of film to an extent where subsequent penetration of photons become more and more difficult, THEN you would increase the latitude of exposure because in any given spot that, were response entirely linear, would have been entirely exposed so as to result in a fully developed grain but that, now, experienced less productive photon hits and so therefore will stop just short of being fully reduced to a grain of ENTIRELY metallic silver.

    In which case, exposure latitude of the film would change due to its nonlinear response.

    Happy?

    And no, I didn't know the TERM 'soft clipping', though I know very well what it is. I'm not afraid to admit it. What am I an EE? Chemistry & biology, dude, but at least I take a stab at subjects out of my field, so don't poke fun because I don't know your terminology. I'm here to learn & maybe, just maybe, in the process contribute. Your arrogance is truly astounding. When's the last time you picked up a Chem Rev. article like the one I posted and asked you to read in the last thread? Mees? Baines?

    Finally, you dog me for misunderstanding 'dynamic range'. WTF? Where have I mentioned 'dynamic range' in my post?! I was speaking of 'exposure latitude'. Read before you attack.

    And while we're on specifics -- fine, it's not a 'logarithmic' response... that's just the term people seem to be throwing around & I jumped on the bandwagon. It's a sigmoidal shaped curve with a toe & a shoulder... i.e. the Hurter Driffield curve.

    The more interesting question to answer is this: is the greater exposure latitude of negative film due to the shoulder of the H&D curve? And is this shoulder replicated on digital sensors?

    Finally, Vijay, re: your comment about how silver that doesn't exist blocking light -- don't be a smart alec. How can it? It *can*; now the extent to which it affects exposure, I couldn't tell you (but it's an interesting question I'd like to find out the answer to). We learned from all our research in the last thread that you could have hundreds to thousands of sensitivity specks per grain. Each of these specks could, upon exposure, yield up to 50 or more (I don't think we ever determined the largest number) reduced silver metal atoms clumped up around the sulfide electron trap. Multiply this by all the sensitivity specks per grain. Multiply this by all the billions of grains in an emulsion. These clumps are BLACK because they consist of metallic silver. So yes they *could* theoretically block light from reaching an underlying sensitivity speck. This MIGHT explain that shoulder that gives the extended exposure latitude of film... I DON'T KNOW, but don't discount it with some smart comment and no thought.

    Rishi
     
  264. And Mauro, what you are talking about (non-linear capture that extends dynamic range) is actually called companding. See the relevant wikipedia article on companding.
    Companding needs a variable gain amplifier in the path - which amplifies small signals a lot, and large signals less. Film don't have no variable amp. If it did, you couldn't have nice geometric apertures and shutter speeds (geometric as in geometric progressions).
     
  265. What I meant to say at the beginning of my last post was: Re-read my post -- I was asking for explanations... NOT ATTACKING YOU.

    Forgot to add that last bit in CAPS.
     
  266. "And Mauro, what you are talking about (non-linear capture that extends dynamic range) is actually called companding. See the relevant wikipedia article on companding."
    YES!! That's what I was talking about! Sorry I didn't know the technical term back then. Now do you retract your comment that linearity (or non-linearity) of capture is absolutely unrelated to the range of intensities recordable??
    Sure a digital sensor would need a variable gain amplifier in the path that amplifies larger signals less -- duh, that was already stated above that that's what'd be needed to extend the exposure latitude of digital sensors. Canon's trying their own implementation in 'highlight tone priority', but I believe it only tries to simulate the effects of a variable gain amplifier.
    The other way to achieve it, in film, would be to employ exactly the technique that I mentioned before -- clumps of silver increasing local density at sensitivity specks, blocking photon penetration to sites directly below. Or negative charge buildup at sulfide electron traps, though, in my opinion, that effect would be small.
    Rishi
     
  267. Vijay, "Companding needs a variable gain amplifier in the path - which amplifies small signals a lot, and large signals less. Film don't have no variable amp. If it did, you couldn't have nice geometric apertures and shutter speeds (geometric as in geometric progressions)"

    Vijay, you are getting confused to a more simple level now. Yes you can have geometric apertures (each lets twice/half the amount of light through with every single stop) - THAT DOESN'T MEAN THAT WHEN THEY IMPACT THE FILM WITH THE SAME PROGRESSION. i.e., the negative does not get twice as dense when when you go from f11 to f8. It gets denser but not twice.

    Please read the publication I gave you the link for.
     
  268. Canon's trying their own implementation in 'highlight tone priority', but I believe it only tries to simulate the effects of a variable gain amplifier.
    Rishi... last time I saw HTP beaten to death on a forum (LL), it was agreed that HTP was nothing more than underexposure by 1 stop, then boosting the shadow data. You can implement it manually by dialling in -1 EC and then applying a bit of "fill light" in LR/ACR.
    I am going to tentatively have a crack at this non-linear shoulder thing, but until I get home and check the references, I will leave it as a discussion point. Hopefully Rishi you have covered it in your reading and can confirm if my memory is serving me correctly. I seem to remember it stated somewhere that as more silver ions are converted to silver metal at sensitivity centres, the less sensitive those centres become. When I get home I'll try and find out if this is a correct recollection. Cheers.
     
  269. Top of page 5.

    http://www.kodak.com/global/en/professional/support/techPubs/e4046/e4046.pdf
     
  270. Bernie, you are right, Canon's highlight priority is nothing more (so far) than underexposure with curves.

    There is no debate there. To simulate film you would need a log gain AT THE PIXEL LEVEL.
     
  271. OK Rishi - your tone was quite, umm shall we say - hostile. If that wasn't your intention, then I misread it and
    apologize for that. I wasn't being arrogant or trying to make fun of you. I was pointing out several realities
    that don't quite jive with the negative film/exposure latitude theory being floated on this thread.

    I wasn't specifically pointing out at you that you didn't know about dynamic range. It was a general comment -
    dynamic range is a ratio.

    Now you deflect to exposure latitude. Let me take a stab at defining that. Let's say a scene has a dynamic range
    of 6 stops. Let's say your film has a dynamic range of 10 stops. The difference is the exposure latitude. You can
    err by 4 stops (2 stops in each direction), and will be able to recover a perfectly fine image. This is because
    you can place the 6 stops (zones) of the scene in any part of a 10 stop available range.

    On the other hand, if the scene has a 15 stop range, and your system has 13 stops of dynamic range, then your
    latitude is negative - no matter what you do, you will lose detail.

    Latitude is not a characteristic of a system; it is just what happens when a signal (light) has less dynamic
    range than the channel (film in this case). Talking about latitude as if it were a system characteristic (i.e., a
    property of film) is an exercise in meaninglessness.

    Now for getting back half stop from the curved portions of the Hurter Driffield curve - look at the curves for,
    say Ektar 100. How much is the length of the toe before the density hits Dmin? Barely a third of a stop. This is
    all you could theoretically straighten out. The rest is already linear - the linear part of the sensitometric
    curve. You need a reference for something as straightforward as this?

    And lastly, if the accumulated silver blocks light to a significant degree, then it would make the response of
    film non-linear. As we know from the Hurter Driffield curve you pointed out, that is not the case. The already
    formed latent image can't significantly affect further formation of the latent image because that would make the
    response non-linear, which it is not.
     
  272. Mauro: "Vijay, you are getting confused to a more simple level now. Yes you can have geometric apertures (each lets twice/half the amount of light through with every single stop) - THAT DOESN'T MEAN THAT WHEN THEY IMPACT THE FILM WITH THE SAME PROGRESSION. i.e., the negative does not get twice as dense when when you go from f11 to f8. It gets denser but not twice. "
    You miss the point that if the negative gets (say) 1.5 times as dense going from f/2.8 to f/2, it also gets 1.5 times as dense going from f/11 to f/8.
    Companding would require that the negative gets (say) 1.8 times as dense going from f/11 to f/8 but 1.2 times going from f/2.8 to f/2 (more light, less gain).
    Sorry Mauro, the former is not a variable gain amp - it's just an "amp" with gain of 1.5. The latter is; but you know as well as I do that film does not behave like that.
     
  273. Bernie:"Rishi... last time I saw HTP beaten to death on a forum (LL), it was agreed that HTP was nothing more than underexposure by 1 stop, then boosting the shadow data. You can implement it manually by dialling in -1 EC and then applying a bit of "fill light" in LR/ACR."
    Sweet. That was my first inclination when I read about HTP. Because they said it only works at ISO 200 or above. Which made me think "Hmm... that probably means that if you set it at ISO 200, it actually sets the sensitivity to ISO 100 then applies a non-linear gain".
    But if each row of pixels (or is it columns? I forget) has its own amplifier, perhaps some non-linear gain can be done to the amplification on a row-by-row basis by somewhere implementing in the circuitry a way to determine gain based on total voltage across the row? I don't even know if that'd be worth it because it might help landscape photogs but certainly wouldn't help trying to photograph a bride's white dress (random highlights) filling up a frame with shadows everywhere else.
    Bernie:"I seem to remember it stated somewhere that as more silver ions are converted to silver metal at sensitivity centres, the less sensitive those centres become."
    Yup, good call. It's in Baines, "The Science of Photography". I think that's part of the reason behind why a clump at sensitivity speck post-exposure but pre-development can only get so big... additionally, they can't get too big because it's energetically very unfavorable to disrupt the crystal lattice. Another reason why clumps don't usually form in the center of grains but rather on the surfaces where imperfections already exist.
    Vijay:"OK Rishi - your tone was quite, umm shall we say - hostile."
    Yeah, ok, with all my 'thanks' and 'that'd be appreciated', clearly, so easy to see how I was being 'hostile'. You, on the other hand, with your: "You don't know what you are talking about at all, do you?"... that, yeah now that's what we call hospitable, huh?
    If you're talking about the part where I was asking you to explain observations when your own comments/theory go against observation, I stand firm on that. In other words, instead of being so negative & shooting everyone down (which, sure, is necessary from time to time), offer a theory to EXPLAIN the observation then after you've knocked down a previous attempt to explain the observation. I still don't see ONE explanation from your end that explains the exposure latitude of negative film. Or the lack thereof of slide film. I understand shooting down theories (bad ones anyway) is part & parcel of good science, but, damn dude, try being on the other side for once. Contribute a theory.
    Rishi
     
  274. Bernie: "I seem to remember it stated somewhere that as more silver ions are converted to silver metal at sensitivity centres, the less sensitive those centres become. When I get home I'll try and find out if this is a correct recollection. Cheers."
    Sure that is a correct recollection. That has to happen, no doubt, otherwise film would have an infinite dynamic range. The point I'm making is that this is linear. Meaning that if you go from 10 ions to 20 ions, the effect is the same as if you were going from 100 ions to 200 ions (a decrease of sensitivity of 2x in either case). Nonlinearity is different - it implies that going from 10 ions to 20 ions will have a larger effect than going from 100 to 200.
    If the latter were true, you wouldn't have the S-shaped sensitometric curve.
     
  275. Vijay:"The already formed latent image can't significantly affect further formation of the latent image because that would make the response non-linear, which it is not."
    Maybe after a certain density is reached, it can. Which would explain why the shoulder exists at higher log E. Maybe the clump of silver metal atoms has to be >20 (arbitrary) to start blocking visible light in any meaningful way to layers beneath. Just throwing it out there. Contributing an idea, y'know?
    Vijay:"Latitude is not a characteristic of a system; it is just what happens when a signal (light) has less dynamic range than the channel (film in this case). Talking about latitude as if it were a system characteristic (i.e., a property of film) is an exercise in meaninglessness."
    I think you mean "when a signal (light) has more dynamic range than the channel (film in this case)."
    Furthermore, here we go with semantics again. Yes I understand that latitude in the sense of relation between the medium and what it's trying to record is what's important. What is the meaning of your comment? Latitude is still relevant to this discussion because apparently the latitude for the 40D is lower than that of Ektar 100 (or, less negative). Jeez man.
    Finally, why are you speaking of the 'toe' and 'Dmin' for Ektar? We're interested in the 'shoulder' and 'Dmax', none of which even exist on the Ektar data sheet, but do exist in typical H&D curves. I guess Kodak just didn't go that far out. What we should be looking at is the log E range over which the 'shoulder' displays non-linear behavior. That, tied in with the linear region, would by its very definition simulate a variable gain amp. Certainly not an ideal one, but perhaps a crude one that would explain the exposure latitude that we see for negative film.
    Rishi
     
  276. Vijay, "You miss the point that if the negative gets (say) 1.5 times as dense going from f/2.8 to f/2, it also gets 1.5 times as dense going from f/11 to f/8. "

    No no no.... I know you must understand it but it's just late.

    Please go to the link I sent you, pull a few points from the graph, and plot them in Excel on a linear axis. Then please, let me know what you conclude.
     
  277. Vijay:"as if you were going from 100 ions to 200 ions (a decrease of sensitivity of 2x in either case)"
    What are you talking about??
     
  278. Going to bed now...

    Vijay, I thank you and enjoy the debate, I hope I don't sound offensive.
     
  279. Rishi: "You don't know what you are talking about at all, do you?"... that, yeah now that's what we call hospitable, huh?
    Sorry. I got a little worked up. It's not like this is the first time I've faced hostility for what I said. Nevertheless, unconditional apologies.
    About theories - well, do you observe a significant extension of dynamic range with software processing of film? Nope - maybe a full stop more. How is this observation in conflict to what I'm saying? OK - maybe you could extract a stop rather than the half stop I said. The point was that compared to the 13 stop linear range, this is not that significant.
    Contribute a theory? Last I remember, I explained the whole process end to end - with gray grains and all; something that covered all observations - with support from information theory and whatnot. Please. Anyway, lets go to the other thread if you want to hash this out.
    As for this thread, I've already defined exposure latitude in precise terms, and I'll gladly explain what I know. Sometimes I may refrain from an immediate explanation because it may not be necessary for the discussion at hand. If you assume based on that that I'm ignorant, so be it. If I don't know something, I'll tell you I'm speculating, or I'm guessing. So what do you want to know?
    Why slide film has lesser dynamic range? I don't know. I speculate or I guess it has something to do with the reversal process. I wonder what happens when you cross process E-6 in C-41 chemistry - maybe the resulting negative has more dynamic range?
     
  280. The point I'm making is that this is linear. Meaning that if you go from 10 ions to 20 ions, the effect is the same as if you were going from 100 ions to 200 ions (a decrease of sensitivity of 2x in either case).
    But do we actually know this?
    Even if we assume this is true, the grain with 200 atoms (I assume you mean atoms, not ions) will be far less sensitive than the grain with 20 atoms. Let's do a simple (arggh, I hate to say it...) thought experiment and say that after time t one grain has 200 atoms and the other has 20 atoms. So after time 2t, lets say the more exposed grain has 250 atoms, and the other one has 35 atoms. ie. the more exposed grain reduces a lower ratio of silver atoms than the less exposed grain. Now we develop. Remember, development will begin sooner in the grain that has more silver at the sensitivity speck. But how much sooner comparetively would this occur, than say we developed after time t? I am speculating here, but lets assume that development initiation is directly linearly related to number of silver at a sensitivity site. So in the case of the less exposed grain, after exposure time 2t its development would begin in nearly half the time that it would after exposure time of t. But in the case of the more exposed grain, after exposure time 2t, its development time would begin in only 80% (or something like that if I could be bothered to work it out) of the time it would in after exposure time of t. So: double exposure time, doesn't lead to an across the board (tonal range) doubling of print tone.
    Did any of that make sense?
     
  281. Rishi: "Maybe after a certain density is reached, it can. Which would explain why the shoulder exists at higher log E. Maybe the clump of silver metal atoms has to be >20 (arbitrary) to start blocking visible light in any meaningful way to layers beneath. Just throwing it out there. Contributing an idea, y'know? "
    Sure - that may well be the explanation for the shoulder - although it is more likely that once an atom is already ionized, you can't do much more with it. A new photon has to find another atom to ionize. As more atoms get ionized, it gets more difficult for photons to find un-ionized atoms, so that could as well explain the shoulder. It is definitely some kind of saturation process.
    I think you mean "when a signal (light) has more dynamic range than the channel (film in this case)."
    No, I meant what I said. If the signal has more dynamic range, there is no latitude at all; the signal will clip. Latitude is the equivalent (in electronics) of headroom. If the signal is already larger than what the system can handle, you have no headroom. Here is the wikipedia article on latitude. http://en.wikipedia.org/wiki/Exposure_latitude. Even they clearly state (about latitude), "It is not to be confused with dynamic range, which is, in photographic context, the range of light intensities a medium can capture."
    And this is not about semantics at all. It is about accurately defining what we are talking about. Dynamic range is a ratio. Linearity is a curve shape. Latitude is an error margin. They may all be related, but they are quite different things.
    Toe and Dmin and shoulder and Dmax are symmetric phenomena. You can linearize both and extract detail at both clipping ends. So, whatever applies to one in general applies to the other. Since the Ektar curves don't show a shoulder (wonder why, Kodak?) I just used the toe curve to illustrate. That doesn't mean that a shoulder doesn't exist or that it behaves different from the toe.
    That, tied in with the linear region, would by its very definition simulate a variable gain amp.
    Of course there is a variable gain amp there - but it is the nonlinear part, not the linear one. That's exactly why I've been saying that you would extend dynamic range if you linearized that region. But from all the curves, what you see is a huge linear region and a small non-linear region, that if recovered would give you a half stop (or maybe a full stop). Compared to 13 stops of linear region, this is not the significant factor in dynamic range comparison - such as 13 stops for film, 8 for digital. If you linearized, you'd get what 13.5 for film, and 8 for digital. Significant difference? Not from where I stand.
    Then I pointed out that digital relies on a charge accumulation device as an optical sensor - this will also have soft clipping - or a curved shoulder, so you could probably extract the same half stop or so from digital, which means 13.5 stops for film and 8.5 stops for digital.
    Now to tie it all together - if you're going to compare a 13 stop range to an 8 stop range, both in the linear portion, then the theory that explains the difference can't rely on non-linearity (log response etc). What is the actual theory? In my mind, none - dynamic range is a characteristic of the system design itself; the 8 stop range for digital is a limitation of current design, and soon you'll see 13-14 stop ranges if not more.
    As for film - the whole thing - silver halide, development time etc., forms a system with some dynamic range. Why it can't be more is something that I have no idea about, but sure, if you made a blacker black (denser Dmax) and decreased base fog, why, you'd have more dynamic range. How to do it? No idea.
     
  282. Bernie:"but lets assume that development initiation is directly linearly related to number of silver at a sensitivity site."
    You mean 'inversely linearly related', yes?
    Bernie:"Did any of that make sense?"
    Yes, actually, it did. Assuming that you meant that the linear relationship was, in fact, inverse. That is, the more 'developed' a speck is, the lower the initiation time. Which I'm sure you did. I can read in between the lines :)
    However, your premise that "development initiation [delay] is directly [inversely] linearly related to number of silver at a sensitivity site" is a big IF. Chemically, I would have to say that the reduction potential at a sensitivity site shouldn't change that much based on 10 vs 50 vs 100 reduced silver metal atoms making up a 'clump'. What would change, however, would be the number of 'paths' along which electrons could travel to reduce other AgBr ions in the crystal. The more silver metals atoms you have in the clump to begin with, the more paths the electrons from the reducing agent (developer) could travel upon to reduce more AgBr ions within the crystal. Therefore, the faster and more extensive would be the resulting silver speck growth upon development.
    Now, if there's a non-linear relationship between the number of silver atoms (metal) at a clump and the number of paths available for electrons to travel along into the interior of the silver halide crystal (directly proportional to 'access' to internal AgBr ions for reduction), which itself (the # of paths) is proportional to the surface area of the clump (of reduced metal silver atoms) exposed to non-reduced silver ions within the AgBr crystal, then this would explain the non-linear response of film. If that's true to begin with :)
    Whoa, my head hurts. Again. But that may be a lead. I need to think about it. Or someone else does. Damn we should form a think tank.
    Rishi
     
  283. Bernie: But do we actually know this?
    Well, we know that the response of film is linear - so I'm saying that it is axiomatic that the response of a single silver halide crystal is also linear, because I can't imagine how non-linear elements could, when put together in an emulsion become collectively linear. Beyond that, I make no statement. That whole 10/20 and 100/200 ion thing was just a contrived example.
     
  284. Vijay:"although it is more likely that once an atom is already ionized, you can't do much more with it. A new photon has to find another atom to ionize."
    Oh, boy, Vijay. Ionize? Are you kidding me? There's no ionization going on. There's the reverse of ionization. Knocking an electron off a bromine ion and making it a bromine atom. What we care about is how many silver ions are 'reduced' to silver metal by accepting said electron. Technically, it's called 'photolysis'.
    Vijay:"As for film - the whole thing - silver halide, development time etc., forms a system with some dynamic range. Why it can't be more is something that I have no idea about, but sure, if you made a blacker black (denser Dmax) and decreased base fog, why, you'd have more dynamic range."
    Huh?? No one here cares about the dynamic range of the film. We care about the range of intensities discernible that were recorded from the original scene. The dynamic range of the film itself we can stretch out to whatever the heck we like in a 16 bit file in Photoshop. Or, heck, a 32 bit file.
    Vijay:"No, I meant what I said. If the signal has more dynamic range, there is no latitude at all; the signal will clip."
    Right, but, in this thread, we're only interested in scenes that have a range of luminosities greater than that recordable by the medium. So, yes, the signal will clip. Either in the highlights, or the shadows. So, in all cases here, we're talking about negative exposure latitude, if I understand the terminology correctly. The question is: which system has a lower magnitude of negative exposure latitude (i.e. closer to '0')? I think the answer is negative film.
    Rishi
     
  285. Mauro: "No no no.... I know you must understand it but it's just late.
    Please go to the link I sent you, pull a few points from the graph, and plot them in Excel on a linear axis. Then please, let me know what you conclude."

    Well, I don't use M$ stuff, so I'll dig up some graphing tool for this if I can, but I did a quick plot on paper. I expanded the X axis by taking 10 to the power of the numbers; and did the same for the Y axis. What I have is the X axis reads - 0.001, 0.01, 0.1, 1.0, 10 and 100. The Y axis reads: 1, 10, 100, 1000, 10,000. This corresponds to densities of 0, 1, 2, 3 and 4. I draw this, and it is still a straight line.
    The point is, density is a logarithm itself in the plot, so I expect to see a straight line, and I do. What am I missing?
    Maybe you're not expanding the Y axis? If so, then you will see a logarithmic response, but that is because the Y axis is already a logarithm. Don't believe me - tell me, is the ratio of the blackest black to the whitest white 4:1? Because a density of 4 is the blackest black, and this is certainly not 4 times less bright than the whitest white - it is in fact 10,000 times less bright.
    Based on the plot, what I see is if I increase the exposure 10 times, my density increases roughly 8 times or so (rough paper plot); same slope throughout. That's linear, since y/x = a constant.
    Maybe I'm missing something else - will wait for you to explain.
     
  286. Rishi: "Oh, boy, Vijay. Ionize? Are you kidding me? There's no ionization going on. There's the reverse of ionization. Knocking an electron off a bromine ion and making it a bromine atom. What we care about is how many silver ions are 'reduced' to silver metal by accepting said electron. Technically, it's called 'photolysis'. "
    I stand corrected. You are right of course. The point remains though; once you have silver atoms and bromine escapes, newer photons have to find bromine ions; which are getting scarcer as exposure progresses. This was actually in support of your point about non-linearity. I was merely stating that this process could very well be responsible for the eventual shoulder.
    Huh?? No one here cares about the dynamic range of the film. We care about the range of intensities discernible that were recorded from the original scene. The dynamic range of the film itself we can stretch out to whatever the heck we like in a 16 bit file in Photoshop. Or, heck, a 32 bit file.
    My turn.
    Huh? What? The "range of intensities discernible that were recorded"? This is the very definition of dynamic range. A range is a max and a min. Dynamic range expresses this as a ratio. It is called Dynamic range for a reason.
    Right, but, in this thread, we're only interested in scenes that have a range of luminosities greater than that recordable by the medium. So, yes, the signal will clip. Either in the highlights, or the shadows.
    You brought up latitude. As far as I have been following this thread, I have always spoken about dynamic range. The title of the thread is also "Dynamic Range".
    The question is: which system has a lower magnitude of negative exposure latitude (i.e. closer to '0')? I think the answer is negative film.
    The question is which medium has greater dynamic range (you may frame this as "which medium gives you the greatest margin for error?", but that just implies a greater dynamic range. No semantics - this is just about correct definitions.) And you are right about the answer - it is unquestionably negative film.
     
  287. Come on you two, let's not go crazy like in that last thread. I think you are actually in agreement here. What Vijay says about film being linear is right (I think), except for the tiny bits of the graph which correspond to the toe and shoulder. So he wasn't being exactly precise in his language, but he is right to say (I think) that to linearize the non-linear sections wouldn't add much to the DR of the film. Surely we all agree on these points? And we all agree that negative film has more DR than a 40D. So I don't even know exactly what you two are arguing with each other about. Now, make up and be friends, before Vijay starts playing reductio-ad-absurdum :)
     
  288. Now, make up and be friends, before Vijay starts playing reductio-ad-absurdum :)

    ROTFLMAO :))))
     
  289. Sorry, I thought perhaps you were being tricky & talking about the dynamic range of the dyes on the film. Is there a different
    term for that?

    As for bromine ions becoming scarce with increased exposure? Not a chance. There are orders of magnitude more bromide
    ions than even the highest number of reduced silver metal atoms possible for a fully exposed latent site; indeed, if it weren't so, then
    a fully exposed negative would look black prior to development, & we know that's not the case (reductio ad absurdum
    anyone? :)

    Bernie, sure Vijay & I are in agreement because Mauro did a convincing job here! But, we aren't in agreement as to the
    HOW (actually, I haven't heard a 'how' theory from Vijay yet). I agree that in the most usable range the response of film is
    linear; but I doubt it is on the upper end. If you & Vijay are claiming that the non-linear response on the upper end doesn't
    afford you much usable information, then how do you explain the presevation of highlight detail in negative film?
     
  290. Some post holiday bonding again? Yeepe.
     
  291. Vijay, by the way you are right, I was looking at the first toe and the line straightening afterward so I was assuming the density axis as linear. Then the actual real toe is in the shadows and the highlights are then pushed out?

    Like Bernie said, I think we all agree on the facts, what is you explanation for them?

    Yesterday I tool some TMX nature pictures I developed in Xtol. Which expands the midtones and cuts the shadows and the highlights. Well I realized that I had taken the DRange test with TMX in the same roll and although I was expecting the highlights/shadows to be gone, a quick scan revealed they are like almost the same as the Ektar shots. (I will post later). That means you can comfortably add at least 1-2 stops on each side when developed with Tmax.
     
  292. Anyone in the Atlanta area with the 5DII yet? I would like to do a follow up on both dynamic range and resolution comparing Ektar, Velvia and TMX to the latest FF DSLR.

    Ellis?
     
  293. It is interesting to see that this forum seems much more mature than fotosidan.se (Sweden) where mentioning Ken
    Rockwell immediately made me an outcast in the same thread I started there a month or so ago. Here somebody
    thought KR shouldn't be the only source of info. That is much more sensible and I agree on that but despite that he's an
    excellent photographer and his level of knowledge surpasses most of the members at photo.net (how many of us have a
    US patent in the field of photography?).
    Please keep up the good habit of not killing the threads by being "besserwissers" (smart-asses)!
    Enjoy your thinking!
    Cheers
     
  294. Rishi: As for bromine ions becoming scarce with increased exposure? Not a chance. There are orders of magnitude more bromide ions than even the highest number of reduced silver metal atoms possible for a fully exposed latent site; indeed, if it weren't so, then a fully exposed negative would look black prior to development, & we know that's not the case (reductio ad absurdum anyone? :)
    OK, here's the reductio-ad-absurdum, since you asked for it. If there are orders of magnitude more bromide ions than reduced metallic silver, then that metallic silver can't be significantly blocking light and causing non-linearity now, can it? Unless it were at the shoulder, if then. (I don't think that this is true even at the shoulder, but that's my "theory", and I'll explain it in a bit.)
    Anyhoo, the question is moot, since we are all in agreement that film has linear response, as I said a long time ago. Mauro agreed to this, didn't he?
    Bernie, sure Vijay & I are in agreement because Mauro did a convincing job here! But, we aren't in agreement as to the HOW (actually, I haven't heard a 'how' theory from Vijay yet). I agree that in the most usable range the response of film is linear; but I doubt it is on the upper end. If you & Vijay are claiming that the non-linear response on the upper end doesn't afford you much usable information, then how do you explain the presevation of highlight detail in negative film?
    Whoa, whoa, slow down. Mauro did a convincing job about demonstrating that negative film has the greatest dynamic range, absolutely. But that's suggesting that this wasn't already known. Look at any of dpreviews tests - digital has barely 9 stops, and anyone who has used negative film knows that the dynamic range is at least 12-13 stops if not more. Indeed some people routinely overexpose portraits (controlled lighting - subject DR is significantly lower than film DR) by 2 stops to get "creamy skin tones".
    Preservation of highlight detail? Well, if it falls within the dynamic range of film it will be preserved. What's so difficult to understand about that? With soft clipping, you can recover a little extra information by linearizing the clipped region, but I don't see how that is more than the half odd stop or so. Once light increases to the point where you get into the flat part of the curve past the shoulder, all detail is lost, because density has hit max. Alternately once light decreases to the point that you cross the toe and get into the flat part of the curve, all detail is lost because density has hit min.
    Dynamic range is a characteristic of the system itself, so there is nothing to explain here. If film preserves highlight detail for another half stop, then by definition, dynamic range is a half stop greater. Once again, this has nothing to do with the curve shape - it is the ratio of the max to min; if by linearizing or whatever, the max can be increased and the min decreased, then by all means, calculate the ratio of the increased max to the decreased min - and that is your dynamic range.
     
  295. Mauro: Vijay, by the way you are right, I was looking at the first toe and the line straightening afterward so I was assuming the density axis as linear. Then the actual real toe is in the shadows and the highlights are then pushed out?
    Right - the density axis is itself logarithmic. So Ektar 100 has a 13 odd stop linear range. The toe is in the shadows and the shoulder is in the highlights, but there is a vast 13 stop range that is linear. As long as your subject's dynamic range is less than 13 stops, this will fit neatly in the 13 stop linear region (assuming you get your metering right, of course). If like your dead roses picture, the dynamic range of the scene exceeds the linear plus toe and shoulder regions, then some parts will either go all black or all white. This is clipping.
    Now you can "extract" some highlight detail and some shadow detail because a part of those are on the curved heel and toe. In a scan you can use photoshop to selectively manipulate highlights or shadows; and with conventional, enlarger based printing, you can burn or dodge. These are all techniques to linearize the shoulder and toe; to extract that last drop of detail.
    As I said, this may get you a half or even a full stop of "extra" shadow or highlight detail but my points are that
    a) That extra half or full stop is also part of the total dynamic range (after curve shaping or whatever) by definition. So if you actually manage to get 14 stops out of Ektar after linearizing the toe and shoulder response, then its dynamic range is 14 stops by definition.
    b) That extra half or full stop is not much compared to the 13 full stops of linear range. (It may be important because you don't want to blow out the bride's dress etc., but it is not significant magnitude-wise.)
    c) Digital uses charge storage (capacitive effects), so it may be possible to extract a half stop extra from digital as well, but that is moot, given that the inherent dynamic range for digital sensors is only 8-9 stops compared to 13 stops for film. Expressed as a ratio, film has a DR that is (4-5 stops = ) 16 to 32 times greater than digital even if you were to ignore the toe-shoulder non-linearity effects.
    d) Point c) is true regardless of curve shapes.
    That's all I was trying to say. Hope it makes sense.
     
  296. Vijay, "Anyhoo, the question is moot, since we are all in agreement that film has linear response, as I said a long time ago. Mauro agreed to this, didn't he? "

    I said if both axis are logaritmic, and there is a kink in the graph, then the logaritmic response happens earlier inbetween the shadows and the highlights.

    Here it is with both axis linear:
    http://shutterclick.smugmug.com/gallery/6616619_YJEwK#428460601_tWtTf-A-LB

    "Mauro did a convincing job about demonstrating that negative film has the greatest dynamic range, absolutely"

    No arguments there, ha ha

    "But that's suggesting that this wasn't already known. "

    Ektar was just released, so a) there is no direct comparison I know of b) I haven't seen other test that graphically show photographers the problems of the -known- limited dynamic range.

    "Dynamic range is a characteristic of the system itself, so there is nothing to explain here. If film preserves highlight detail for another half stop, then by definition, dynamic range is a half stop greater. Once again, this has nothing to do with the curve shape"

    True. You could have a wider DR with a system that is also linear. It just so happen that in this case it is not linear.
     
  297. Vijay, I just saw your follow up. I think we are all pretty much in agreement.
     
  298. If you & Vijay are claiming that the non-linear response on the upper end doesn't afford you much usable information, then how do you explain the presevation of highlight detail in negative film?
    I don't know (but take that with a grain of salt, as I am certainly no expert). I know virtually stuff all about sensitometry, and I am wondering what actually does happen after the shoulder. Vijay is suggesting that after the shoulder kink, the slope of the line zero's out. Is that the case? If it didn't zero out, but maintained a very small upward slope, then that could explain why adding exposure doesn't lead to a directly proportional increase in density. Certainly from the aspect of chemistry it seems to be a plausable explanation.
     
  299. Bernie, because I don't know. But both the actual results from my test and the Ektar publication (10^ -2.8 to 10^ ~3.1) show at least 14 stops. That is 50% more than the 40D. The 40D reaches the wall sooner, the exact explanation I'm not sure.
     
  300. Actually I think we're all making this more complicated than it needs to be. Thanks Vijay for reminding us that the shape of
    the curve doesn't have to have anything to do with maximal attainable dynamic range (though it can). Even within the linear
    range of response, negative film can record a larger maximal vs minimal input signal. That says nothing about the max vs
    min 'output' signal of the film, though, right? Correct me if I'm wrong Vijay. In other words, we could extend the dynamic
    range of digital by having a photocell that can hold or record an even larger charge than whatever the 40D's photocells
    currently can, yes?

    I have a theory for the shoulder of negative film, along with some graphs from Baines... Will wait till I get home from the
    airport to write more as this is just getting annoying on an iPhone...
     
  301. Okay, I am skipping some of the recent discussion as it doesn't appear related to the original one. I downloaded the full size scan and raw file. I processed both with Lightroom and then used Photoshop to put them together as a jpeg (I had to stretch the digital image a bit to match the film's dimensions). No sharpening or noise reduction was applied to either file. I attempted to roughly normalize the two scans using the white and gray patches from the QP Card and a bit of the black background so that there's some reference for comparison. Doing that also meant that I couldn't completely optimize the raw scan to preserve highlight detail- increasing digital midtone values to match the film, retaining tonal relationships between highlights, midtones and shadows, and not blowing the highlights are mutually contradictory. I'm not sure what this test shows other than negative film has a gradual roll-off of highlight detail but extreme highlights may not be color accurate (see what looks like a red cast). I don't see any scanner noise.
    00RfN5-93987584.jpg
     
  302. In other words, we could extend the dynamic range of digital by having a photocell that can hold or record an even larger charge than whatever the 40D's photocells currently can, yes?
    Just a bit of info related to this concept - the signal off a digital sensor usually doesn't occupy the full range possible (ie. in the 40D's case that would be 2^14-1 = 16383). I can't remember off hand whether it is the photocells that are filled to saturation, or if the ADC clips the results. I suspect it is the latter, because digital sensors (well, those from a year or two ago when I last researched this), are not fully linear. At the highlight end they can become non-linear. The clipping, I seem to remember, was to get rid of this behaviour (perhaps because the non-linearity was different depending on what colour channel you were talking about). In reality, the 40D only has values from 0 to about 13-odd thousand.
     
  303. OK, Arcane Post Alert. I'm going to try and answer some points raised, and may get into the really arcane stuff. Hopefully it is going to be interesting for some people, others please ignore as desired.
    Bernie: Vijay is suggesting that after the shoulder kink, the slope of the line zero's out. Is that the case? If it didn't zero out, but maintained a very small upward slope, then that could explain why adding exposure doesn't lead to a directly proportional increase in density. Certainly from the aspect of chemistry it seems to be a plausable explanation.
    Given that max density isn't infinite, there will come some point on the characteristic curve where more exposure won't do any good - all halide is already silver etc. Regardless of whether the slope after the shoulder is flat or slightly upward, a point must exist where the slope will become completely flat. That is your "max" point for dynamic range calculation. Same for the toe.
    Rishi: Actually I think we're all making this more complicated than it needs to be.
    Thank you. My point exactly. Dynamic range is a max/min ratio. No more, no less.
    Rishi: Correct me if I'm wrong Vijay. In other words, we could extend the dynamic range of digital by having a photocell that can hold or record an even larger charge than whatever the 40D's photocells currently can, yes?
    Oh yes indeed. In an electrical circuit dynamic range is limited by only two things (no surprise) - the noise floor and the peak signal output. For the active pixel sensor circuit I liked earlier, there is something called the Johnson noise (or Johnson-Nyquist noise) which is the noise floor. Below it, you can't distinguish signal from noise. The upper limit is a hard limit - no output signal can be larger than the supply voltage for the transistor amplifiers.
    There is always a pressing need to reduce power consumption, so reducing the supply voltage is almost a given. The only ways left to increase dynamic range therefore are to lower the noise floor, or to do companding. Companding is easily implemented, but takes up area on the sensor die, area that should really be used by the photodiodes, reducing the "fill factor" for the chip. Thus the benefit that companding offers is negatively impacted by reduced pixel size, or greater noise. A possibility is to implement pseudo-companding; some people are trying that too.
    There are ways to lower the noise floor - a simple one is to have larger pixels. Unfortunately, there is a limit to this too, since you want higher resolution (requiring smaller pixels) and greater dynamic range (requiring larger pixels).
    Ultimately though, newer semiconductor technology will easily hit the 13 stop dynamic range of film (or 80 dB as it is called in the parlance) and possibly even exceed it to 120 dB (a 20 stop range!). I predict 13 stop DR for digital in 3 years or so, and 19-20 stops in about 5.
    Hope I answered some questions, and hope it helps.
     
  304. Hey friends!
    Don't worry about film or digital and just take good photos,some time you see people spend a lots of money in cameras and they take horrible photos,bad taste.I like this page but in general almost all photography are cheese and ugly,so you talk to much and spend to munch time in from the computer,go out and take good picture.
    You can take good picture with any ship plastic camera, went you see a guy with big digital camera i almost sure that he take bad photos.
    Thanks,Juan.
     
  305. wow, lots of confusion and lots of lack of knowledge,. here are few very simple answers,. Canon 40D uses CMOS technology, and it is rather cheap camera,. while profesional systems such as Phase One ( I use P45+) have full 12 stop range,.. secondly,. scan should have been done on some top line scanner to compare those two,. juch as X5 from Hasseblad/Imacon and then you could see that in some ways digital back such as P45+ exceeds capabilities if film, even when converted to b/w. I recetly posted few portraits. done digitally,. you are free to check
    00RfSt-94039584.jpg
     
  306. Above shot was done with existing lighting using PhaeOneP45+,.. no fill in lighting was used, pay attention to deatil,. both in the hightlights and in the shadow
     
  307. "wow, lots of confusion and lots of lack of knowledge"
    fair enough, but you contributed to reversing this 'lack of knowledge' how...?
    p.s. i like this side of Vijay :)
     
  308. To be more clear, Mario, if you think the digital back has more DR, any ideas fundamentally how? Higher bit A/D converter? Photocells capable of more charge?

    These are the sorts of questions we've been trying to answer of late -- some of the more fundamental ones that might help us discern why we see certain behavior of film vs. digital when it comes to DR.

    Sorry if Mauro's thread wasn't initially intended to go down this path... but it's entirely relevant and very interesting (for some).

    Rishi
     
  309. Vijay:"Companding is easily implemented"
    How, in a digital sensor? Just curious if you have any ideas...
    Rishi
     
  310. And just to add to my earlier points, take a look at these DP Review curves for the 1Ds Mk III - they all show soft clipping; but the DR of this camera is barely 9 stops (8.6 stops).
    EOS 1Ds Mk III curves
     
  311. I forgot say that in some digital photos the image don't look real,is like Hollywood movies, fantastic, is to much,is because some people think like that in this days(plastic)
    I like film just because look more realistic,like Hollywood movies from the 50' the good time in North-America.
    Thanks.
     
  312. Nice shot Mario,

    But silver based films can easily exceed 12 stops. No biggy there. I have obtained a very usable 14 to 15 stops with dilute solutions and stand development. Twelve stops is hardly a badge of honor for a $30,000 piece of equipment!
     
  313. ok,. sorry if I wasn't so clear,. if you compare film which is scanned with much higher quality scanner with CCD sensor and then you compare that to Canon 40D which has CMOS sensor,. that test will be greatly in favour o film. Also, development of technology is such that every few months new things come out. What I was proposing was to test a film scanned on X5 Hasseblad scanner, which is about year old technology,. versus some digital back like PhaseOne P45+ Both use CCD technology. I shoot both film and digital at the same time,. film both color and b/w and I do about 7-8,000 shots on location each month, so I can freely say that I have experience both scanning and processing digital. Stil,. in some cases film is way to go,. but I found dynamic range of CCD digital back, pulling of the shots that film will never be able to do. I am including one recent shot,. attention is to back lit windows on the sides, and the dark face and white dress of Madonna at the middle of the shot,. only existing lighting i sbeing use
    00RfU1-94043584.jpg
     
  314. I agree Dave,. I have been shooting for 35 years and I am sad to see things like Rodinal dissaper,. which was developer of my choice,. I still shoot b/w,. changing back on my Mamiya allows me to do so,. and then have Velvia 50 for certain situation,.. which I scan with X5
     
  315. Sorry, Mario, where is the 'confusion' and 'lack of knowledge'? You still haven't set us hicks straight on that one yet. Couple of points:

    Firstly, you seem to be criticising Mauro for not using a P45+ and a hasseblad scanner. It would be great if we all just had one of these lying around, wouldn't it?

    Secondly, whatever the p45 can do in relation to DR, negative film can obviously do it better. 12 stops is short of the 13-odd stops we are talking about for Ektar. So what's your point, other than a shameless plug for your own photography?
     
  316. I say someone use a Fujifilm S5 at 400% DR and compare it to Film.

    Canon 40D is not the best of choices.

    Wouldn't it be better to figure out which DSLR have the best DR first then compare it film, instead of using a DSLR that is not known for DR???
     
  317. James and Mario, can you please post a 14 stop graduated example like the mine so we can appreciate the dynamic range of your systems?
     
  318. Bernie, I wasn't taking about Mauro,. I meant lots of other posts that I have read following his,. so do I apologize to
    Mauro,. I am not trying "to plug in" shamlesly my photography, I am way to busy proffesionaly for that. I was just
    trying to explain to some other members not to mix apples and oranges and having engineering background, I know
    how to set up accurate test. I was just trying to explain that when you do comparison,of two tests, film scanner and
    camera should be in the similar price range, similar class,. sort of speak,. that gives more realistic test,. also that
    both pieces of equipment are of similar generation. You can't scan film on CCD scaner and then compare it to CMOS
    camera,.. that is my opinion,.. regarding P45+ and X5, it was just an example to explain myself,. two similar
    technologies, that came out at about the same time
     
  319. Mauro,. from two images I have posted, and complexity of those lighting situation,. should be quite enough,.I believe photographers that dealt with similar situations will know what to look for. I do apologize if I have ofended you, I was trying to some other members explain difference in technologies and how aproach a test
     
  320. You have not offended. No worries. It is just not possible to see how many stops a scene had by looking at a picture of it.

    Without any complicated setup. Could you take a one picture at f8 of a dummy metering f64 (with the meter to the camera) and post it?
     
  321. Vijay:"Companding is easily implemented"
    How, in a digital sensor? Just curious if you have any ideas...

    Well, a sensor chip architecture is something like this: there are individual pixels; which are analog charge accumulators whose charge varies with the photons striking it. This can be implemented with various photoconductive materials (passive Light Dependent Resistors or LDRs) or photodiodes, or phototransistors etc, which have a p-n semiconductor junction etc.
    Then there is an amplifier that works as a buffer - a mechanism that allows the charge value to be detected - i.e., read out, without destroying the original charge. It does deplete the charge slightly - but being an amplifier, it takes a very small current (in picoamperes or less) and reproduces the same voltage as the photodiode at its output. That picoampere (10^-12 ampere) current is allowed to flow for just a few nanoseconds; this does not deplete the charge in the photosensitive device by any appreciable amount.
    The way this is organized, an entire row is read out at the same time. This is the usual architecture of a dynamic random access memory (a DRAM) - you provide a row select and all columns are simultaneously available to you.
    At this point, the pixel values are fed into an analog to digital converter that converts the voltage read out from a sensor to a digital representation. This is a process called quantization. There may be one ADC per column, one for the entire chip, or any number in between. This is dependent on the analog to digital conversion time, and how fast you want to read out the entire pixel array; and on the area of the die, power requirements and so on. This affects the frames/second specification of the camera, so some appropriate number will be chosen based on the overall data throughput that the sensor designer decides is appropriate.
    The digital representation of the analog value can be 8 bits, 14 bits, 16 bits or whatever. This increases the precision of the representation, not the dynamic range. The actual dynamic range is a voltage ratio, the "dynamic range" of a 16-bit vs. a 32 bit representation is a meaningless number, since the max value - FFFF for 16-bit, or FFFF_FFFF for 32-bit represents the same voltage - the saturation voltage of the photo cell.
    Once you have a digital representation, what you end up having is a matrix. As in a mathematical matrix of numbers. You can apply several algorithms for post capture processing at both the hardware and firmware levels and eventually get your JPEG or TIFF or whatever image.
    With this background, you'll see that companding must be implemented at the level of the pixel itself - i.e., a pixel must have a nonlinear response characteristic. There are semiconductor techniques that let you do this: see this IEEE paper.
    You could implement nonlinear amplifiers if needed, and here is a reference for this. These are called "waveshaping" circuits, but I am not sure if these will help if the actual photosensitive element itself has limited dynamic range.
    I'm hearing of a new technique that converts light to frequency inpixel - this seems to give a linear dynamic range of 115 dB and overall dynamic range of 130 dB (21.6 stops!!!). See this IEEE paper. This is far more promising than putting companding circuits in the sensor - and it operates at a peak voltage of 1.2 volts. Really impressive.
    Finally, here is some excellent reading material if you want to understand more of the nitty gritty. Noise, Dynamic Range and Bit Depth in Digital SLRs.
     
  322. I will try,. I am on location on a big project,. how do you want to set it up? I don't have f64 on Mamiya,. it goes to f22,. but I can play aorund with the speed
     
  323. It is just not possible to see how many stops a scene had by looking at a picture of it.
    Exactly! He berates you for not conducting a test properly and then posts some shots with no info or reference point on which to judge them. I could easily post a 5D shot of my daughter in window light and claim it had higher DR than film. He's wasting our time.
     
  324. I meant set the Mamiya at f8 and set the light so meters f64 with the sphere to the camera. If you happen to have a doll handy (I hope not) it'll do perfect as a subject. Should take just a couple minutes - nothing fancy.
     
  325. Mario,

    Is that sunlight through the stained glass? If so, that's amazing. Like, unbelievably amazing.

    Rishi
     
  326. I'm not bothered at all. I'm interested to see the results.
     
  327. "Mauro Franic , Dec 01, 2008; 11:42 p.m.

    James and Mario, can you please post a 14 stop graduated example like the mine so we can appreciate the dynamic range of your systems?"
    ::

    I never claimed the S5 had 14 stops. I figure if the test is going to be done then should a camera that was made with a wide DR as far as DSLRs are concerned be used.
     
  328. Mauro,. I will do it tomorrow,. it is 6 am where I am,. and got to hit some sack before we continue today,.

    Bernie,. by checking your portfolio,.I don't think you have enough experience to understand imagery I have posted. Anybody who has good knowledge of Ansel Adams zone VI system, can read from image how wide the range is, especially in b/w photo. And you have no right to insult people like you do,. especially people that have way more experience than you will ever have and that paid their dues on the road.
     
  329. Rishi,..yes,..this is sunlight through stained window and altar is lit by the candles,.. especially check that Black
    Madona didn't loose detail in the face, neither white dress burned out
     
  330. Vijay, could you give us non-engineers a jargon free summary of this companding thing? I'm not really sure how it works (or what it is actually beneficial for).
     
  331. Bernie,. by checking your portfolio,.I don't think you have enough experience to understand imagery I have posted.
    Mario, you are an arrogant twat (sorry Vijay, couldn't resist...). You lob in on the discussion, telling us all that there is heaps of confusion and misinformation, but don't bother telling us what exactly you are refering to. You suggest the test is flawed because the OP didn't use a P45+ and a Hasselblad scanner. You post two images that have an unknown scene dynamic range and expect us to gain something from this. You are the one who doesn't understand the technical aspects of imaging. But please, explain to us how from your black and white photo we are supposed to know what the scene dynamic range was?
     
  332. In fact, don't stop there, tell us with the colour photo as well...
     
  333. forget it Bernie,..
     
  334. forget what? I am willing to hear from a great master how you can tell scene dynamic range from an image with no
    other info to go by. Are you now admitting that you made wild claim that can't be backed up?
     
  335. Mauro,.. when I have mentioned P45+ and X5 I didn't meant that you need to use them,.. it was just an example of scanner and digital back on the same level of technology,.I don't know how better to explain.

    You can do just as well with 40D,. but to use the scanner that is similar technical range,. and off hand I cannot tell you which scanner to suggest, since I am not that much in that level of market.

    Point that I was trying to make, was that results can be more flawed if the scanner is of much different and better technology than 40D,. which is reletively low end camera,. no insult intended. Then you would get more accurate comparison,... so if camera is $ 1,000,..use teh scanner that is 1,000 and that is of teh same CMOS technology,.. that is my opinion and point that I was trying to make. My first comment for some other posts that I have read.

    Also, you can Google my name,. there will be interesting results that you will find,.
     
  336. Bernie,. your sarcasam is out of place,..great Ansel Adams have developed Zone VI sistem,. of judging prints and negatives on scale of 12 stops,. where VI ( or 6 ) is middle "Gray Value". That is very good described in three books that Ansel Adams did,. "The Camera" , "The Negative" and " The Print" Although many people don't shoot film anymore,. thos boosk are very valuable to study, because lots of those principles can be very well applied in digital photograpy and in processing digital photograpy
     
  337. Also Bernie, Rishi,. few posts above,. understood what he was looking at,. you didn't
     
  338. What sacrcasm? You say I insulted you, yet it was you who barged into the discussion claiming we were all stupid. You say I am sarcastic, but what of your "forget it" comment? It's clear from your shameless self promotion, and your condescending manner, that you believe yourself above us. Well, I am here to tell you, we aren't as stupid as you think. So let's hear your explanation of how we can tell what the dynamic range of your images was. Go through it step by step so we can learn how to do it. I bet you can't because it is nonsense.
     
  339. Bernie,.I just did with Zone VI,. but it was apparently over your head for lack of your knowledge of some basic things
    in photography,.. besides,.I never, ever called anybody stupid on any of the forums,. so don't put words and things in
    your post that I have never said
     
  340. So mentioning "Zone VI" is you giving us a step-by-step guide how to do it? So what is the answer (let alone the steps) to your two images? Seriously, try explain it. You might be suprised at how smart I am.
     
  341. Well this thread's going off the rails. I'm not sure the theoretical engineering debates are adding much to photo.net, and aren't bringing out the best in our posters.

    Mauro, in case you missed it I posted my renditions of your files at 7:48. Feel free to contact me directly.

    Mario- nice light.
     
  342. wow.....and lots of lack of knowledge
    Sounds like 'stupid' to me.
     
  343. Roger: Well this thread's going off the rails. I'm not sure the theoretical engineering debates are adding much to photo.net, and aren't bringing out the best in our posters.
    Roger, photography is as much an art as it is a science; and the entire photographic process is a product of engineering. So if you can appreciate that aspect of photography, great; but please don't deride it by saying that theoretical engineering debates aren't adding much to photo.net. A subject like the comparison of dynamic range of different media is a technical subject, with more to do with the science and engineering part of photography than the art. This also goes to those who say "just go and take pictures" - we evidently do, but we also want to understand more about the underlying processes. How is that a bad thing?
    That said, I agree that these discussions aren't bringing out the best in the posters.
     
  344. Well this thread's going off the rails. I'm not sure the theoretical engineering debates are adding much to photo.net, and aren't bringing out the best in our posters.
    I don't think it is the engineering aspects which are bringing the worst out in posters. It's one particular poster who barged down the door and told us we were all idiots, and then proceeded to show us nothing. The rest of us are enjoying the 'engineering' aspects of this discussion.
     
  345. holy smokes... I just scanned this entire thread. I'd just like to say 'thank you' to Vijay for providing all the information.

    Question for you, Vijay: When you say:

    'The digital representation of the analog value can be 8 bits, 14 bits, 16 bits or whatever. This increases the precision of the representation, not the dynamic range. The actual dynamic range is a voltage ratio, the "dynamic range" of a 16-bit vs. a 32 bit representation is a meaningless number, since the max value - FFFF for 16-bit, or FFFF_FFFF for 32-bit represents the same voltage - the saturation voltage of the photo cell. '

    Does the number of bits impact the 'tonal range'? what is the metric used to refer to the granularilty *between* the shadows and the highlights?
     
  346. Mauro,. I don't have it handy here since I am on location,. but explaining principles of Zone VI, when reproducing art, I
    forgot to mention that I quite often use color and gray scale strips calibrated by Kodak.

    If You photograph gray scale strip ( which has squares from pure white to pure black in increments of 1/2 stop,.You
    will get simple and very accurate test ,. you will see at which point you don't see difference between two squares at
    oposite end of the strip,.for film and for digital camera and that will give you accurate difference in range
     
  347. Mario: Rishi,..yes,..this is sunlight through stained window and altar is lit by the candles,.. especially check that Black Madona didn't loose detail in the face, neither white dress burned out
    No offense, Mario, but here's my take on your Madonna image.
    First, the altar is lit by what looks like electric "candles". Far brighter than wax candles.
    Second, the Black Madonna is lit by several halogen spotlights barely ten inches from her face, again very bright.
    Third, there isn't direct sunlight shining on the stained glass. You've used a wide angle lens, and it may not be immediately apparent, but the stained glass windows are at right angles or so to each other. All the stained glass is more or less equally illuminated - that can only be possible on an overcast day or when the sun isn't shining directly into the camera.
    Fourth, stained glass is pretty absorptive of light. If some part of the stained glass appears as a midtone on zone VI, it has already absorbed 82% of the light falling on it (18% luminosity is mid gray etc).
    Fifth, I see shadows of the electric candles on the altar; meaning that there is a source of bright light, hidden from the camera, that is illuminating the entire altar. My guess is a halogen flood. This is also apparent from the sharp shadow of the trim on the wall, and the reflections from the gold edge trim on the wall.
    Sixth, there is evidence of clipping where there is clear glass, so your metering was such that it dealt correctly with the interior and most of the stained glass parts, letting those highlights blow. Could do that with any camera, any film.
    Needless to say, what the first five things do is significantly compress the dynamic range of the scene, and the sixth demonstrates that highlight detail preservation wasn't that important to this image anyway.
    My guess is the scene dynamic range of everything except the clear glass parts of the windows is no more than 7-9 stops, so most digital stuff will cope pretty well. If you did this with a P45+, your scene was well within its dynamic range. You could have used color negative film and you'd still have been OK. For this scene, I'd dare say that even a 40D may have been fine.
    So, I'm sorry, but this picture does not demonstrate a wide scene dynamic range nor does it demonstrate that digital deals with wide dynamic range well.
    It is a great picture, nonetheless.
     
  348. Does the number of bits impact the 'tonal range'? what is the metric used to refer to the granularilty *between* the shadows and the highlights?
    In digital sensors, 'Bit Depth' directly relates to the granularity between tones. Think of bits as steps in a ladder. The top and bottom are fixed, but the number of steps in between can vary. The more steps you have, the shorter the distance (granularity) between steps (tones). Of course in film, there are no steps as such as it is a continous analog system. But in digital more does not necessarily equal better. See this article for a good explaination why.
    Just a comment on Vijay's statement about bit depth not influencing dynamic range - this isn't the complete story. Bit depth does have a relationship with dynamic range. It is as follows: A 14-bit sensor if recording a scene of the same or wider latitude than the sensor's dynamic range, is capable of capturing a MINIMUM of 14 stops. This is because each successive bit is double the value of the preceding bit (ie. 2^0=1, 2^1=2, 2^2=4, 2^3=8, and so on..). So when you double exposure, you are increasing your dynamic range by 1 bit. ie. 1 stop exposure = 1 bit. So why don't 14-bit dslr's have 14 stops of DR? Because in the lower few stops, noise overwhelms the signal. The other point to remember, as I mentioned above, is that most DSLR's don't capture their full bit-depth of data. Most clip somewhere short of the theoretical maximum.
     
  349. David Punsalan: Question for you, Vijay: When you say:
    'The digital representation of the analog value can be 8 bits, 14 bits, 16 bits or whatever. This increases the precision of the representation, not the dynamic range. The actual dynamic range is a voltage ratio, the "dynamic range" of a 16-bit vs. a 32 bit representation is a meaningless number, since the max value - FFFF for 16-bit, or FFFF_FFFF for 32-bit represents the same voltage - the saturation voltage of the photo cell. '
    Does the number of bits impact the 'tonal range'? what is the metric used to refer to the granularilty *between* the shadows and the highlights?

    Yes, the number of bits tells you how many tones you can represent between white and black. Lets say that you use 2 bits to represent all the tones; then you can represent only four tones: 00 = pure black, 01 = dark gray, 10 = light gray and 11 = white. This is an arbitrary representation, I could choose any combination of 1's and 0's to represent any tone. Now if I go to a 3 bit representation, I can represent 8 tones; 000, 001, 010, 011, 100, 101, 110 and 111. This is a simple demonstration that the number of tones you can represent with N bits is 2 to the power of N. Thus, with 16 bits, you can represent 2^16 or 65,536 tones between black and white.
    How many tones is appropriate depends on the sensitivity of the human eye to distinguish between two "adjacent" tones; thus if you have 65,536 gray tones, and you print these on paper but you can't distinguish two adjacent tones, then using a higher bit representation is of no use. Most people can't distinguish between 8 bit and 16 bit representation in color. Most people can distinguish between 8 bit and 16 bit black and white; this is quite apparent with Adams' zone system - he expanded the tonal scale to 10 bits or more (10 or more zones); and nearly everybody sees that enhanced tonal range in his prints immediately.
    I hope I answered your question satisfactorily.
     
  350. Mario Novak , Dec 02, 2008; 05:08 a.m. Mauro,. I don't have it handy here since I am on location,. but explaining principles of Zone VI
    Must have missed that explaination. I recall you only MENTIONING "Zone VI", not explaining it.
     
  351. Vijay: No offense, Mario, but here's my take on your Madonna image Vijay.... You da man!
     
  352. Bernie: Just a comment on Vijay's statement about bit depth not influencing dynamic range - this isn't the complete story....etc
    It is just a representation; a bit is not necessarily equal to a stop; for instance I can encode a 9 stop dynamic range as 14 bits; this is commonly done with the current crop of DSLRs; so a bit does not imply a stop. Similarly, if I take an 8 stop dynamic range and encode it as two bits, then each bit represents more than a stop of change. Of course I lose tonality, but number of tones has nothing to do with dynamic range. Film has 13 stops of dynamic range, but theoretically infinite tones within those 13 stops. I could use 128 bits to represent that if I wanted.
    Finally, there is no reason to assume that the representation goes like 00 = darkest, 01 = dark, 10 = light, 11 = lightest; this is just one of many representations. I could very well do: 00 = darkest, 01 = dark, 11 = light and 10 = lightest. This representation has a one bit change per tone change, and is called the Gray code (Gray after a person) and is advantageous for many many applications, especially asynchronous data transfer. In fact, you don't even know what encoding (representation) is used at the hardware layer, this is a designer's prerogative, and you can freely convert between encodings.
    So sorry, 14 stops does not mean 14 bits. The former is a dynamic range, the latter the quantization step (1/(2^14) steps make up the dynamic range. Only in this specific case does a bit mean a stop; but that is generally not the case.)
     
  353. I can encode a 9 stop dynamic range as 14 bits; this is commonly done with the current crop of DSLRs
    Yes you can, but you will have 5 empty bits. You could have captured the same scene with just a 9 bit camera (if noise didn't play a part, of course).
    Similarly, if I take an 8 stop dynamic range and encode it as two bits, then each bit represents more than a stop of change.
    I agree. Nothing I said goes against this.
     
  354. Bernie: Vijay, could you give us non-engineers a jargon free summary of this companding thing? I'm not really sure how it works (or what it is actually beneficial for).
    It is quite simple actually. The concept is that when the signal is small in amplitude, it can get lost in the noise of a transmitting channel (the signal comes from each pixel, the transmitting channel is the entire electronics of a digital camera), so amplify the small signals more. When the signal is big in magnitude, it can take care of itself - i.e., won't get lost in the noise, so you can amplify that less. This is called "compression". It allows you to use a channel with limited dynamic range to carry a signal which has a wider dynamic range, but has been compressed to fit the channel.
    Once the signal passes through the channel, it goes through the opposite amplifier - one that attenuates the small signals more than the large signals. This recovers back the dynamic range of the original signal. This process is called "expanding" giving you the term "companding" - "COMPression + expANDING".
    This technique has wide application in electronics - such as in tape recorders. Magnetic tape has limited dynamic range, so you have to compress the signal while recording, and expand when playing back.
    With digital cameras, it will have application as soon as the pixels have larger dynamic range than the other electronics. Right now, the pixel DR is itself limited, but there are semiconductor processes that make a pixel itself generate a larger voltage change when less light falls on it but a smaller voltage change when more light falls. A sort of non-linear compression right at the pixel level. Of course, the signal will be expanded later to recover a larger DR.
    Hope this helped.
     
  355. Thanks Vijay. This is what I understood it to be, but I don't know how you can get the original signal back from a compressed signal. When the signal is compressed, wouldn't you get different tones mapping to the same output? Like in the Levels command in Photoshop: if i compact the levels, but then later expand them out, there will be gaps in the histogram. How does this work?
     
  356. Bernie: Yes you can, but you will have 5 empty bits. You could have captured the same scene with just a 9 bit camera (if noise didn't play a part, of course).
    Nope. Suppose I have 1 stop of dynamic range: the max value is just twice the min value. What do I do to get quarter stop or half stop changes?
    I could represent a stop at 3 bits, then I'd get one-eighth stop tonal differences. For a 9 stop DR, if I wanted eighth stop tones, I'd need a 9x3 = 27 bit representation.
    No empty bits in this case either - and if an engineer is dumb enough to put "empty" bits in an electronic circuit, he is liable to get fired. Empty means that there is electronics out there that takes money to manufacture, that consumes energy from the battery but does nothing useful. Yep, if someone who worked for me did that, out he'd go.
    So then, there is no inherent relation between "stops of DR" and "bits of representation". Whatever relationship you might see is incidental.
     
  357. Nope. Suppose I have 1 stop of dynamic range: the max value is just twice the min value. What do I do to get quarter stop or half stop changes? I could represent a stop at 3 bits, then I'd get one-eighth stop tonal differences. For a 9 stop DR, if I wanted eighth stop tones, I'd need a 9x3 = 27 bit representation.
    You've lost me. Try that again.
    and if an engineer is dumb enough to put "empty" bits in an electronic circuit, he is liable to get fired.
    That's hardly a good analogy. The better analogy is this: If an engineer is dumb enough to not put enough bits in an electrical circuit, he is liable to get fired.
     
  358. Bernie: When the signal is compressed, wouldn't you get different tones mapping to the same output?
    No, that wouldn't work. What you really want is that the difference between the tones be reduced. This is difficult to understand with digital (because it is analog compression), so imagine it with... say water in a cup. Say you wanted to represent four levels - quarter full, half full, three quarters full and full. You could do this by using a smaller cup, but saying that half full of the new cup is a quarter full of the old cup; 3/4th's of the new cup is 1/2 of the old; 7/8 of the new is 3/4 of the old and full of the new is the same as full of the old.
    You don't lose the individual tones, you just space them closer together. Different tones must map to different outputs, otherwise you lose information and can't recover it. In the analog world this is possible because there is no quantization (or it exists at an atomic/quantum level, so the quantization step is too small to matter). Much harder to do the same with digital bits.
     
  359. You don't lose the individual tones, you just space them closer together.
    But there are effectively no 'individual tones' in analogue, as the signal is continuous and hence has theorectically infinite tones. Now I know that in reality there will be individual tones, but wouldn't they be so narrow in band that they would be lost amongst noise. So I can't see how you could not get different tones mapping to the same output (I'm not doubting that this is actually true, I just can't get my head around it).
     
  360. Bernie: If an engineer is dumb enough to not put enough bits in an electrical circuit, he is liable to get fired.
    Ha, ha - that too. Bernie, realize this: one extra bit per pixel translates to a total of (for the 1Ds Mk III) 21 million extra bits, or roughly 3 MB of useless data. 5 bits is 15 MB of data that has to be represented as electrical circuits (that consume power, generate heat and contribute to global warming and what not) and processed by computers, stored on hard disks etc; and all that is per image.
    As engineers, we don't want one iota less than what is necessary, and not an iota more. We sometimes "overbuild"; but that is because we want to prepare for unanticipated loads. My dad is a civil engineer - he designed bridges, but is retired now. He would overbuild by a factor of 3 - because what if there were an earthquake with a train on the bridge? But if we could guarantee loads, we rarely overspecify or overbuild.
    Now a particular camera could saturate at say 12 bits, but have a 14 bit ADC - that is usually because designers keep the same chipsets across product lines as far as possible to keep costs down. A separate 12 bit ADC may not have been justifiable when the next gen would use 14 bits anyway, and a 14 bit chip were already available etc.
    Like I said, this relationship is incidental; not an inherent property that n bits represent n stops.
     
  361. Bernie: You've lost me. Try that again.
    OK. Say you look at the marks on a ruler - 1 inch, 2 inches, 3 inches... upto 8 inches. This is a dynamic range of 8:1 = 3 stops. Correct? (1->2 = one doubling; 1->4 = two doublings, 1->8 = 3 doublings, or 3 stops)
    I could represent this by three bits - 000 = 1 inch, 001 = 2 inches, 010 = 3 inches, 011 = 4 inches, 100 = 5 inches, 101 = 6 inches, 110 = 7 inches, and 111 = 8 inches.
    But how do I now represent the half inch marks? Like how do I represent 1.5 inches, 2.5 inches etc? I could do it by increasing the bit depth (or reducing the quantization step) to 4 bits.
    Then, 0000 = 1 inch, 0001 = 1.5 inches, 0010 = 2 inches, 0011 = 2.5 inches etc.
    I've doubled my "tones" from 8 to 16, but the dynamic range has still remained 3 stops. There are no empty bits, since each 4-bit value now represents a half an inch of change.
    Hope this helped.
     
  362. bernardwest