Quality of negative scanners?

Discussion in 'Film and Processing' started by stuart_henry, May 15, 2008.

  1. I just got a Nikon N65 and I'm looking for a cheap practical solution for
    developing the film. It seems from my research online that scanning negatives
    or slides are good options. People seem to say that scanning negatives gives
    better results but it may be more difficult to get the color right, etc. I
    don't mind having to adjust the color, I just want good resolution with no grain
    or blurring. My question is, how good is the quality I can expect from a home
    film negative scanner?

    Just as a reference point, I looked at this scanner...which scans at 4800 x 9600 dpi
    http://www.circuitcity.com/ccd/productDetail.do?oid=184510&WT.mc_n=4&WT.mc_t=U&cm_ven=COMPARISON%20SHOPPING&cm_cat=GOOGLE&cm_pla=DATAFEED-%3EPRODUCTS&cm_ite=1%20PRODUCT&cm_keycode=4

    To scan a 35mm area, this would produce a 6614x13229 image...which is 87 MP.
    That doesn't sound right...!

    Ive been browsing around trying to find images that other people have produced
    to get a sense of the quality. Some of them have weird artifacts and distorted
    colors...others are crisp and look better...
     
  2. In general flatbed won't do any good on film scanning, but the actual quality depends on the scanner, film, exposure and developing, and then there is calibrating and postprocessing. I haven't done scanning for at least 2 years but can tell you that if quality is what you are after, be prepared to spend lengthy hours just to experiment until you work things out (check out Vuescan), and then you will need an expensive model if you want to scan slides. For colour negatives, just overexpose the film a bit and the scans will not be as grainy, but grain free scans are unreleastic expectation. For that you will be better off with DSLRs.
     
  3. The Nikon Coolscan V is much better than consumer-grade flatbed scanners. You should expect to spend about $550. And to have a bit of trouble finding one in stock.
     
  4. The latest color neg films are made for scanning, Kodak Portra and Fuji 160s Pro. They are significantly better than earlier versions. Scanned with a good scanner, they are the equal of dslr files.

    Scanning works if you know what you are doing, but it is a lot of work compared to a digi cam. With the digi cam, the file goes right into the computer and you are finished. 95% of photogs have come to this conclusion.

    My advice is to get a digi slr with the money you save from not getting a good scanner like the Nikon which is out of stock everywhere I am told.

    Or buy a cd with the film images scanned onto it when you have the film developed. They will not be great high rez scans.

    The flatbed scanner of choice is the V700 Epson if you do not get a dedicated film scanner. One of the Nikon models if you want a dedicated film scanner.

    Consumer flatbed scanner resolution figures are notoriously over rated. The one you picked will get you a digital image, but the quality is really unknown as nobody here uses one that I know of.
     
  5. Indeed for a good flatbed scanner: (Epson) V700 or V500 when scanning negatives or use a dedicated Nikon film scanner.

    Looking at the OPTICAL resolution of the flatbed scanners max. 6400 dpi divide it by two and you have a practical scan resolution (for 35mm). Almost all flatbed scanners are producing only data over 4000 dpi, not more details.....
     
  6. Stuart, color is, indeed, the bugaboo with color negatives, and getting the color perfect does take some skill, experience and time, but it can be done. On the other hand, with the right software (eg software which supports profiling) transparencies CAN become almost a turnkey operation, but be aware that getting the most from denser (eg Kodachrome) or higher contrast (eg Velvia) transparencies does require better quality equipment. The bottom line is that, given good equipment and experience, I'd still choose my film type based on shooting requirements and personal preference, not scanning requirements. Both CAN scan very well.
    Take a look at this page from my website for some additional examples. I reference this page as all but 2 of the shots on it are scanned 35mm negatives (Nikon LS-50 dedicated film scanner). Of the remaining two, one is from a 35mm transparency (Astia 100F), and one is from a digital SLR. I use Vuescan for color negatives, and Silverfast AI Studio (with IT8 calibration) for color transparencies.
    As for flatbeds, well... horses for courses. I'd also recommend a dedicated film scanner (which these days pretty much means Nikon if you're talking reasonable prices) over the comprise represented by consumer level flatbed scanners. That recommendation, BTW, is based on direct experience, not internet hearsay.
    Scott
     
  7. I use a Nikon IV scanner which i bought used last year. When i have a properly exposed slides (i almost exclusively shoot slide film, some BW negs, almost never color negs) i love it to death. Extremely sharp and clear IMHO. If something is not exposed just right, its better to just give up, though, especially kodachrome. That is where it gets really hard to get the colors right, when it is under or sometimes over exposed.

    The quality of the film you use makes a big difference. I get a lot more keepers and straighter colors from say, Provia 100 than Sensia 100. I shoot both film and digital, and both have their uses. I would say go for the dedicated scanner, and if it is something important you are shooting, shoot good film. I personally shoot slides because i think that they usually scan better, but just be ready to accept the inherent disadvantages of a narrow exposure latitude etc.
     
  8. Scanning film is not a casual endeavor. It takes skill and practice to achieve consistent results, and a lot of time (2-4 hours/roll just to scan). That said, the results are superior to anything I've done in years of darkroom experience.

    The best way to achieve good photographic results are with a dedicated film scanner. The Nikon V (LS-50) gives outstanding quality at an affordable price (~$500). The resolution is about twice that of the best flatbed film scanner (Epson V700/750), with better contrast and sensitivity. It is also much faster to use the strip scanner than to use a film holder, and requires only minor registration adjustments.

    In the end, you may find that a 10+MP DSLR gives better results. The choice to scan or go digital is based more on personal satisfaction and involvement than anything else.
     
  9. If you think the Canon delivers a true 4800x9600ppi, I'd like to sell you a bridge in Brooklyn.
     
  10. Stuart, I have a Nikon CoolScan 4000 ED that I picked up for $350 off eBay. It works GREAT. I shoot Fujicolor Reala or Pro S, both extremely fine grained, and get fantastic results. Although I have a Nikon D200 system I still enjoy my manual focus Canon FD film gear (great glass!). The enjoyment of these cameras is reason enough to own a film scanner for me. Pick a color neg film that you like, shoot a few of test shots of a skintone and a macbeth color checker under miday sun. Find the best exposure, (tip: don't over expose too much like you would for optical printing, I use box speed) use the scanners curve controls and the color checker values to dial in a curve setting. Save the curve. Shoot another roll of average subjects and scan and tweek the curve as needed. Yes, you have to learn how to scan, but it's not brain surgery. There is plenty of help here on Photo.net and other places on-line. Depending on crop, at full resolution, I get about a 14MP output that prints beautifully. Color neg film gives a great exposure range plus you can use Digital ICE while scanning to reduce artifacts. By converting in Photoshop, I can get a nice B&W image from the same scan. A good film scanner breathes new life into older film systems and gives access to digital tools like Photoshop and Inkjet printing. BTW, I also own an Epson V750 scanner. It not quite as good on 35mm as the Nikon but it's dang close and the best flatbed scanner for 35mm I have seen plus it does MF and LF films superbly. A good flexible tool. Good Luck and enjoy!......Lou
     
  11. Thanks so much for all your prompt feedback! I certainly wasn't expecting to get so much good advice so soon.

    Scott, those images on your website look decent..but they are only 0.6 MP...so I'm guessing you have down sized from the digital scan considerably. What I'd like to see is a raw image of the scan output...so I can see how much total resolution it has, see how much work would need to be done in Photoshop to clean it up etc. As I said I am no stranger to Photoshop so that's not a problem for me.

    Louis, how does the quality of that 14 MP film scan to the output of a 14 MP DSLR? I haven't owned a DSLR but my friend gave me a bunch of raw files from her Rebel (I think it was 8 MP but not sure) and I was pretty shocked/disappointed to find that the "effective" resolution was much smaller...they felt more like 1 MP images that had been scaled up.

    What is the physical difference between a "flatbed" and "dedicated" film scanner?

    Don't bother trying to persuade me into a DSLR...I have so many reasons not to go down that route. For one thing, my primary impetus for getting a camera is for my upcoming 1 month trip into the African bush where there will be no electricity. I can't afford a good DSLR and if I could, I wouln't want to carry around some big expensive fragile thing anyway. Digital cameras stink...I don't want to lose photo opportunities because my camera has to be "turned on" first, or couldn't pull the trigger fast enough..or ran out of batteries...or pictures werent as good because it needed a longer exposure time. I just want cheap, simple, lightweight, reliable, sharp film camera. It doesn't matter that my intended medium is a digital file -- I still think its better to make the conversion to digital in the lab rather than in the field.
     
  12. One more thing...

    Another option for me is to get slides, so I can inspect them and see which ones I like, and then send the good ones off to be professionally scanned. This would make sense for me if the quality of a professional scanner is going to be much higher megapixel/quality than what I could do by myself.
     
  13. Hi Stuart,

    May I ask you a question, what will you do with the scans you make?

    If you will view them on your monitor or post them to the Web, then a good flatbed scanner like the Epson V500 or V700 is all you may need. The limiting factor to the displayed image will be display monitor.

    If you will make smaller sized prints - 4x6 or even 5x7 - the flatbed will still suffice.

    If you will make larger prints or crop your images in smaller prints, then scans will benefit from dedicated film scanner.

    The advantage film users have is we always have our negatives or slides. If you need an occasional high resolution scan, you can always either re-scan that particular frame at higher resolution, or send it out for professional scanning.
     
  14. A flatbed scanner has a glass bed on which you lay the film (or print), generally in a film holder. There is a light in the cover and the film is scanned from beneath the glass. In the latest generation of high-quality consumer scanners, there is an array of microlenses on top of a full-width sensor which slowly moves down the length of the bed. The fact that the imaging is comparable to the compound eye of an insect means that the actual resolution is significantly less than that which the sensor cell spacing would suggest (the "optical" resolution).

    A dedicated film scanner has a fixed, high quality. multi-element lens. The film is held in the optical path and the lens forms an image directly on the sensor. To scan, the lens assembly may move (e.g., LS-50) or the film holder may move (e.g., Nikon LS-8000). The optics are optimized for a fixed reproduction ratio, which gives corner to corner sharpness. Front surface mirrors are used to make the optical path more compact.

    High-end graphic arts flatbeds use a single high-quality lens and have highly stable scanning mechanisms. The price is measured in $10K increments. Imacon scanners are in between - using a moving film holder with a fixed lens, and no mirrors in the path.
     
  15. Hi Stuart, on a purely resolution level the scans compare favorably to my digital cameras. However not only do the scanned images have visable grain, but come nowhere near the image control available with a digital RAW capture. A lot depends on your final output size. On 8x10 prints I don't see much difference. A digital SLR may not be suitable or desirable for your needs but they, in and of themselves, are wonderful tools nor are they especially fragile. A flatbed scanner is an all purpose scanner device digitizing both reflected (prints, documents) objects and transparent (film) objects. They range from cheap office type units to sophisticated (read expensive) high-end units. Film scanners only scan film. Different units offer options for different sizes originals, resolutions and price points with professional drum scanners at the high ens and small desktop LED scanners (like my Nikon CoolScan 4000ED) at the consumer end. To wit, a professional drum scan created by an experienced scanner operator is going to be significantly better than a scan by a scanning novice on a consumer-level scanner. I hope this helps you.......Lou
     
  16. "... how does the quality of that 14 MP film scan to the output of a 14 MP DSLR?"

    It depends. Let's assume you're using a Nikon dedicated film scanner. You'll do better with a 4MP DSLR if comparing to 800ISO color negative film. On the other hand, you can approach 20MP (albeit a noisily) if you're comparing against 100 TMAX or Fuji Acros.
     
  17. All pixels are not equal, and it is my experience that one pixel in a DSLR with good glass is equivalent to 3 pixels using a Nikon film scanner. A 4000ppi scan is grain-sharp. With greater resolution, the results are limited by grain and inherent diffusion of chemistry and light within the emulsion which reduces edge contrast.

    In short (and based on experience), a 6MP camera (D1x) or D2h (4.1MP) is comparable in sharpness to a 24MP film scan on a Nikon LS-4000ED, and has much less "grain" and better color. A 12.3MP D2x far surpasses 35mm film, and approaches medium format quality. When I shoot film, it is medium or large format only.

    The main advantage of a scanner, particularly a medium format scanner, is that you can achieve the results of an high-end DSLR at a fraction of the initial cost. The challenge is to find a processing facility for medium format film other than mail order.
     
  18. Does the ISO speed of my film have a large impact on scan quality of the negatives? I know that faster film is going to be more granular by nature, but I'm not sure if the effect is compounded during the scanning process. I was planning on using 500 ISO film.

    Brooks, 99% of my pictures will be viewed in digital format on the computer screen. I would probably be happy with just 4 MP of information content -- by that I mean, the amount of sharpness you would expect from taking a 100 MP image and downsampling it to 4 MP in Photoshop, so there is no blurriness.

    For some of my favorite pictures, I may want to make prints -- probably just 8x10. For those I would probably have it sent somewhere to be scanned professionally.

    Edward, you mention that all pixels are not equal..and I already know what you mean. I was given a bunch of RAW files from my friends DSLR Rebel taken at 8 MP...but I don't believe they really had more than 1 MP of actual "information content."

    It is disappointing to hear that equal MP from a negative scan are worse than DSLR of the same resolution -- but I guess that's just a fact I will have to live with. This discussion makes me want to consider medium format film cameras in the future...but not for this trip.

    Edward, can you recommend some places where you can send negatives to get truly high quality professional scans?
     
  19. Stuart,

    The reason I pointed you to that particular page was because you seemed to have concerns about color quality from negatives. The scans I pointed to were largely of color negatives, so I thought they'd serve to address that point. Yes, of course, they're down-sampled - from original files which are about 22Mpixels.

    But further to your original, and follow-up, questions, I would caution you about the whole mega-pixel & 100% view thing. You simply can't compare film scans to digitally sourced images on a pixel by pixel basis. The answer to how well-scanned film compares to digital images is far more complex than that.

    I would also caution you not to set unreasonable expectations. In your original post, you say "...I just want good resolution with no grain or blurring..." Sorry, the grain-free part just ain't gonna happen. Even fine grained transparency film like, say, Velvia 100, will show visible grain in a 100% view of a true high-resolution scan. Film has grain - it's a fact of life, and you seem to be expecting film to be something that it isn't. If you don't see grain, you're not talking about a high res scan - or there's something wrong with the scanner. In fact, one of the challenges in learning to post-process film scans, as opposed to digitally sourced images, is in how to PP to bring out the best in a film scan without accentuating grain.

    Using my Nikon scans, I can make 13x19 prints from fine-grained 35mm film with no grain-reduction whatsoever which appear grain free even from nose-on-the-glass distances. But that's very different from saying, "I want a grain-free scan". Scan or no scan, we're still talking about film images which have grain.

    Rather than get overly concerned about how images appear in full-sized scans, I'd look at some of the comments like Edward's above, with which I fully concur. I get better prints from 35mm film using a Nikon LS-50 than I have ever gotten from either my own, or a professional lab's, traditional prints from the same film. Hands down, no question. In other words, a well-executed scan workflow gets more from film than has ever been possible.

    Isn't that what you're really interested in?

    Scott
     
  20. "Brooks, 99% of my pictures will be viewed in digital format on the computer screen."
    OK, Stuart, now I'm really confused as to what your expectations really are. If that's what you're concerned with, why do you care what 100% views look like? Higher resolution images only matter if you're going to be printing to large sizes. The fact that a 100% view 22Mpixel scan doesn't look like a 100% view 1DsMkIII image, truly does NOT matter for your purposes. Not at all.
    Scott
     
  21. Les Sarile,

    Nice..thanks for that excellent comparison! The Coolscan 5000 scans are about on-par with the quality I'm looking for in my general purpose shots. It's out of my price range right now, and I guess that means I can stop bothering to look at the lower end scanners. I'll just have to get all my film scanned professionally until I can afford a scanner of adequate quality.
     
  22. Scott, I wanted to see 100% resolution views because that's what I'm going to be dealing with if I buy the thing. Looking at a down sampled image, I don't know what kind of information was lost (or not lost) during the down sampling process. The pictures you linked were a bit small for on-screen viewing.

    The images that Sarile linked were (in my opinion) just barely good enough for on screen viewing purposes. Yes, they are 3600x5500 raw...but I would end up doing a lot of post effects to hide the effects of grain...bilateral filter, maybe a median filter...then down sample it to a much smaller size to get a crisp image. If my raw scan was already 2000 pixels, then I'd be playing with thumbnails.
     
  23. Stuart, you may want to look at this page on my website that I dedicated to some of your questions:

    http://www.boeringa.demon.nl/menu_technic_optimalscanningresolution.htm

    I think you will find some answers, and yes, the ISO speed does matter for final quality. Sometimes it's simply better to scan at a lower than maximum resolution, instead of trying to apply PS filters to hide grain. It will not improve quality of your final scan. Again, see the webpage...


    Marco
     
  24. Also, if you expect to capture grainless images comparable to 10-12MP digital camera's, it can only be done using all but the least sensitive (ISO 100 or 50) film types.

    You just can't expect a 35mm negative at ISO 400 to have that cary that amount of true information... If you want to shoot at that ISO and still get detailed - or better said large - scans (e.g.> 12 - 30MP), you need to upgrade to MF or LF.

    On the upside, you plan for Africa, so ISO 100 at daytime should not be an issue...
     
  25. I can give another thumbs up for the nikon scanners. it's advantage over say an espon
    flatbed, is the ability to focus all parts of the slide to ensure consistent sharpness over
    the whole slide/neg.
     
  26. Nikon Coolscan V ED here. Unless you plan to do extreme crops, everything to 400 ASA will be perfectly useable.

    400 ASA at full resolution will look a bit grainy, but so will an enlargement from the same negative.

    I've had great experiences with Kodak 200 HD film, myself.

    On the whole, if you're dedicated to film, any of the Nikon scanners is well worth the investment.
     
  27. Stuart said, "Edward, can you recommend some places where you can send negatives to get truly high quality professional scans?"

    Considering the effort I find necessary to get good negative scans, I doubt you could expect much from a commercial institution at a reasonable price. Reversal film is another matter. Scanning reversal film is straightforward with a calibrated scanner and CMS-compliant software (e.g., Silverfast), but negative film remains highly subjective.

    You would find it more satisfying and economical to buy and use a scanner yourself.

    The main advantage of the LS-5000 over the LS-50 (V) is the ability to use automatic slide and roll feeding attachments. This is a great time-saver. With an older (slower) LS-4000, it takes about 2 hours (or less) to scan a 36-38 exposure roll of 35mm film with the feeder, but 4-6 hours if you are limited to 6 frame strips. AFIK the image quality is identical - perhaps Les or others could comment in this regard.
     
  28. -Save your money, time, and effort. Go digital. If you are ending up scanning everything you shoot anyway.....
     
  29. Scanning is like buying your first shovel. You can dig your own flowers; foundation for a new house; a boat slip; the Panama canal

    . Like scanners shovels also emit a gas that blinds the new user. This gas makes them believe shovels and scanners require little time; and that weeks worth of work will be instant.

    Part of justifying the purchase of a scanner is ignoring the time involved with transforming thousands; tens of thousands of images into the digital domain.

    Typical buyers of scanners quickly reduce the amount of scanning once the time warp honeymoon phase wears off. One can shoot a roll of 36exp in six seconds with a motor drive and spend an hour or two to scan the damn stuff.

    Scanners might be fun if in jail; or a desert island; or if you scan custom high end stuff. With massive scanning the time sink hole of scanning dozens of rolls takes its toll in time. Many times one can buy tools such as scanners cheap after the honeymoon phase; due to some folks have a life; or have to work.

    Folks still under the gas must attack the more practical dlsr route; since the film scan times are debunked as trivial.

    If you plan on a scanner route for 35mm get a real film scanner; not a flatbed. get one that will bulk feed and scan quickly. At some point the excitement of scanning wears off.

    At first the excitement of digging a ditch a football field long is real; you have a single shovel; your time is worthless; it doesnt matter if you have no life; you are saving massive money by not using a ditchwitch; ie modern tools. In practical terms here I shot a total eclipse of the moon with kodachrome in the 1970's; I shot about 2 dozen 36exp rolls. Transforming this into the digital domain is still ongoing; each slide has to be cleaned; custom scanned.

    In scanning "stuff" for the general Joe and Jane Public their input are not perfect; they are often a shoe box of crap; slides spilling out of boxes; all mixed up; faded; moldy; dusty; covered with cig. smoke. One can spend twice the time cleaning; vacuuming of the crap than scanning. Folks still under koolaid/gas think their "stuff" is perfect' when its often toxic crud that fouls the scanner; slides with mold growing on them.
     
  30. Stuart,

    Based on your comments, expectations and experience, I'm afraid you're going to be disappointed. What you're expecting just isn't IN the medium you're dealing with.

    Here's the bottom line. For real world shots, a good quality 4000dpi film scanner gets everything there is to get from 35mm film (everything that counts, that is - you can always scan more detailed grain down to microscopic levels and get more bits in the file, but there's just no point to it). Fine detail in digitally shot images simply ENDS where-ever the anti-aliasing filter ends it. Fine detail in film fades slowly into the matrix of grain/dye clouds. If you don't get grain in the scan, you're not getting all the detail. If you want more (eg, more detail with less grain - higher signal/noise), and you don't want to go digital, your only choice is going to be medium format.

    Of course, most would consider MF severe overkill for screen display. Screen display at any reasonable size just isn't a particularly demanding use for well scanned film, and any 35mm image (at least any ISO 400 or less) should appear crisp, sharp and grain free at any rational screen size, if properly processed.

    Given your dissatisfaction with what you're seeing, my suggestion to you is to try it for yourself, and see what YOU can get. Go to a reputable pro lab that still offers drum scanning, and have one of your images scanned. Of course, you'd be best served to start with slow transparency film, as it's ISO for ISO finer grained to begin with, and most drum scanning operations are tuned for transparency films not negatives. I'd recommend West Coast Imaging in CA. See their scanning web page at: http://www.westcoastimaging.com/wci/page/services/scan/scanning.htm

    A good drum scan at the resolutions you seem to feel you need will cost you about $25-$50 or so depending on lab, but it'll give you something to play with, and answer most of your questions once and for all.

    One final thought, and then I'm oughta here... You've already made some assumptions based on what you see in Les' full-res scans. I've been scanning and post-processing scanned film for years and, your PS experience notwithstanding, I'm afraid those assumptions are a bit off. People always over-react to the noise/grain of 100% views of scanned film. Some never get past dealing with it differently, and move quietly to digital where the PP task is far easier. Processing film scans IS different, and will require some time to master in order to pull out all that's there.

    In particular, you're going to need to re-think grain. It's always going to be there, but it's not as problematic as you seem to think. For fine-grained film at screen resolutions, you're going to need to think less about reducing it (though there are some good techniques for selective grain reduction with minimal impact on detail) and more about not emphasizing it in post (think masking, selective sharpening...). I typically do NO grain reduction on 35mm ISO 100 film scans, and get no visible grain on 13x19 prints. Screen sizes are easy by comparison.

    You're also going to need to think about proper down-sampling to screen size. With clean digital images, you can literally get away with just about anything, but with large film scans with both fine detail AND noise/grain, you can't get away with anything. Every digital shooter's favorite of "bicubic-sharper" is a disaster, leaving you with a 'scrunchy' looking mess. Heck, any form of bicubic is a disaster.

    Good luck,

    Scott
     
  31. If you have Nikon Lenses buy a Nikon D300 and save yourself lots of hassle and time. I used a dedicated slide scanner and a Minolta 700si for several years. I eventually got tired of the the scanning hassle and now have a Sony Alpha. I should have done this years ago.
     
  32. "Does the ISO speed of my film have a large impact on scan quality of the negatives?"

    Sure. Just look at the film data sheet. Look for the 50% point on the MTF curve. This gives a good idea of what resolution the specific film is capable of recording. In fact, it's enlightening to survey a broad spectrum of films. You'll see why an honest 4000dpi is about all you'll need from a scanner.

    Think of the scanner as the modern equivalent of the wet darkroom enlarger. There's no magic there. The best equipment only means only less degradation relative to the source image. The limitation really is the film itself, and there's not a whole lot of reason to use the 135 format today.
     
  33. Go film IF you are skilled or want to take the time to be skilled at color correcting the results. Even scanned, film has a wider color gamut and exposure latitude, and if done right can provide spectacular images.

    I'm a trained color darkroom printer and I still have trouble nailing the colors in post-processing.
     
  34. Been there and done that....the short answer is: Do not go the 35mm + scanner way. You'll spend more money and waaaaayyyy much more time getting decent results when compared with a medium priced DSLR (I would be thinking of the Pentax K20D with the II - New generation lens).
    Now - If you are looking to print 20" or more inch photos and you do not have 40k under your mattress, you may get a decent MF camera and an Epson V500 scanner coupled with Vuescan software. A 645 neg will get you up to 25" prints and a 6x7 wil almost double the size and quality.
    Now if you are just looking to display your photos on a PC screen... pick a P&S Digital. Anything will do the trick.
     
  35. A picture is worth a thousand words. Here's a link to a 11.7 Megapixel image derived from a scan of ISO 200 film with a Coolscan V:
    Scanned film image. Click the magnifying glass (and wait) to see at full resolution.
    I think this is comparable to a DSLR image. Now, to reduce the grain in this, I used Neat Image, which can produce a much more DSLR-like image than the raw scan. (Whether reducing the grain in this way is actually desirable is a matter for subjective assessment.) I think many comparisons of film versus digital are comparing apples and oranges, since the digital images are always digitally processed, but the film images may not be. If you want a DSLR-like image from film, you need to do DSLR-like processing.
    Some details: This was taken with Black's ISO 200 colour negative film (probably rebranded Fuji), a Minolta SRT-201, an MC Rokkor 85mm 1:1.7 lens, and bounced and off-camera flash.
    By the way, the unsharp images you saw from a Digital Rebel were probably unsharp for reasons other than its sensor resolution. An 8 Megapixel DSRL image isn't really 8 Megapixels, because of the Bayer array interpolation, but it should certainly be much better than 1 Megapixel, if other things are all fine (which often they aren't, of course).
     
  36. This thread has really gotten some attention! I've enjoyed reading it. I am an unabashed analog photography fan, even though I own a digital SLR, among many cameras. I prefer shooting film for its look, and so I went on a scanner hunt to try to get the best of both worlds. In my journey I learned consumer film scanners just don't cut it for image quality. I've shared the following example with many friends and photographers, and I think it is quite revealing. Just click here. There is another link on this page that shows the advantages of not only the Scitex EverSmart Pro II scanner I chose over consumer level scanners, but also the advantages of fluid mounting which I now do routinely.
     
  37. Marco, thanks for making that page on scanning resolution. What you say makes sense to me if you do not plan to do any digital post processing. However, if you are doing digital post processing then I think it may be better to scan a bit more into the grain and then remove the grain digitally. Of course, I can't quantify this without having a scanner to play around with.

    I can see from your chart that ISO film speed does have a very significant effect on grain size...just going from ISO 100 to ISO 200 loses 1/3 of the information. Since I don't have much experience shooting film I will have to research this aspect separately...I can see how

    ISO 100 film would be fine for daytime shooting of static scenes outside, but I'm not sure that it would work well for example...for shooting moving animals, or from a moving vehicle, or in low light conditions. Since I will only have 1 camera, I will have to determine the best compromise ahead of time.

    Les Sarile, you have mentioned that ICE is pretty important...yet looking at your comparison shot, it appears to me that Normal ICE is barely perceptible difference, and Fine ICE is significantly worse than No ICE. Neat Image, on the other hand, looks like a significant improvement -- but this is a photoshop plugin, so it doesn't effect scanner choice.

    I have made a little test of my own. I found a 4x5" print and scanned it in on my crappy old flatbed scanner. I cannot compare it exactly to what I've been seeing on here, but the quality is definitely better than the lower end film scanners...and looks comparable to the higher end film scanners to me. Since I cant afford a high end scanner, and its easy to get prints developed, and easy to look at them before I scan, and easy to scan...I'm pretty convinced now that's the option I should take.

    Scott, you say that bicubic sharper is a disaser for down sampling when you have grain. If you do a median filter first than it gets rid of the bad noise that messes up a bicubic sharper filter, and since you're downscaling it you don't really lose anything geometrically from doing it. Do you have a better option?
     
  38. ICE is for removing spots due to dust, etc. If you don't have any dust, then ICE is at best going to do nothing (and maybe makes things slightly worse). Of course, you probably do have dust...
     
  39. I have made a little test of my own. I found a 4x5" print and scanned it in on my crappy old flatbed scanner.
    The results of this may depend a lot on how this print was made. If it was an optical print, there might be a lot of information there. If it was made recently in a consumer photo lab, however, it was certainly done by scanning the negative and then printing digitally. Very likely, they scanned at no more than 300dpi in terms of the print size, for a resolution of 1200x1500, or 1.8 Megapixels. If so, no scanner is going to get more than 1.8 Megapixels of actual information out of the print.
     
  40. Scanning prints is an option if you don't have the original film, which is often the case for family albums. However the dynamic range on paper is extremely limited (2.5 stops) compared to film, and the resolution subject to many common faults. Minilab prints are in fact digital, printed at about 300 ppi.

    It's easy to be overwhelmed by the number of film choices, each with it's own fan club. The solution is to settle on two or three emulsions and learn how to work with them, rather than jumping on the latest rave. It's especially bad to look for bargain or outdated film - not that it's bad, rather that you can't get enough of it to work out the details. So choose a film you can get consistently and, if necessary, in quantity.

    The contrast and color of negative film is subjective. However, you can nearly always get good color balance and any contrast and saturation that you want in a scan. Besides availability and reliability, the most important consideration is grain, especially with 35mm film.

    I use Fuji Reala (100) with a tripod for landscapes and closeups for it's fine grain, and Fuji NPH400 for hand-held and flash shots, for its acceptible grain and easy to attain flesh tones. The third emulsion is Tri-X or TMax 100, because B&W is a lot of fun and easy to process. Your taste may vary, but you get the idea - simplify and explore in depth.

    No post-scan plugin can work as well as ICE. ICE, as implemented by Nikon, uses a 4th, infrared light to differentiate between objects in the image from dust or dirt. The dyes in most film are relatively transparent to IR, but dirt is opaque. ICE-Normal has only a slight effect on sharpness, especially compared to ICE-Fine (which I never use).

    I have both an LS-4000 and LS-8000. I find ICE is much more effective with the latter, due to its shallow DOF and softer light source. It is still necessary to do a touch-up using the healing tool (clone tool for tough cases) in Photoshop, which takes but a few minutes (<10).
     
  41. Stuart,

    Think logically about your strategy for scanning a print for a moment, and I believe you'll see the flaws. A print is a compromised second generation copy, and a copy of a copy can never approach working with the original. To make matters worse, as others point out, if it's a mini-lab print, it's printed digitally from a very mediocre scan itself to begin with. The dynamic range of the original is long gone, detail has been thrown away, and the color gamut has been significantly reduced.

    Again, you seem to be drawing some provably erroneous conclusions based on no direct experience. If you're forced by finances into taking a compromise approach well, hey - you do what you have to do. But don't kid yourself as to the results you'll get.

    Compared to well-made scans of the original film, scans from even good prints will look flat, one-dimensional and muddy. It's a strategy of last resort at best - something which would become painfully obvious to you once you actually started developing scanning & film post-processing skills. Painfully obvious

    As for down-sampling, I'd commend to you this article by Bart van der Wolf: http://www.xs4all.nl/~bvdwolf/main/foto/down_sample/down_sample.htm The big potential problem in down-sampling anything with high spatial frequency detail (whether desired or undesired) is aliasing. Yes, some sort of low-pass pre-filtering is required - the question is what kind and just as importantly, how much? Too much, and your down-sampled image is too smooth, too little, and it gets unpleasantly scrunchy. But the filtering method used in the down-sampling itself is also critical, and even with appropriate pre-filtering, bicubic does not distinguish itself. Over time, I've found I generally prefer Lanczos, and I also get best results using a package with a built-in pre anti-aliasing filter.

    As for grain reduction, I'll repeat my recommendation that, if PROPERLY TREATED, for almost any purpose, grain in ISO 100 level films is a non-issue. Don't over-react to 100% views of scans. If you didn't see the grain, you wouldn't have the details. You seem to want scans to look like digital images. Not only can you not expect 100% scan crops to look like digitally sourced images, YOU DON'T WANT THEM TO! If that IS what you want, then the answer really is to save yourself time and money, and go the fully digital route. Honestly.

    For higher ISO films, I'd recommend one of the better profile-based noise reduction packages. I've used Neat Image for years, and with experience, it CAN do a very good job if used carefully, particularly in combination with masking, and/or the history brush. Use globally and with default values, however, it easily creates Plasticville images. Lately, I've been moving from Neat Image to Noiseware Pro, another profile-based package. While the results possible are similar, Noiseware seems to be a bit easier/faster to get to quality results. Again, it's best used in its PS-plugin form rather than standalone, for flexibility in masking and selective reductions.

    Again, good luck to you.

    Scott
     
  42. Les Sarile,

    First of all, yes -- there is an overwhelming amount of information here :) But I'm doing my best to take it all in.

    My comment about ICE was with respect to this image here:
    http://www.fototime.com/4BA7AABC5FD205A/orig.jpg

    The scratches are removed, but at the expense of making the whole thing a lot blurrier. Those scratches could have been removed in other ways without making it so blurry.

    In your latest image of the black woman, the ICE does clearly remove dust and scratches -- but the image is too small to see how much blurriness it might be adding. Also, since
    you have not compared it directly to digital techniques for removing the dust and scratches, I can't say that it's any better than digital techniques.

    Edward says that "No post-scan plugin can work as well as ICE. ICE, as implemented by Nikon, uses a 4th, infrared light to differentiate between objects in the image from dust or dirt."
    It may be true that ICE has additional information that is not available for digital post processing, and that theoretically gives it an advantage over digital post processing. It doesn't mean that it DOES perform better than any digital post processing..

    It may be the case that a smart post processing algorithm can still manage to do a better job. Also, you have not considered that in post processing you have human intervention -- and a human can probably identify a scratch as well or better than ICE can. This gives an advantage to post processing that ICE doesn't have, because then you can use image inpainting techniques in small localized areas that are directed by a person.

    I don't see any improvement through GEM in the crayon picture. The Grain Surgery does look like an improvement. I think your image intending to show NeatImage was mis-linked.

    The image that I tested scanning in was probably produced optically, not digitally. That's what I would expect to have done again, if making prints. You guys sound very incredulous that this method would actually look good. It makes sense to me, though -- the limiting factor seems to be in the digital scanners, so by blowing it up optically first, it allows the scanners to do better. Yes, there's probably a loss of dynamic range from print -- but dynamic range is insignificant compared to spatial quality, which is much less perceptible and can be compensated for by playing around with contrast etc.

    Have a look for yourself, and tell me if you don't think it looks better than a lot of the digitally scanned negatives...

    Small:
    http://img383.imageshack.us/img383/9441/lapsmvj9.jpg

    Full size:
    http://www.mediafire.com/?dxoijluzopn
     
  43. "You guys sound very incredulous that this method would actually look good. It makes sense to me, though -- the limiting factor seems to be in the digital scanners, so by blowing it up optically first, it allows the scanners to do better."
    It's not a matter of being incredulous. What you're suggesting is both logically and provably fallacious.
    In any case, no, the limiting factor is NOT in the scanner, it's in the film to begin with. What's bothering you in looking at full-rez film scans is the fact that they really are what the film looks like at those magnifications. What you're doing now appeals to you because the process of printing has already done the post-processing work for you. You're seeing the finished product as opposed to a blow up of the negative as you see in the raw negative scans you're viewing. You're simply making a second generation copy of that finished product which reflects all of that work - unfortunately both the positives and the negatives.
    Give me the negative that print was made from originally and allow me to scan it, process it and make a print of the same size, and it'd absolutely blow your example out of the water for color, sharpness, detail, contrast and 3-dimensionality. This is the point several of us having been trying to make to you for some time now. The thing you're missing is that the full size scans you see ARE BLOWN UP, ACCURATE REFLECTIONS OF THE ORIGINAL NEGATIVE, warts and all, and that's the best place to start. It's the same thing that your print (which you then copied) started from. If that print had started from your scan, it wouldn't have looked nearly as good, (and, of course, neither would a scan of that subsequent print). That's the critical element. No one looks at 22Mpixel 100% crops, except to pixel peep. You'd have to look at a well-made print made from an original scan and compare it to your print in order to see the differences.
    "Have a look for yourself, and tell me if you don't think it looks better than a lot of the digitally scanned negatives..."
    Frankly, no, I don't. It's a nice image, BTW, but in technical terms, your digitized version is soft, lacking in fine detail, flat, and tonally blocked up. If I were to take one of Les' better images, for example, properly process it and print it, and then compare it, well...
    Look, I'm not trying to run down what you're doing here. If it works for you, and produces results that are easy for you to get and are acceptable to you, then you've got what you need. But there are far, far better ways of working if you're looking for top quality.
    Scott
     
  44. Les Sarile,

    The ICE does look a lot better on the African woman. I think that the reason it works on this picture and not on the map photo is because the image is blown up beyond the detail level of the original film grain, whereas the resolution of the scratches is not limited by film grain -- so a small amount of additional gaussian blur caused by the ICE is imperceptible and does not lose any true image information. This is not the case in the map photo, where the highest frequencies in the image content are similar to the frequency of the scratches...probably because you photographed it in such a controlled environment with perfect focus and perhaps lower ISO film. For this reason, I can see that ICE would be a very desirable property when scanning higher ISO film.

    The GEM is much more pronounced in this harbour shot..I see the difference in the crayola picture now, too..although it is less pronounced because it doesn't have such high contrast salty noise as the harbor one does in the sky.

    I would disagree when you say that the GEM is hardly significant at screen resolution but effective at 100% resolution. I think it is a pretty significant improvement at screen resolution, and LESS important when viewing at 100% resolution...at full resolution, the noise tends to blend into the image more naturally but its very obvious when you zoom out.

    Indeed, I cannot do a controlled comparison of a digital-scan-from-optical-print to digital-scan-from-negative without having a digital negative scanner...but there is such a big change in quality level, that I think I can safely say it's probably worse than the CoolScan but still blows the Epson V500 completely out of the water. This is a comparison you could surely do, though. Just make sure that you are using a print that was produced optically so there's no controversy there.

    "I would guess that the reduction from the original is perhaps as much as a GEM setting of 4?" I'm not quite sure what you're asking, and I don't actually know what GEM is, but I didn't use it. My particular scanner (Mikrotek ScanMaker 4700) is of very low quality, and it always adds a very obtrusive color gaussian noise to anything I scan. The actual print is so much better than this...there is no perceptible grain at all in the print, and contains a lot more detail than my scanner captured. In order to get rid of that terrible noise I used a median2 filter and then I just adjusted the levels a bit.

    Here was the raw scan output:
    http://img509.imageshack.us/img509/5337/postue5.jpg
     
  45. "In any case, no, the limiting factor is NOT in the scanner, it's in the film to begin with. What's bothering you in looking at full-rez film scans is the fact that they really are what the film looks like at those magnifications."

    Scott, you have just misunderstood me. I understand that the information content of the film presents an upper limit on the quality that can theoretically be achieved.

    I do not require images with greater detail than is theoretically attainable with 35mm film. I think I've already said that. My problem is that I cannot afford a digital scanner like the Coolscan that allows the full amount of information to be extracted from a 35mm film. As you can see in Sariles comparison here,

    http://www.fototime.com/037129A6767519B/orig.jpg

    Using a cheaper scanner like the Epson doesn't even come remotely close to getting the full information out...and that doesn't just mean its good for small prints, because honestly this scan is so terrible...I think the effects would be visible even if you shrunk it down 10x.

    So, the limiting factor is indeed scanner quality *for me* for price reasons, not for theoretical reasons.
     
  46. The median filter thing was just a quick hack...I will probably cook up some custom software specifically designed to remove this kind of noise from my scans, if I use that option.

    But I have another question...does B&W film of equal ISO have less grain than color film?
     
  47. Les Sarile--I look forward to your e-mail. The thread continues to focus on consumer-
    grade scanners which, as I've posted in my example link above, are remarkably inferior
    to professional-level scanners. Once you've seen this difference, the debate seems to
    fade away--at least for me!

    Take care--

    Kip
     
  48. "I do not require images with greater detail than is theoretically attainable with 35mm film. ... I cannot afford a digital scanner like the Coolscan that allows the full amount of information to be extracted from a 35mm film."

    A really nice option is to use larger format film. A cheap, film capable consumer flatbed has half the linear resolution of the Nikon. However, a frame of 6x7 MF is roughly double that of 35mm on a side. The flatbed/MF scans are at least the equal of, and almost always superior to that from the Nikon/35mm film combination.

    For a real example, look a the flower and vase crops near the end of this thread: http://www.photo.net/bboard/q-and-a-fetch-msg?msg_id=00MkgS

    Keep in mind that the price of film MF equipment has depreciated to almost nothing. For the amateur (and basically anyone not under time and production pressures) it doesn't get any better than this.
     
  49. Stuart,

    It's a little frustrating when an newbie argues authoritatively based on one thing he's read as opposed to another. I refer specifically to your opinions on Digital ICE, but there are other things as well. Until you have used a film scanner to the extent Les and I have, I suggest you listen a little more carefully. I'll shut up now and let you get to work while you still know everything.
     
  50. Les Sarile, I didn't take the photograph so I'm not totally sure (actually, the baby is me). It's a 3.5"x5.5" print on Fujicolor Crystal Archive photographic paper...so now that I think about it, it was probably a digital print made from an old slide that came from a 35mm film camera.

    Why would slides by finer than negatives in terms of grain? Aren't the slides made from negatives? Thanks for the comparison of B&W ISO's. I'm still not clear though, does B&W film contain less grain than equal ISO color film? The reasons I ask is because I thought color film might have to use different types of phosphorus or something, sort of like a TV screen, where you sacrifice spatial resolution for color depth.
     
  51. Robert, yeah..medium format film looks pretty cool. I've never done that before but it's something I may check out in the future. I don't think its a practical option for this trip, though, because I need something small that I can just throw around my neck and go hiking around with without being a burden.
     
  52. If you want to control grain then I would recommend medium format.

    Also, with Medium Format you can scan using a V500 and print quite nice
    8x10's and 11x14's. A nikon 9000 scanner is around 2 grand and that is without the mandatory 300 dollar glass holder. Unless you are doing a lot of printing for exhibition at medium print sizes like 11x14 or 16x20 then save for your cash for a professional drum scan. The nikon is sharper than the v500. However, at what cost? think of your needs carefully. How often and at what print size will make enlargements? big, medium or small size?

    I shoot medium format for maximum quality and have the option of large prints if needed. If I need a 30x40, I get a Drum Scan. IMHO I feel there is a big difference in massive print sizes between one of these scans and the Nikon. If the Nikon was a little more reasonably priced I would probably get one however even in this scenario I would still be sending out my best work to be drum scanned.

    If you plan to stick with 35mm then a Film Scanner is just about mandatory even for 8x10's.
     
  53. Stuart, from my experience with Nikon 5000 scanners, negative scans often show more grain or grain aliasing, that does slide film. Depends on film type, ISO etc.

    If you plan to scan negatives, expect to need to run the GEM (grain reduction) feature of Ice. Be aware GEM adds time to a scan and sometimes leaves swirl artifacts in its wake.
     
  54. There reason "consumer level scanners" are the topic of discussion on photo.net is that most all folks here are consumers who scan for fun; not as a way to make a living day in and day out. The scanner is consumed and considered a trendy tool thats typically not used long and upgraded based on emotion and not return on investment. In professional scanning for the public many times so called consumer grade scanners are used if they make economic sense. The typical publics input for scanning is not a clean room kept negative; but ones stored just average; or worse stuff thats a moldy mess stuck together in a shoebox that went under salt water in katrina. Crap like this can foul up a high end film scanner. Much of this stuff is so poor even a flatbed works well; plus the mold and dirt dont go into a delicate optical chain.<BR><BR>Here my excitement over film scanners has waned. We got our first high end scanner back in 1990 that ran under DOS with a GUI like window. Then we got B&W scanners that are 36" wide; and then color. We have been thru a dozen or two flatbeds and about 6 modern type 35mm scanners.<BR><BR>The general public doesnt shoot with tech pan; with a summicron at F8; with the camera on a granite block with a calbe release.:) There is alot of "stuff" that folks bring in for scanning that's just the common iso 800 35mm print film of soccer; sunsets etc. Sadly an old 2700 dpi film scanner is a vast overkill.<BR><BR>Here we have been scanning for the public now for 18 years. The bottom of the market went out years ago; some folks were sending the bulk/shoebox stuff to Iowa; then to Ireland; then to India. That cool internet deal with the lowest price of a few cents might be just sending it overseas. This has been done for almost a decade at some firms.<BR><BR>Either the film (ie the input) or the scanner can be the limit. All it takes is a tad of missfocus; a tad of movement and those best case resolution numbers are bunk. Nobody wants to think they stink; or their slides either. Thus the lay public often wants to Monday night quarterback a scan job. They may want many thousands of dpi; but their input is an old faded 126/Kodpak slide shot with a single element lens.
     
  55. Stuart: there is one other link you may wish to check out. It's an article by Tim Vitale and it's available from I think is the Stanford University website:

    http://aic.stanford.edu/sg/emg/library/pdf/vitale/2007-04-vitale-filmgrain_resolution.pdf

    It explains a lot about what film grain really is and scanner design. I found it very enlightening...

    Marco
     
  56. By the way, Canon CMOS DSLRs are very miserly with batteries. I used one standard capacity battery for about 5GB worth of photos on a week and a half trip (with some chimping my shots on the LCD in the evening).

    It's easy to buy a few extra large capacity batteries and such things as portable solar chargers are even available. Canisters of film took up much more room in my luggage.

    "For those who think scanning is a "massive sinkhole", perhaps the Noritsu 1700SA will serve you better? It's specs claim 6305 X 4181 pixels and at 16bit, has a capacity 415 frames per hour."

    Scanning speed is only one issue and not the major one in my experience. Touch up (dirt, grid, slide mounts, etc) and processing to restore the look of the original slide or to interpret the negative are. The raw output I have from an unprofiled LS-5000 is not something I'd consider a finished file but a good quality first step.
     
  57. You're welcome Les,

    Anyway, from the article it seems Kodak Portra 160NC (Neutral Color) might actually be Stuarts best choice. From all the common films listed in the articles, it has the best film speed at high resolution.

    It even rivals Fuji Velvia 50 ISO film, yet has 2/3 of a stop speed gain even compared to regular 100 ISO film.

    Stuart, if you would like to consider buying this professional film, just make sure you have it processed at a professional lab!

    I had a normal lab screw up two test films of 160NC, probably because they processed it as 100ISO. The film was hence underdeveloped and the colors and grain were severally screwed up.

    And one other remark:
    If you want the quality of 100 ISO film and still be able to shoot at low light conditions, why not consider a fast, fixed focal length, prime lens as an addition to any other lenses you may want to carry?

    From 100 to 400 ISO is just two stops, you can easily have the same gain if for example you buy a 1.7/50, 35 or 28 mm lens. Many inexpensive zoom lenses have at best 4.0 and sometimes even 5.6 as their biggest aperture. From 1.7 to 4.0 is a two and a halve stop gain. It will allow you to do much more in low light conditions, especially with short focal lengths as 35 or 28 mm.

    And lastly, how "bad" is film grain actually??? I have here a beautiful book of the world famous Sebastiao Salgado with some very grainy B&W shots of Africa printed at sizes of up 8x10 or larger, probably mostly originally shot at 400 ISO TriX. They are stunning...
     
  58. And one other remark: If you want the quality of 100 ISO film and still be able to shoot at low light conditions, why not consider a fast, fixed focal length, prime lens as an addition to any other lenses you may want to carry?

    Indeed, that was my conclusion too. I also wanted a fixed zoom lens so that it would protrude less and be more rugged. I've purchased a Nikkor 50mm f/1.8 D...it just arrived in the mail today :) Thank you for your suggestion about the film -- I will definitely give it a shot! It looks like this may be a pretty good compromise all around.

    Yeah, I suppose you could have some nice grainy shots...but I've never been touching up a shot and thought to myself, "gee...if only I had some more film grain." Of course, it's always possible to add grain, but you can never take it away.
     
  59. "Of course, it's always possible to add grain, but you can never take it away."

    You can actually. I've been a long time user of NeatImage. One of the best executed programs in this genre. Check their website and give the free trial a whirl.
     
  60. Gotta say, I love film, and dont have a DSLR, but what you are looking for IS a dslr. Sorry, better face it. If slides for projection or prints for viewing are not your end goal, but rather a digital image is, its a no brainer.
     

Share This Page