DSLR vs. 4000 dpi, what happens to all the Megabytes ?

[holger]

Heaving checked the threads and read the great comparison DSLR - 2900

dpi scanner, I am up to decide now which way to go. My brother has

bought the new Nikon V scanner, and results seem perfect. Problem is,

I have no way to compare similar pictures.

 

I shoot people, nature and landscapes, use both MF (Mamiya RB67) and

35 mm gear (Nikon F3), b/w and colour (Portra). Until now I got my

pics printed in a professional lab. With costs exceding my limits, I

would like to switch to either slides + 4000 dpi scanner (35 mm only,

MF stays b/w and pro lab) , or to DSLR and in both cases only get the

best pictures printed. What interests me most is picture quality, not

speed. Are prints (professional lab) of scanned slides comparable to

DSLR prints ? Most pictures I see in the german magazin

Naturfotografie are obviously scanned Provia and Sensia slides, and

they don't look bad at all.

 

The one thing that seemes strange to me is the difference in file

size: colour slide scanned at full 4000 dpi approx. 65 MB (TIFF),

Nikon D100 approx. 14 MB (TIFF, as stated in the manual). Is the

difference just dust and film grain ?

 

Why not switch to DSLR right away ? --> I love my F3, my manual

lenses, can't afford a DX1, feel safer with a neg or slide in my

archive than with an unvisible file on dvd or cd, love big

enlargements, and hate to depend on batteries.

 

Thank you very much for any kind of input, Holger

[daniel_taylor]

Before getting a Canon 10D I shot Fuji pro slide film, scanned the best shots on my film scanner, and printed to an Epson Photo printer. Prints from scanned slides can be excellent. I have a stack of sharp, beautiful 8x10's produced with this workflow.

 

How do they compare to images from my 10D? The 10D has a solid two stop exposure advantage over slide film, i.e. I can set my 10D to ISO 400 and still have less noise/grain than scanned Provia 100F. The 10D produces more accurate colors. And 10D prints appear slightly sharper due to the lack of noise and higher acutance. So I find my 10D outperforms slide film and scanner, but the quality of my old workflow was none-the-less very high. The two workflows are comparable in terms of final print quality, though the 10D does perform better. I also feel 10D images enlarge better.

 

Of course, the 10D workflow is turbocharged compared to film scanning. Don't underestimate this, you will shoot a lot more with a DSLR, and probably enjoy the work more.

 

Why the mismatch in megabytes?

 

1) With average contrast scenes a good slide film (i.e. Provia 100F) only records about 60 lpmm. That translates into roughly 10 MP regardless of scanner resolution, when you take into account that the extreme edges are generally cropped out of the scan. The real reason there are 4000 ppi scanners is to avoid grain aliasing which introduces more noise into the scan than there should be. There isn't actually 20 MP (again, those edges) worth of data on 35mm (high contrast test charts under strobes not withstanding).

 

2) Film's clarity drops significantly as resolution increases. Once you get around 50-60 lpmm things are very fuzzy on film. Details at that resolution add little or nothing to the photograph and get lost in the grain.

 

3) The scanner itself introduces degradation, again limiting the resolution of really clear and useful details.

 

4) A digital sensor records with tremendous clarity all the way to its resolution cut off point, and noise is virtually non-existent on a DSLR at low-to-medium ISO speeds.

 

5) The details that are important in most photographs are well within the resolution range of a 6 MP sensor.

 

That's why DSLR's are actually edging out 4,000 ppi film scans which produce much larger files. The extra data is pretty much "dust and film grain." More film grain really. You can quite frankly extract all the usable data with a 2700 ppi film scanner, but 2700 ppi doesn't quite render the grain shapes correctly which can result in aliasing and noise.

[beau 1664876222]

When you ask a question like this, the digital proselytizers and film reactionaries always jump in with prefabricated, emotional arguments.

 

From the preferences and facts you've stated, it sounds to me like you would be happier buying one of those new Nikon scanners and sticking with the cameras, lenses, and films you've gotten used to.

[l._nelson]

Actually even at 4000dpi (Coolscan 4000) I still get grain aliasing when scanning Provia (sometimes). At least that's what I think it is: it looks like noise in medium light, smooth toned areas that's not visible on the slide.

 

Also about the manual lenses, I recently started using a D100 with

my manual lenses, and it's a lot easy than I thought it would be,

even without the meter working. Basically I make sure to underexpose

a little to give myself working room, shoot NEFs and tweak it in photoshop.

 

Of course you could also have your manual lenses chipped by Roland Elliot, which will give you metering, and which I will probably be doing soon.

[neal_shields]

In objective scientifically valid, automated, quantifiable testing, film will resolve between 80 to 200 lp/mm mtf. Digital SLRs can't get close. (Note even 4000 dpi is only about 80 lp/mm best case) Many people don't believe in MTF testing but probably wouldn't buy a stereo system that couldn't demonstrate a measureable ability to hit the highs, just because the salesman said he thought it sounded good. Beyond that, if not MTF testing, what? Opinions?

 

However, for most photography you won't be able to tell the difference or care. If you don't use a tripod, cable release and very careful focusing at optimum aperture, you probably won't capture anything that can't be scanned at 2700 dpi.

 

However, I have made slides where signs that were over a mile away were legible at 6000 dpi and not legible at 4000 dpi.

 

One reason that I stick with film is that if I want the extra information that film captures, I have it. If I want to make a photograph taken with film look like it was taken with a digital camera, I can. In fact some of the new quick print machines will do it automatically, pixels and all when you get prints made.

 

If however, I leave information at, what I call, the scene of the crime when I make a photograph, and decide I want to crop it later, I may not have suficent information to do it. I have never been happy with any of the interpolation program's ability to create information.

[ky2]

<p>I used to shoot with a Fuji S2/Pro, which actually has better resolution and lower noise than the Canon 10D. I now shoot only film.</p>

 

<p>DSLRs are not for everybody. I found that my pricey DSLR turned out to be an expensive point&shoot. I made much better shots on my Fm3a, and that is why the DSLR had to go. Keeping something that expensive (with depreciating value) for snapshots was ridiculous. </p>

 

<p>Unlike most DSLR users, I enjoy film grain. Apart from the workflow (I hated spending hours in front of the computer sorting my shots) this is one of the reasons I couldn't adjust to DSLR snapshots. I find that grain brings much character my shots, especially in black and white photography. As for file size-- do you really care? it all depends how big you intend to print. </p>

 

<p>I never owned a 4000 DPI scanner. I did scan several medium format scans on a drum scanner, some on my dedicated film scanner, and some on my Epson 3200. They all print well. Just as my DSLR snaps. Enlarge them, and weird things happen. IMHO enlarged DSLR shots (12mpixel shots on the Fuji) turn out to be fuzzy and unreal. Enlarged 35mm b/w at ISO 400 enlarge beautifully- grain is evident, but I'm not trying to hide it-- its there for a reason. Maybe it is my eyes, but I like to see texture in film grain. </p>

 

<p>These are all subjective feelings. Please don't see this as a "Film-Is-Better-Than-Digital" flame. You pick your own tools. I had both to compare, and I like film much more. You may choose otherwise. </p>

 

<p>For a more scientific approach to scanned resolution vs. digital captures, you may want to <a href="http://www.clarkvision.com/imagedetail/scandetail.html#testarea1"><b>read this</b></a>.</p>

[frank_nikolajsen1]

Hi Holger,

 

<p>Here is a photo I snapped a while ago, while strolling around down town.<div></div>

[frank_nikolajsen1]

...and here is an excerpt from near the center, cropped and scaled up by a factor of

two to a total size of 222x222 pixels. As the original, shown above with scan

boundaries, warts and all, was scanned at 2820dpi, then this represents precisely

1x1mm on the film.

 

<ul><li>Scanner: Minolta DiMAGE Scan Dual III</li>

<li>Film: Fuji Sensia II 100</li>

</ul>

 

<p>I think I've hear about better film out there... You be the judge. Only image tweaks

after scanning was the adjustment of levels in PS. No sharpening.

 

<p>Remember that (all but 1?) of the digital cameras out there today make up

two/thirds of the information in the image. Their contrast suffers as a result.

 

<p>-- Frank.<div></div>

[neal_shields]

Much more constructive suggestion:

 

Go take the best 35mm picture that you can of a subject with diminishing detail. A long street scene with signs further and further away, a shopping center parking lot with lots of cars and license plates, etc.

 

Shoot with a good prime lens, at F8, focus very carefully at infinity, and if you use a tripod, use a cable release or self-timer to avoid shake. Bracket the exposure and use something like Fuji Velvia 50 or Realia 100.

 

When you get the negative or slide back, have it drum scanned at 12000 dpi with a photo multiplier tube scanner. (you may want to limit the scan to about ¼ of the slide to save money) It should cost you $20 or $30, but that is a pittance compared to what you will spend on a good digital camera.

 

Get them to give you a tiff file and take it into the computer and interpolate down, until you begin to see a loss of detail that you care about. (note look at both highlights and shadows)

 

Now you know how many pixels it takes to make you happy. If you go with a digital camera, you�re there. If you decide to go with a scanner you will have to add a fudge factor to compensate for the fact that a 4000 dpi scanner will not deliver the same quality as 4000 dpi interpolated from 12000 dpi.

 

If you don�t intend to use a tripod in your photography, disregard the above and go by a 6 meg digital camera and have fun.

 

Note: what you are probably seeing in magazines are from drum scans, worlds of difference in my tests.

[scott_warn]

Daniel has provided an excellent summary of the advantages to shooting directly with digital vs. scanning slides and producing digital prints from the slides. The only problem is that if you are used to analog prints made directly from slide film the analog process is still the quality leader. You should expect a slight loss in quality switching to a digital based process, most cannot see the difference unless you compare an extreme enlargement such as an 11 x 14 or 16 x 20 side by side, and even at that level of enlargement the differences are pretty minor.

 

My only concern about the digital process at this time isn't the resolution, it's the number of levels per color channel. With only 256 levels per channel I wonder if we don't loose some tonal rendition in the digital process. Proving out that concern would be difficult but I would be a lot more comfortable if they would increase the levels per channel to 1012 or even 2024.

 

As to your question about file size difference. I am rounding the numbers a bit, but at 4000 dpi a 35mm scan will yield a 6000 x 4000 pixel image. The D100 produces a 3000 x 2000 pixel image. Will the larger image file create a better picture? Maybe, as Daniel pointed out, there are a lot of factors in that 6000 x 4000 pixel image file that compromise the final quality. I would be inclined to call it a toss up with a slight tilt towards film at this point in time.

 

Whether you choose scanning versus just spending the money and shooting digital directly is a choice that should be determined by what you currently have. If you have a large stock of 35mm slides you are better off using the scanner because at some point you are going to want to convert those slides to a digital format. This means that no matter what you will end up buying a scanner. Also, digital cameras currently on the market have a problem shooting wide. The APS sized sensor used means you either buy an expensive wide angle zoom dedicated to that format or you just don't use the digital for wide angle shots. If you only shoot 35mm or wider perspectives and don't have a large stock of slides, then you may do better with the digital camera. They are VERY convenient and get more convenient every day.

 

As for which format has the better longevity, I believe that the digital format will out last film by a huge margin. Today's CD's may last for 1000 years, I am sure they will last 100 years. I am sure that the storage media and formats in 100 years will be much different that today's CD or DVD but if the files are copied/converted to the latest media/format at each evolution, the file will still be readable in 100 years. The big advantage to digital storage is that there is zero loss at each generation of copy. Film does fade over time, even Kodachrome, and there is not much we can do about that. Museums in 100 years will probably be able to read all those ancient medias. Remember each time media and storage formats evolve the legacy formats are carried over for a while. We can still read bmp files and those go back to MS-DOS, anyone remember what "copy c:\whatpic.bmp A:\whatpic.bmp" does?

 

I am waiting to convert to a DSLR. I payed over 1000 for a 2mp digital over 3 years ago and learned my lesson. Today a simlar camera can be had for 150.00. I plan on waiting until a full format 10-15mp camera breaks the 1500.00 price barrier, probably only about 2 or 3 years.

 

Now a completely unrelated question and totally off topic. I ask here because most photographers are also into music and am sure that there are some serious audio junkies who read this board. I have never found CD audio to match the quality of vinyl, I find listening to CD's tiring compared to vinyl. Is the new 196 mhz super audio DVD comparible in listenability to vinyl?

[vgoklani]

"The one thing that seemes strange to me is the difference in file size: colour slide scanned at full 4000 dpi approx. 65 MB (TIFF), Nikon D100 approx. 14 MB (TIFF, as stated in the manual). Is the difference just dust and film grain ? "

 

I don't know where you got 65M (TIFF) from, most of my scans are around 175M from a Nikon Coolscan LS-4000 (scanned at 4000DPI and 16bit) and 280M from a Minolta 5400 (5400DPI and 16bit). Keep in mind that scanning is very time consuming, as you have to worry about film curvature and where to set the focus point. There are other factors as well, you should do a search (look for posts by Charles Miller). I don't have any experience with DSLRs, so hopefully others can comment about the procedure necessary to process RAW files.

[frank_nikolajsen1]

<a href="http://www.cdfreaks.com/news/7751">Reality check</a> for Scott. Also note

that some very expensive studio digital TV cameras use a complex optical system with

three CCD chips in order to increase the dynamic range of the digital output. This is still

not enough to avoid burning out the areas around the highlights.

 

<p>I believe I've read that an <em>additional</em> 10 bits for a grand total of

around 18 bits of data per channel would be needed in order for digital cameras to

properly be able to simulate

the logarithmic response to light as found in film. The electronics within needs to be

linear to within the

same range as well, of course.

[ky2]

<p>To Scott:</p>

<p><i>"Today's CD's may last for 1000 years, I am sure they will last 100 years..."</i></p>

<p>Dunno. I got CDs and DVDs that don't have a year on them, and stopped playing. So far for durability. One of them is even a TDK armor media. Perhaps my DVD players' laser ray is a bit too powerful. I did however, scan 55 year old film last weekend.</p>

<p>As for Vinyl- it does sound better than CDs. Way better. Perhaps I am going retro :)</p>

[kelly_flanigan1]

Here I have had a few of my high 5 dollar each Gold Name brand archival CD's that have failed; after about 5 of 6 years; stored in perfect conditions.. Mayber the majority of the others will last 1000 years? :) who knows!. I wonder how long the "no name brand" ammo pack CD's will last; that most all buy today; "to save money".

[david_h._hartman]

Once you buy a DSLR(s) each shot is not quite free. A friend told me last night that DSLR(s) are coming with shutter problems due to very high number of cycles. If you can write the camera off comfortably as a business expense don�t worry beyond being backed up. Hobbyists on the other hand may want to shoot their DSLR(s) somewhat more like film to make them last. I understand that some cameras are coming in with blown shutter that aren�t even a year old. This can include professional models.

 

Some film cameras used to be driven high miles. This same friend put three shutters in an FM w/ MD-11 used by a photographer in Hollywood, California. He tried to get her to switch to an F3 w/ MD-4 but she didn't like the weight. The camera was used to shoot portfolios for hopefuls.

 

B&W film v. whatever: if you really want it to last conventional B&W film and prints will outlast it all. That�s not a reason to never shoot color but I�m quite sure many families will not have photos of great grandparents and some may not have photos of grandparents as children as previous generations did. I told my sister and cousins to shoot a roll of B&W of their kids each month and send it to me and I�d process it and proof it free. None (Zero) took me up on the offer.

 

Fortunately I have some preference for B&W over color, about 60/40. I still want a Nikon D2H. I�m just not willing to spend $3,200.00 plus spare batteries and memory on one. I put an extra big stocking by the fire place this year but I didn�t even get a lump of coal.

 

Best,

 

Dave.

[daniel_taylor]

Frank,

 

Attached is a 222 pixel section from an airshow shot I took, 10D, 70-200 f/4L, ISO 200. This would be a 28x42 inch enlargement. If Bayer color interpolation is causing any contrast loss or color errors, I would like to know where. I guarantee Provia would not have recorded that lettering with that kind of clarity. It would be there, but grain would be interfering at this level of enlargement.

[daniel_taylor]

Here's another 222 pixel block of a photo I took of my friend's cat. 10D, 28-105 f/3.5-4.5, ISO 400.

[daniel_taylor]

Just so you don't think I'm a total film bigot, here's a section from a recent 35mm flower shot. Rebel 2000, 28-105 f/3.5-4.5, tripod, remote release, Fuji Astia 100.

 

Now mind you, the whole shot looks great at 8x10. But it's got more noise than the ISO 400 cat shot, and is a bit softer. Compared to the plane...well, the plane shot has no discernable noise what-so-ever at 8x10.

 

Again, you can get excellent quality from a slide film/scanner workflow, and I highly recommend it to anyone shooting film. But I've got to give the edge to the DSLR's.

[daniel_taylor]

Neal -

 

"In objective scientifically valid, automated, quantifiable testing, film will resolve between 80 to 200 lp/mm mtf."

 

Yes, but film's ability to record detail changes with scene contrast. Film manufacturers will typically list two resolutions for their films. One is a 1000:1 contrast "objective scientifically valid, automated, quantifiable" resolution. The other is a much lower contrast "average scene" resolution, the one you can expect in your day to day shooting.

 

Velvia does 160 lpmm at 1000:1 contrast, 80 lpmm at average contrast. With an average scene, the contrast of those lpmm on the film is dropping off pretty rapidly by the time you pass 50 lpmm. Provia 100F does 140/60 lpmm, and its response has dropped off considerably by 40 lpmm in an average scene.

 

"Digital SLRs can't get close."

 

Actually the 10D sensor is rated at 65 lpmm. BUT it's an APS sized sensor so you will be enlarging more, i.e. fewer lp per inch on the final print than with a 35mm sized "sensor".

 

What a raw lpmm measurement misses, however, is signal to noise ratio. Back when I only shot film and DSLR's were just going mainstream (i.e. D30) I tended to disbelieve reviewers when they noted how important this was. (I was positively irritated with Reichman for daring to suggest a 3 MP D30 could compete with my precious Provia!) Now that I have a DSLR in hand and can do comparisons, I'm seeing first hand just how important it is.

 

With film you have a sloping curve. The higher the lpmm, the lower the contrast until you can no longer distinguish line pairs because they're too fuzzy. With a DSLR sensor you have a very sharp drop off point where the line pairs turn grey. But right up until that point they are recorded with extreme clarity. High contrast with virtually no noise. A very, very shallow slope on your response curve. (In fact, I wonder if it would be a flat line with a theoretically "perfect" lens.) I'm finding I prefer the digital curve.

 

I don't doubt that some prefer the film look, nor fault them for that choice. Heck, I'm keeping 35mm gear around for Velvia 50. How do you reproduce that magical film in Photoshop? There's also a point to be made for B&W and films like Portra (more exposure latitude than a DSLR), though I personally never cared for the lower contrast films and just dabbled in B&W.

 

But the measurements and theories which make 35mm film sound vastly superior aren't reflecting the real world. Sometimes you'll find a little more detail on 35mm film, but more often the shear clarity of a DSLR will win in a comparison.

 

Again, I don't care what someone chooses to shoot, and I think the orginal poster will be well served by a slide film/scanner workflow if he doesn't want to add digital to the mix yet. Like I said, I have a stack of beautiful 8x10's from my film scanner.

[frank_nikolajsen1]

Daniel,

 

<p>You may have missed it, but the cutout I presented was digitally scaled up in size

to be easier to see. The original was only 111x111 pixels. 2820dpi/25.4mm/inch = 111

'dpm'.

 

<p>If you would like to present an equivalent view from the 10D, then you need to

show us an excerpt, which is an 1/36 of the image sensor/frame width. For a 10D that

would mean 85x85 pixels.

 

<p>Also be aware, that the scan shown was not done under perfect conditions. The

slide wasn't mounted, and the DiMAGE film strip holder is not the best, so at this scale

the scan is bound to be somewhat out of focus. Additionally the very nature of the

subject means that this particular part is probably not in sharp focus either. I think it is

reasonably close for the section I choose though.

 

<p>Anyone aware of a proper mathematical treatment of precisely what is possible to

'predict' using Bayer color interpolation? As far as I am able to see, then it is impossible

to properly reproduce a step function at an arbitrary angle, unless the algorithm knows

what we are looking at.

 

<p>-- Frank.

[daniel_taylor]

Frank,

 

"If you would like to present an equivalent view from the 10D, then you need to show us an excerpt, which is an 1/36 of the image sensor/frame width. For a 10D that would mean 85x85 pixels."

 

OK...attached are two 85 pixel clips from the cat shot, scaled 2x. They appear a little dark on the web to me, the original image is brighter. We're now at a 56x84 inch enlargement from an APS sized chip at ISO 400! We're peering into the very structure of the image, the pixels themselves at 2x, and the thing still looks OK! Individual hairs are easy to see (mouth clip). The shapes in the cat's eye, including the window reflection, are still pretty well defined given the extreme enlargement size. There isn't any real noise to speak of even at this extreme enlargement. The "grain" you see is caused by the pixel structure itself.

 

There is no doubt in my mind that at this view Provia 100F (two stops slower) would show much more noise and softer details. Individual hairs would still be visible on the mouth, but I doubt the shapes and tones of the eye would be as clear. They would be intermixed with grain. And ISO 400 film? Forget it...

 

I'm at the disadvantage of not being able to show identical shots side-by-side. This site has several such shots:

http://www.sphoto.com/techinfo/dslrvsfilm.htm

 

I originally intended to do a test series of identical 35mm/10D photos when my 10D arrived, but the first prints from my printer made me realize it wasn't worth my time. I know the best I can do with scanned Provia, Velvia, Astia, etc. My 10D surpassed that level from day one. (Well, Velvia has a unique look. But otherwise, the 10D out performs that to. That grain again.)

 

Again, high quality 35mm prints are clearly obtainable using a film/scanner workflow. If you like your film body and don't want to go digital, or if you like the look of certain films (B&W, Velvia, etc.), by all means, shoot and scan your heart out. But you're kidding yourself if you think a DSLR can't meet or beat 35mm film because of a numbers game involving file sizes. They have arrived.

 

"Also be aware, that the scan shown was not done under perfect conditions."

 

Yet another reason I'm so fond of the digital workflow. Perfect processing, perfect film handling, perfect scanner focus, and perfect film flatness are not easy to do. Dust is always an issue. And I've lost shots to poor handling and scratches/blemishes. None of that is a concern with a DSLR.

 

"Anyone aware of a proper mathematical treatment of precisely what is possible to 'predict' using Bayer color interpolation? As far as I am able to see, then it is impossible to properly reproduce a step function at an arbitrary angle, unless the algorithm knows what we are looking at."

 

Any real world examples of Bayer failing on a modern DSLR? From everything I've seen, the system is ironed out, and it works fine. It worked fine on early digicams to for the most part. Even with those it was rare that you would see a color photo with fine patterns that would "fool" Bayer and cause some aliasing. Shouldn't be surprising since the human eye uses a very similar system of "color interpolation", and it's the most advanced camera in the world.

 

I don't doubt that some creative mind somewhere could set up a precise test which fooled Bayer and exhibited aliasing. I also don't doubt that this test would bear no relation to real world shooting, just like the 1000:1 resolution tests of film.

[arnabdas]

Whenever I see people splitting hairs over comparing a generation 1 digital photograph with a generation 2 digital photograph (typically shot with a desktop scanner) of a film photograph -- I feel so glad I still shoot film.

[frank_nikolajsen1]

Thanks Arnab, I was beginning to wonder how long it would take. ;-)

 

<p>Daniel,

 

<p>The point is that I was using a long lens, hand held while shooting a moving

subject on consumer grade slide film, scanned on an inexpensive desktop scanner,

which I knew was probably out of focus to begin with.

 

<p>Whenever the results from digital matches my Ilfochromes, then I'll switch. In the

meantime I'll just rescan my slides as the technology improves.

[ky2]

LOL. a 56x84" print from a 6mpixel CCD? Sounds much like Michael Richmann and his D30 climax :). The 1Ds tops at about 20x30", and in my opinion, still loses to any 645 kit that shoots b/w. With color, its very close.

[daniel_taylor]

Frank,

 

"Thanks Arnab, I was beginning to wonder how long it would take. ;-)"

 

Note I wasn't the one to start the hair splitting...

 

"The point is that I was using a long lens, hand held while shooting a moving subject on consumer grade slide film, scanned on an inexpensive desktop scanner, which I knew was probably out of focus to begin with."

 

The cat shot was 168mm equivalent with a consumer grade lens (wide open even). Cat wasn't moving but the plane was when I shot it at 320mm equivalent (L glass this time). Both were hand held. The plane was shot at ISO 200 in JPEG mode. The cat at ISO 400 in RAW. Though these images weren't originally intended for a "film/digital" comparison, had they been I'm not sure how I could have given a greater advantage to ISO 100 slide film. Should I choose a lower JPEG quality and ISO 800? What if I had shot that cat with the L glass, ISO 100, RAW?

 

BTW, "consumer" Sensia is an identical emulsion to "pro" Astia 100 (not 100F), it's just not aged as long. Provia 100F is better, but not by any large margin. Velvia 50 is sharper, but has more grain. A shot on Sensia is already on one of the absolute best, most capable color films in existence. And if an "inexpensive desktop scanner" can't compete, where do we turn? At what point do all the "film is vastly superior" theories pan out? $30 drum scans? Thanks but no thanks. At that cost per shot I'll drag out a 4x5 rig.

 

Like I said to the original poster, film scanners make beautiful enlargements from 35mm at the sizes one would normally enlarge 35mm to. But there's no inherent superiority in shooting film over a DSLR. All this information that 4,000 ppi scanners are supposedly pulling out of film is really just grain.

 

I think the point is you posted an extreme film enlargement (Arnab's idea of hair splitting) assuming nothing from a DSLR could compare, not at that size with only 6 MP to play with. The fact that you're now back peddling shows you're having second thoughts. And you should. DSLR images hold up very well even at extreme enlargements. The sensors are that good, Bayer and all, and a lot of what degrades a film image is gone from the workflow.

 

"Whenever the results from digital matches my Ilfochromes, then I'll switch. In the meantime I'll just rescan my slides as the technology improves."

 

If you live in southern CA I'll be happy to meet with you and compare your best Ilfochrome print to my best digital print, judging purely on the basis of print quality (sharpness, grain, color, etc.).

 

Yaron,

 

"LOL. a 56x84" print from a 6mpixel CCD? Sounds much like Michael Richmann and his D30 climax :)"

 

I wouldn't recommend trying to print a 56x84 from a DSLR. I was simply illustrating the extreme nature of the enlargements in my hair splitting contest with Frank.

 

However, I would not hesitate to go 16x24 from a 10D. GOOD 16x24 from 35mm is really hard. I have no question that a 16x24 from my 10D would beat a 35mm 16x24 slide enlargement via any workflow.

 

Again, I don't give a cat's wisker what anybody shoots. But the "35mm film is superior theories" are just plain wrong. Which is why DSLR's are competing so well despite the smaller file sizes.