BJP report on scanning film v digital

[Philip Freedman]

The latest British Journal of Photography reports that a scan made on a Fuji

Frontier 570 of a 35mm Kodak Ultra 200UC negative taken on "an ageing Nikon

FM" produced a photograph that has "much higher contrast and sharpness" than

a digital image taken on a Kodak DCS760 6 megapixel digital camera. The

magazine reproduces the two images and there is a noticeable difference. Up to

then, I had believed many reports that scans of 35mm negatives taken on film

faster than ISO 50 were generally poorer than 6 mp digital images.

 

Sounds as though there is still use for my M6.

 

Any thoughts?

 

Philip

[mike dixon]

I'm thinking: Why are they making comparisons to a five-year-old digital camera?

[Ken Katz]

What are their results using a D70, or a D200, or a Dx2 using the same glass as they used on the FM?

[Spearhead]

They should do a re-test with the Kodak digital camera and a Lomo. I think they might get different, but just as worthless, results.

[erik_ohlson1]

I have found just the opposite - even re-sampled 3MP pictures trump any 35mm neg or slide I have, from a whole range of 35mm cameras from Leica thru Olympus OM-2s SLR, and a variety of lenses.

 

This is based on actual print tests, not guesswork.

 

Possibly it's my technique, but you'd think that sometime over the last 50 years I'd have accidentally gotten a sharp 35mm, if only by accident.

[pics]

50 years and nary a sharp 35mm? I'd say something must have gone awry along the way lol. I've toyed around with a 3MP Canon and never got anything that approached the quality of a 35mm film frame. So much of it has to do with the skill of the operator and I don't think you can make a blanket statement equating the quality of a 35mm frame to a certain number of pixels. I've gotten medium format scans from a Frontier that were horrendous and of significantly lower quality than what I could get on my home flatbed. Using them as a guideline, it would look as if a 3mp digi point and shoot had reached the level of MF film.

[dickhilker]

I've made 16 X 20 enlargements of pictures taken by experienced photographers with new Nikons and Rebels on an Epson 7800. Compared with current Fuji and 40 year 0ld Kodachrome slides taken with Zeiss lenses and scanned on an old Minolta Dimage Scan Multi II, the digital images are less sharp and the colors are anything but subtle. I'd love to make the switch to digital, but think I'll wait until the camera makers stop claiming their latest, greatest gear is "nearly as good" as 35mm.

[Spearhead]

That sounds scientific.

[dickhilker]

"That sounds scientific"

 

Not in the least, Jeff -- just an artistic appraisal. I'm no technician, just an artist who works with a camera instead of a brush.

 

For my own work, subtlety and the ability to make large, sharp images are most important. If I were covering sports, I'd do exactly what you're doing and go for the flexibility, speed and low operational cost of digital.

[Spearhead]

We can all do "artistic appraisal," but that's not very useful for drawing conclusions for anyone but yourself. The article referenced is obviously supposed to prove something and is just as obviously a flawed test.

 

Sports is a tiny bit of what I do. I have printed digital, 35mm and medium format images up to 20x24. At 13x19, the digital (from a 10D) is significantly better (in my "artistic appraisal") than the 35mm and slightly inferior (in my "artistic appraisal") than medium format.

 

But that's all meaningless to anyone else. I have specific ways of processing things, and I assume others do, and without some type of standard, or at least rigorour, test, it's not necessarily appropriate to what anyone else does.

[bobatkins]

The CCD sensor in the DCS760 was developed a decade ago and used in several of Kodaks early DSLR attempts. It has an upper limit of ISO 400. I suspect that maybe if they'd tried more recent equipment they might have gotten different results.

[Spearhead]

The Canon 20D "enhanced" looks a whole lot better than most of the film scans. What is this supposed to prove?

[eugene_scherba]

Yeah, my 20D beats them all except MF.

[erik scanhancer]

Although I am generally voting in favor of scanned film, this BJP test doesn't make sense.

Higher contrast and "sharpness" are easily to obtain by electronic post-production. And

the Fuji Frontiers really make low quality scans, believe me. What about resolution, color

gamut, exposure latitude, noise (or grain) and gradient smoothness? Those are the real

values to look after if you want to compare digital capturing to film. Everything else can be

dealt with electronically after the shot has been made.

 

Having done many tests myself I am still convinced that -if properly executed!- film can

beat digital in most cases. But choices have to be made really well. Negative or slide? How

to scan? How to convert? Using what color space? How to sharpen? If you know the answer

to these questions in combination with the circumstances to be photographed under, nine

out of ten times film will still rule. (And oh yes, there is still a lot of use for an M6!)

[marbing]

I have a digital camera and I have a film scanner. The one thing that I have found is that, when I post pictures on this and other websites, nobody ever says "Hey, that looks scanned!" It appears that, unless they are specifically alerted to it, nobody can tell whether an image was directly digital or scanned to digital. However, if you tell people that this is a scanned shot, then all sorts of techno-babblers will come out of the woodwork to explain why the shot isn't as good as the hot 'dSLR du jour'.

 

I once, on another site, put up two images for comparison. One was scanned and the other was digital but I didn't say which was which. To be sneaky, I intentionally renamed the scanned image to DSC5678.jpg and renamed the digital image to Scan096.jpg (or something). Those who peeked at the properties of the link saw the bogus names; assumed the digital file was the scanned one and they tore it apart! This was bad. That was bad.

 

Now the bogus digital file, it had all of the superior hallmarks...it had better contrast, sharpness..everything.

 

However, after I revealed my little joke, THEN people saw the 'defects' in the other shot. So, in the end, most of the negative comments about scanned images have more political motives than technical ones behind them. It was more important to slam the film option than to objectively evaluate the image as a photograph.

 

If people evaluate the image for what it is and comment on the content of the shot, this is much more sensible than speculating on the various processes involved in getting the shot onto the screen.

 

I expect that if a good shot is taken, it shouldn't make any difference whether it started as a data or as a film frame. Is it a good picture? Yes or no?

[pics]

Interesting thoughts all around. I think that what defines image quality for most people (from a technical standpoint) is the apparent sharpness of the final print. Resolution isn't nearly as big of a factor as some people make it out to be. The visual sharpness of today's digital images is a result of post process sharpening in photoshop and has little to do with the resolving capabilities of the sensor itself. The same thing can be done digitally with film after scanning, or through traditional means in the darkroom (unsharp mask.)

Both digital and darkroom sharpening destroy the resolution of the file or film image but make for a very sharp looking print. The difference is that with digital, the shooter can more easily control the aspects of the image whereas with film it takes quite a bit of practice in the darkroom or if you let a lab do the work you are at the mercy of someone elses skills. I'm willing to bet that many people over the years have never experienced the full capabilities of what film can deliver beyond what the local lab does for them. Once they go digital and have control of these aspects of image quality, the conclusion is that film can't do what digital is currently doing for them based off of all the mediocre drug store prints they have accumulated over the years. In terms of the inherent resolving capabilities of film vs. digital, I've yet to see an unbiased test on both sides that shows one undeniably trumping the other and I don't think such a test is a good indictator of how to arrive at the best possible print, which happens to be my goal. The generally accepted rule nowaday's seems to be that 6mp outresolves 35mm in most cases. Then again, when Zeiss recently tested their newly introduced lenses which are supposedly the highest resolving products on the market, they used film and stated that no digital sensor was capable of showing what the lenses were capable of. In the end I think workflow issues should be the overwhelmig factor in choosing digital or film. You can make images of equally high quality with either method.

[dickhilker]

As with most jobs, it's important to pick the right tool for the task: for landscapes, MF or LF will generally yield the highest quality for large-scale images. For posting on the web, weddings, sports, almost any digital will produce adequate quality with the added bonus of "bigger than life" color and low operating cost.

 

In reality, there's no single best format, only the most appropriate for the circumstances. Somehow, I'm more inclined to put faith in what Zeiss says than the claims of those who've just invested big bucks in the latest digital gear: but, if the digital does the job at hand, what difference does Zeiss' opinion make?

[Spearhead]

<i>Somehow, I'm more inclined to put faith in what Zeiss says</i><p>

 

"Someone from Zeiss" is paid to spout the company line. I can't see why anyone would take their statements with faith given that they are pushing certain products. I tend to ignore pretty much any company spokesman on marketing-type issues, which this obviously is.

[colin_thomas5]

Someone mentioned that the digi-cam was 5 years old, well, the Frontier's scanner has been put down by a certain digi-photo expert on here as "inferior to home film scanners", so maybe this is a handicapped test on both sides. Regardless of the outcome of that article, I'm happy with my film results.

[pics]

"Someone from Zeiss" is paid to spout the company line. I can't see why anyone would take their statements with faith given that they are pushing certain products. I tend to ignore pretty much any company spokesman on marketing-type issues, which this obviously is."

 

Zeiss makes lenses not film. Their lenses are also used on numerous digital cameras as well as film models, so I'm not sure what the "company line" would be in this regard. I'd be more inclined to doubt a company like Canon considering they have much more of a vested financial interest in touting the abilities of their digital sensors. Why would Zeiss purposely use film to test their products if it showed the lenses performing at a lower level than they are capable of? If I was in the business of hawking high quality optics, I would use whatever medium that made them perform best.

[Barry Thomas]

Hey Meryl, reading your post one might think you were all for the shot and not remotely a gear-head, but your snaps on p.net are nearly all grouped by camera. Some kind of a disjoint there I reckon.

[dickhilker]

Well said, Daniel!

[bljkasfdljkasfdljskfa]

It all comes down to the results. If you like the results of a particular methodology and equipment, then shut your trap and use the said methodology and equipment.