8MP DSLRs vs ISO 100 film

[Curt]

Well, I use a Olympus 8080 & two DSLR's. I would not again buy the 8080. It is a fine camera but no match for a D70 or Fuji S2. Fuji's 400 ISO is probably superior to the 8080's 50 ISO. This is not based on scientific testing, only on what I see on prints.

[pfigen]

There's more to "quality" than just the presence or lack of "grain". Considering that a 1Ds just barely beats out 35mm Velvia for resolvable detail, according to actual tests I've done with the same lens, tripod and aperture, I have a hard time imagining that a MkII would be better than that. You really have to consider the quality of the scanner when comparing film to digital. There can be a huge difference and most scanner, even some reputable drum scanners, can't capture everything on the film.

[BJL]

There's more to "quality" than just the presence or lack of "grain". Considering that a 1Ds just barely beats out 35mm Velvia for resolvable detail, according to actual tests I've done with the same lens, tripod and aperture, I have a hard time imagining that a MkII would be better than that. You really have to consider the quality of the scanner when comparing film to digital.

I did not suggest that noise and grain are the only important factors in print quality, and indeed I believe think that dynamic range is a more pressing concern than noise. But one thing at a time: this particular question happens to be about noise and grain.

The 1Ds Mk II and Velvia were not on my list, I am interested in 100 speed films and 8MP cameras instead. (This is because I have never used Velvia, feel no great need for more resolution than good 100 speed 35mm films can provide, and do not see any 11MP plus camera in my near future.)

Scanning also has little to do with my question. I am interested in how prints look, not how crops of small fractions of an image look when displayed at extreme magnification on a low resolution output device like a computer screen (i.e. "100% pixel viewing")

[BJL]

This might seem like an old topic, but the recent wave of 8MP cameras in a wide variety of sizes makes me curious again. I have seen arguments that 8MP about matches the resolution of good ISO 100  35mm films like Provia, and if that parity holds, it might be interesting to compare sensitivity, meaning exposure index capability.

So, to all users of the 1DMkII, 20D, or any of those 8MP 2/3" format digicams, how high do you find that you can go with the ISO setting and still have noise that is no more visible or objectionable than the grain of a good ISO 100 film?


P. S. To my mind, the answer to this question indicates something like the true "ISO speed" of the camera; any higher, noiser setting is definitely akin to push processing.

[BJL]

Curt,
   I am looking exactly for comparisons based on what people see in prints! I asked partly because I still have no good feel for what various noise level test numbers tell me. Working with film did not teach me what a 43bB S/N ratio looks like.

Can you compare to 100 speed film as well as between digital cameras? Are you hinting that 400 speed on the S2 is of "100 speed film quality" as far as lack of annoying noise/grain?

I am fairly sure that 8MP DSLRs will be able to match Provia100 quality at higher than 100 speed, and at higher speed than the 8MP digicams, but I am curious about more specific, quantitative comparisons to film.

[didger]

There's more to "quality" than just the presence or lack of "grain". Considering that a 1Ds just barely beats out 35mm Velvia for resolvable detail, according to actual tests I've done with the same lens, tripod and aperture

I'm not sure how you define "resolvable detail", but my personal experience with having shot and scanned thousands of Velvia slides did not show me that 35mm Velvia is anywhere near as good as 1ds.  True, there's more to quality than the presence or lack of grain, but for my taste, if grain is too conspicuously visible, then any other merits of the image don't count.  I found that my 8 Mpixel Velvia scans were generally way too grainy at 100% view.  1ds at 100% is consistently vastly better than this.  For a fair comparison you'd have to uprezz the Velvia scans to 11 Mpixels also and then they'd look even worse.  I scanned with a Nikon Coolscan, which by present standards is not the greatest, and a better scanner would no doubt show less noise and better color fidelity, but the grain would still be there.  Based on my personal experience with 35mm Velvia and 1ds, I definitely agree with the general consensus that 1ds has about twice the real life practical resolution that any sort of 35mm film does.  There's got to be a good reason why Michael dumped his MF film gear soon after he got to test a 1ds.

[Jonathan Wienke]

If you're comparing prints, then 10D RAWs, properly upsized and sharpened, will exhibit better color, clarity, and detail than scanned 35mm film in most cases. A 1Ds is about twice as good as 36mm film, and the 1D-MkII falls in the middle, unless the ISO is above 400, in which case its low noise levels give it enough of an advantage to overcome even the 1Ds.