Can DSLR compete with Velvia?

[taka_anzai]

I was sceptical about digital still camera. Very.

Because I was not sure that digital can offer same color reproduction

of Velvia, which I have been using for over decade.

I bought Nikon D100 at last, and then, tested with Velvia.

I uploaded my test result on web, so that you guys can see it, if you

want.

 

http://www.hi-ho.ne.jp/taka_anzai/photo/

 

I'm landscape photographer and would appreciate if you guys can tell

your experiences with digital.

[peza]

Taka - very comperhensive test, however, there is something what I'm not sure about. Detail of 6MPX DSLR simply could not be and it is not better than scanned Velvia - it was probably scanned on subpar scanner or misfocused during picture taking or during scanning or all of them.<br><br>

I'm 10D owner and did simillar tests myself. Velvia holds (far) more detail, if focused properly.

[steve_chan5]

You should realize that Taka cropped the Velvia image so that the angle of view would be comparable. As a result, the Velvia pictures are only about 1/2 of the total film available.

 

A full frame digital at the same focal length would be a 12mp image. Is scanned Velvia going to easily beat a 12mp image? Even at 4000dpi? Maybe not.

 

Velvia would have clearly come out ahead if he had allowed the focal length vary so that the same scene were framed.

[gsbhasin]

Steve,

<p>Taka cropped Velvia shots to get the same frame. This has to be done since the D100 has sensor size 1.6 times smaller. Alternatively if he had used a lens with focal length x on D100 and another with focal length 1.6x on F100 he would have gotten similar results.

<p>I dont see why Velvia should result double the resolution of D100, at least based on your reasoning..

<p>Regards,

<br>Gurpreet

[ellis_vener_photography]

actually the CCD imaging chip in the D100 is .66x the times of a 24x36mm frame of

film. Which means there is a virtual 1.5x focal length increase. This virtual focal

length increase only affects angle of view but does not affect other optical

characteristics such as depth of field. Perhaps a fairer test would b to make the D100

images at 17mm and the F100 images at 25 -26mm for angle of view considerations

but of course that will change depth of field for any given f-stop and possibly

performanace characteristics of the lens. So either crop the scan or zoom the focal

length -- either way you are going to get compromise.

 

A scan is of course a second generation reproduction of the original image. this will

be true no matter how good the scanner and scan is. A "desktop" scanner is not as

precise an instrument as a drum scanner that is used well. Unlesss the photographer

manipulates the digital image are essentially first generation.

[steve_chan5]

The D100 sensor is 2/3 the size of a 35mm frame on each dimension. When you look at the total area, you have 4/9 of the overall area.

 

So Taka is throwing away about 5/9 of the film when he crops it down to the same area as the D100 sensor to make the comparison. The comparison is fair in regards to color and matching the bokeh, etc... But it is not fair in regards to how much resolution Velvia is capable of vs. D100 when framing the same scene.

[gsbhasin]

Steve,

<p>You might consider reading on FOV. D100 and 35mm film can be compared in resolution for similar FOV. Please read on FOV.

<p>They need to see the SAME SCENE. Hence focal length 1.5x on 35mm and x on D100 will result in same image captured. My earlier 1.6 figure was incorrect. But cropping Velvia is a good approximate thought not accurate comparison. I hope it is clear that just because a lens on 35mm covers more field of view it makes the 35mm film more resolving, is an absurd reasoning.

[steve_chan5]

Gupreet,

 

I think we are talking at cross purposes. The problem is obviously the 12mp number I casually mentioned.

 

My point is that comparing the detail visible in the Velvia crop to the detail in the D100 is not "apples to apples".

 

To me, aside from geeky pedantry, the resolving power per unit area of Velvia vs. D100 is not really that interesting - what is interesting is the level of detail that a frame of Velvia can capture vs. the level of detail a D100 frame can capture for the same scene.

 

On Taka's site, he mentions that they are roughly comparable in terms of the resolved detail (with film being actually somewhat better in terms of detail. but the D100 looking better sometimes), and Peter says that cannot possibly be true since he gets much better detail from Velvia. I toss out the 12mp number because you are comparing half a frame of Velvia against a full frame of D100.

If the D100 sensor were full frame, it would be about 12mp and it would capture the same scene as the full Velvia frame.

 

At that point, the Velvia scan and the D100 images could be compared on a pixel vs. pixel basis. Would the Velvia still clearly beat digital? Looking at Taka's scans, my opinion is "Maybe not".

 

In fact, I'd say that based on Taka's excellent comparison, the Kodak DCS-14n is, for practical purposes, the equal of 35mm Velvia for resolution, and maybe better in terms of the overall image quality returned. Take a look at this photo from a Kodak user: http://www.pbase.com/image/22874478/original

 

There seem to be several landscape photographers who pop up on DPReview to say how happy they are with the Kodak's (so long as low ISO ratings are being used).

 

A good "cheapskate" strategy for a landscape photographer with Nikon lenses is probably to wait until Nikon comes out with a full frame DSLR, and then buy the Kodaks at fire sale prices when people abandon the perfectly functional "old new thing" for the incrementally improved "new new thing".

[vatovec]

The discussion is heading in a strange direction. For me it is funny to compare a ccd with a cropped piece of film - just like comparing 35mm with equal MF crop - everyone knows that 35mm lenses (the best ones) are far much better then MF ones (the best) the only quality factor change is due to FORMAT SIZE, and people seems to forget that when it comes to digital - it`s only matter of the magnification multiplier for a given format.

Example for a 20x30cm print:

Canon g3 (sensor cca. 5x7mm) magnification = 42x

canon 10D (sensor cca 16x22mm) magnification = 15x

35mm film full fram CMOS (24x36mm) magn. = 8,3x

MF 6x7 cm magn. = 4x

LF 8x10 in. magn. = 1x

 

Which one would yield the best print?

Forget about resolution, any 800 ASA film in MF is better than TMX with the best available optics in 35mm or full frame ccd!

 

Apples to oranges is comparing a CCD image with a scan of a cropped Velvia, please spare me testings like this - why not compare a projection of an LCD projector with a projection of a Velvia slide - what would be the result?

 

Anyone out there heard of colour space and Dmax? I live that to another discussion.

[zhi_da_zhong]

Taka,

 

Your scans, particularly the one in resolution test 1, look a bit out of focus to me. I

have the same scanner (Nikon LS4000ED), and I often find that I can improve the

focus a bit by moving the focusing target to a contrasty/detailed area. In this case, it

might help to put the target directly on the area that you're enlarging because the

film may not be perfectly flat or parallel to the focal plane of the scanner's lens. Also,

was ICE on for those scans? ICE can blur the image a bit.

 

Technicalities aside, I agree with Steve, et al that it would be more useful to compare

a DSLR capture to a full-frame scan (unless you always crop your film shots to the DX

size:)). But the color comparisons are still valid and quite useful.

 

Thanks for posting the comparisons.

[ben_lanterman1]

I understand from other forums that the Canon 1Ds would stand up fairly well

to a comparison with film in most all areas.