Dedicated film scanner vs DSLR - what do you think?

<p>Have a look at the two b&w images on this page:</p>

<p>http://www.thedambook.com/scans/</p>

<p>Compare the DSLR and the film scanner. Tell me what you think. IMHO the difference is not subtle and there is a clear 'winner'.</p>

<p>The film scanner's image is clearly superior, but we don't have any details about how the two images were generated other than the model of scanner and camera used. Was the camera set to its optimal ISO? What lens was used? Did he shoot a slide (illuminated how?) or a print?</p>

<p>It's also not fantastically surprising that a device designed for scanning film does a better job of it than a camera that is really intended primarily for other things.</p>

<p>Looking at the mess the D300 made I would rather stick to using my V500.</p>

Stuart, what was the purpose of the image included in your response?

<p>Well it looks better than the mess that the D300 made.</p>

<p>I am surprised at the amount of grain in the D5 image compared to that in the LS-8000 scan. T-Max has a needle-like grain structure, and the D5 image looks close to what you would see under a microscope. The sharpness suffers in the corners of the LS scan, which is likely due to poor film flatness. That's best addressed by using a glass carrier. I think either version could be used to make an acceptible 11x14 print, perhaps a little larger. What more do you expect from 35mm film?</p>

<p>"Winner" is a subjective term. It takes an LS-8000 2-3 minutes to scan a 35mm negative (a V700/750 takes about the same), which the D5 could do in a fraction of a second. Facing a deadline, I'd copy using the D5 without hesitation. But then, why would I shoot film when I had a D5 ;-) Perhaps the real question is which to use if you are archiving a box full of family negatives. Again the D5 would make more sense, based on flexibility and time. I would scan those old, unmounted prints on a flatbed at 300-600 ppi, which takes only seconds. Mounted prints (scrapbooks and frames) would probably be easier with the D5.</p>

<p>I would not expect better from an optical print. It is very hard to achieve grain-sharp prints from an enlarger, due to a combination of negative flatness, paper flatness, lens limitations and vibrations during the relatively long exposures. In "the day", I seldom used glass carriers because of dust. I did use a focusing microscope, which shows just how sensitive the focusing of an enlarger can be. Without a vacuum platten, variations in the paper flatness easily exceed the DOF at f/stops which keep a reasonable exposure time (<60seconds).</p>

<p>I'm sorry, somebody please help me with my ignorance--this is a somewhat interesting topic, but D5? What is that? Where do we see it was used? Also, I can't seem to pull up the white paper. And maybe more importantly, is this about using a camera in place of a film scanner? Because I don't think anyone advocates doing so where quality is important. AFAIK, the camera-as-film-scanner is only useful where you have a box of old family negatives that you want to digitize in a short period of time with a $99 device.</p>

<p>Les, just curious, and not wanting to reopen old debates, but you regard Astia 100F as sharper than Velvia 50 or Provia 100F? Because certainly that is not what Fuji says. Just curious. Thanks.</p>

<p>The referenced thread does indeed compare a scanner to a DSLR as a copy camera. I wrote D5 instead of a 1Ds. I misread the "s" as a "5", and didn't check the results. A Canon 5D (not D5) at 18MP would be a reasonable substitute for a 21MP 1DsMkIII in this application.</p>

<p>Dave,</p>

 

<blockquote>

<p>the camera-as-film-scanner is only useful where you have a box of old family negatives that you want to digitize in a short period of time with a $99 device</p>

</blockquote>

<p>I am not at all sure about that for slides. A good macro outfit (bellows + enlarging lens) or macro lens, would work pretty well I think and I think is a lot better than you imagine. Maybe not quite up to a good scanner quality but pretty close and much quicker - although a lack of dust removal would be a pain - but there again for Kodachrome dust removal does not work anyway in a scanner. Have you tried this set up? Many archivists are using just such a setup because wholesale "real" film scanning is often impractical.</p>

<p>I have a lot of slides I would like to copy to digital for on screen use (no larger than a 1080p television. Scanning them all would take me months. Using a digital SLR with a bellows is ideal for me. I have not, however, found an easy way to use my DX format camera to do this yet (I have the Nikon PB-5 Bellows and slide copying attachment). The 55mm Nikon Micro-Nikkor provides too much magnification. I am thinking to try my Nikon 28mm f2.8 AIS that has CRC for better flat field performance at close range.</p>

<p>The 1Ds mkIII actually did pretty well. Raw resolution appears to be close, just a bit better in the Nikon scan. And the 1Ds3 actually recovered more highlight and shadow detail. The problem with the 1Ds3 scan is that it is over sharpened relative to the Nikon scan, and this just <strong>destroyed</strong> the grain structure. This gives a horrible first impression, but it really is just one aspect of the scan and one problem to solve.</p>

<p>I'm guessing that more careful adjustments and post processing would yield a better result, though I expect that the CoolScan, which is built specifically for the task, would still exhibit an edge. It's hard to say how much of an edge in terms of suitable print sizes (for example). It all depends on what can be done to render a more pleasing grain pattern in the 1Ds3 scan.</p>

<p>You might be surprised at how good it can be, I use a Kodak Pro SLR/n with a Beseler Dual Mode Duplicator and get quite usable results. I use an 80mm Rodagon and a modified macro tube with a cpu mounted to allow aperture priority operation, it takes only several seconds per scan, shooting tethered RAW. Would I try to make a 20x24 from it, no. But for up to 10x15 prints off the Frontier at the studio, it's perfect, and far quicker than the scanner on the Frontier.</p>

<p>My opinion: the 1Ds was awful. Sharp and convenient - but awful. Look at that horrible mess! It reminds me of NR smear from old Kodak Pro DSLRs. Yuck.</p>

<p>You are entitled to your opinion, but I disagree in part. The LS-8000 scan has better quality, but the 1DsIII would be acceptible. I've seen (<em>i.e., </em>made) many less acceptible LS-8000 scans, and many DSLR shots superior to the ones in this example. The image processing for the 1DsIII image could be improved by imposing a stronger curve and using less sharpening. As in "the day", the real work is in the darkroom (or it's modern equivalent).</p>

<p> I think the fiddler guy is the same exact picture. One picture post processed better then the other. </p>

<p>They look pretty similar to me. </p>

<p>I think the Coolscan may have some DoF issues as the far eye looks soft, perhaps from film curvature. Look at the bridge of the violin- it's totally out of scanner focus. For that reason alone I'd take the camera "scan."</p>

<p>I would have sharpened the camera shot a bit differently myself to avoid emphasizing the grain as much. Still it's not likely to be visible in the final print.</p>

<p>Lee,<br>

The same "scan" could have been obtained via a c. $2500 5dMkII. No need to spend the extra for the 1Ds.</p>

<p>"Creamier on the real <em>scan</em> due to actual tonality resolved."</p>

<p>Look at the violin, Les- it's an obvious film flatness issue with the 8000. Look above the violin and watch as the grain goes from grain-like (towards the left) to mush (right) over a continuous background. I've scanned enough to know that is not what the actual film looks like.</p>

<p>I think this is good news for film shooters as it means there will be viable ways of digitizing film after the last scanner has been discontinued. How many film shooters *don't* also have digital cameras these days anyway? I know I do!</p>

<p>The 1Ds MKIII capture looks like some kind of crazy photoshop texture has been applied to it whats up with it.</p>

<blockquote>

<p>looks like some kind of crazy photoshop texture has been applied ...</p>

</blockquote>

<p>Do you mean the "worms?" It's the typical artifact from aggressive unsharp masking.</p>

<p>Try looking at it in Photoshop, not downscaled by your browser. I'm looking at a side by side in photoshop.</p>

<p>It looks like someone smeared a picture, sharpened it and the applied a watercolor and stained glass texture to it.</p>

<p>The person using it the 1Ds III as a scanner would have been better off reducing the size to 5-10 megapixels that sharpening smeared detail.</p>

<p>Not sure what this person will do with this white paper..... no innuendos intended...</p>

<p>In fact, if you reduce the size in photoshop to 6 megapixels or lower the mustache buds don't look painted anymore. This is the only cope with lack of detail.</p>

<p>Interestingly, as I entered many discussions on detail over the years, I know understand that people who don't notice severe difference in detail when comparing capture methods (coolscan vs 1DsIII scan, 5DII vs 7D, film vs digital, etc) are not ignorant or lack evidence; they just simply have a difference in perception (or use very tiny monitors).</p>

<p>More so, since I have an appreciation for detail as well as two 55 inch calibrated LCDs and large format printers, differences are more obvious to me.</p>

<p>Those who can't tell the Coolscan provides more than 4x the detail of this 1DSIII scan should run to best buy. That said, much of this is operator error trying to make up something that is not there.</p><div></div>

<p><em>Those who can't tell the Coolscan provides more than 4x the detail of this 1DSIII scan should run to best buy </em>[best Buy]</p>

<p>The 1DsMkIII has a resolution of 5616 x 3744 pixels, compared to a nominal 6000 x 4000 pixels for the LS-8000 scanning a 35mm negative. How that comprises a 4:1 advantage is a mystery to me. A 6x7 negative scans at approximately 10000 x 8000 pixels - still far short of 4x.</p>

<p>This is notwithstanding the obvious fact that the difference lies in the microstructure, not the pictoral detail in the two images. This suggests a trip to Lenscrafters for some, not Best Buy ;-)</p>

<p> </p>