Dedicated film scanner vs DSLR - what do you think?

[mauro_franic]

<p>Edward, reduce the samples from 20MP to 5MP and then upsample them back to 20MP. You will see that the Coolscan still has more detail - meaning the 1DSIII scan was even lacking to provide high detail at 5MP. (You may do this quick using my crop above).</p>

<p>With this I am not judging the detail the 1DsIII could have provided of the subject directly, just the detail it captured used as a scanning device.</p>

[mauro_franic]

<p>There is just a lot less detail on the 1DsIII scan.</p><div></div>

[mauro_franic]

<p>Still puzzled how the difference in detail is not obvious and why there is even a debate on that particular point. </p>

<p> </p>

[roger_smith4]

<p>The detail is fairly obvious but scanning isn't just about detail. If you gave me two scans, one with blown highlights and serious film flatness issues and the other with less detail and exaggerated grain I'd choose the latter. As a LS-5000 user, I'm sure the LS-8000 is capable of better results than this with a bit more care taken and also think the 1dsIII would look much better with sharpening applied differently.</p><div></div>

[mauro_franic]

<p>Film flatness is not an issue using the glass holder. The Coolscan doesn't bow out highlights, the person processing the picture can. </p>

<p>The 1DsIII scan was indeed sharpened pass the level of detail contained thus creating almost a water painting effect.</p>

<p>Operator's shortcoming aside on both images, the Coolscan produces excellent scans but 1DsIII is not a capable as a serious scanning tool (much less for archival).</p>

<p> </p>

[craigd]

<p>I think I see an emerging consensus here that the comparison is not very well done due to technical deficiencies in both the handling and image processing. This leaves me thinking that as an advertisement for someone trying to sell a book (which the linked page basically is), it's not very effective. I'm certainly not going to buy the book if this is typical of the author's way of comparing and contrasting different solutions.</p>

[mauro_franic]

<p>Agree. It is like George Constanza advertising fitness equipment. </p>

[roger_smith4]

<blockquote>

<p>Operator's shortcoming aside on both images, the Coolscan produces excellent scans but 1DsIII is not a capable as a serious scanning tool (much less for archival).</p>

</blockquote>

<p>That's too categorical a statement unless you meant to say the "1DsIII is not AS capable as a serious scanning tool." Considering the number of people who find flatbed scans acceptable for their purposes, the 1DsIII appears a good bit better than that to me based on this scan.<br>

<br />Based on another document on the site I think the author is actually making a case for using DSLRs to do quick and dirty sorting and categorizing of film rather than using it as an alternative to scanning.</p>

[mauro_franic]

<p>Roger, for sorting and categorizing it may be useful.</p>

<p>I agree also the speed could be a big plus. Probably downsized to 5mp to improve handling without losing much detail. With color, and especially print, I doubt the output would be of any use though even as a simple dirty sorting/evaluation.</p>

<p>I am not sure though whether this would be easier, faster or better quality than scanning multiple negatives at a time with a V700 at 2000dpi.</p>

[Ed_Ingold]

<p><em>I agree also the speed could be a big plus. Probably downsized to 5mp to improve handling without losing much detail.</em></p>

<p>Scanning at lower resolution makes a big difference with a Nikon LS-8000 or Epson V7xx - primarily through reduced processing (e.g., Digital ICE) and transfer time. However the 1DsIII takes the same time, a fraction of a second, regardless of the resolution. If you are going to take the time to set up the job, it makes sense to finish it at the highest available quality, scanner or DSLR. Transfer time for a 50MB raw file from card to computer is less than 3 seconds (FW800 card reader), a fraction of that for a scanned image.</p>

<p>Before scanners, slides and photos were copied to film using the same setup proposed for the Canon DSLR. The Canon clearly has more resolution than 35mm color film, so why not make this logical jump? On the practical side, Canon and Nikon are still making cameras, but not dedicated film scanners.</p>

<p>Why not, indeed? The answer is because an unrepentent filmfan is unwilling to concede that a DSLR can ever excel at anything, notwithstanding that an LS-8000 is nothing more than a digital camera in a big box with a built in film holder ;-)</p>

[mauro_franic]

<p>Edward, my statement compares the DSLR scan versus using a flatbed scanner. There is no argument a DSLR does not compare to a Coolscan as a scanning device.</p>

<p>I was referring to downsizing the DSLR capture to 5MP (not the Nikon scan).</p>

<p>Regarding speed/use for dirty cataloging I posed the question on what would be easier/faster/better quality (1) a DSLR or (2) a page full of negative on the V700. Unrelated to the Coolscan.</p>

<p>As you know, a scanner is not a DSLR in a box. A DSLR produces only interpolated color and antialiased blured outputs. A scanner does not interpolate color nor applies a bluring filter and it does have ICE.</p>

[mauro_franic]

<p>Also I am unfamiliar on the ease or difficulty of lining up individual frames to shoot them up with a DSLR. If anyone is, they can answer how this compares to using a flatbed with sheets full of negatives at a time and ICE turned on.</p>

<p>From what I have seen, a flatbed produces better outputs than the 1DsIII example posted here. If there are better examples of DSLRs as scanning tools I would like to see them. It is always good to try to search for faster and better way to do things.</p>

[Ed_Ingold]

<p>In case I wasn't clear, I see no point in downsizing the DSLR shot either. It's a waste of time to sabotage quality during a scan.</p>

<p>A full page scan with a flatbed is problematic with respect to the correct exposure. The individual frames are not optimized, and have little value should you want to use them for editing and printing. Setting up separate frames for each negative in a full page shot is time consuming in itself. Since film placement is subject to random variations, it is nearly always necessary to fine-tune the framing in a flatbed.</p>

<p>A scanner is not a DSLR (q.v., single lens reflex), but it is in fact a camera. Not all DSLRs have anti-aliasing filters, and some (q.v., multi-shot and scanning MFDs) do not interpolate color. In practice, the AA filter causes relatively little loss. One is hard-pressed to demonstrate how color interpolation has adverse effects on resolution, although it does create color artifacts in highly specific instances.</p>

<p>Once you have a setup for framing and lighting, it is no more difficult to use a copy camera than a slide scanner (one frame at a time). If you don't already own a film scanner, it would pay to familiarize yourself with the process.</p>

[mauro_franic]

<p>No point in beating known limitations of DSLRs to death.<br>

I am still interested to see examples of people that claim DSLRs can provide with good scans because I have never seen one.<br>

Are there any post anywhere on the web available of good results of 20MP/35mm scanning using a DSLR? <br>

The processing in the example on this thread makes difficult to determine whether good results could have been obtained. </p>

[roger_smith4]

<p>I use a holder from an Epson flatbed to "scan" Diana 120 slides using a Canon 20D (8mp). The resulting image is lower resolution but has better color fidelity and lower noise than the middle-range flatbed. I can also jack up the brightness of the light source (flash) to compensate for uniformly dark slides which is handy. </p><div></div>

[mauro_franic]

<p>Could you post a 100% crop for inspection?</p>

[roger_smith4]

<p>Ha- the Diana and its plastic lens is not made to be "inspected" at any size. I'll see if I can get you a crop when I get home.</p>

[Robin Smith]

<blockquote>

<p>Why not, indeed? The answer is because an unrepentent filmfan is unwilling to concede that a DSLR can ever excel at anything, notwithstanding that an LS-8000 is nothing more than a digital camera in a big box with a built in film holder ;-)</p>

</blockquote>

<p>I completely agree with Edward. If a DSLR can essentially take a cleaner and to many a "better" image than 35mm film (in general appearance much the same as a MF film image), why can it not essentially take an image of a slide that is at least as good as an MF dupe of a 35mm slide would be? For many people this is quite good enough: in fact I suggest it is probably good enough for all but the most minute pixel peepers.</p>

[Ed_Ingold]

<p>The main impediment to using a DSLR for slide/film scanning is the lack of an high quality device to hold the film in juxtaposition to the camera. You need something to hold the film accurately and parallel, provide for illumination while excluding incident light, and allow for fine focusing and (ideally) positioning.</p>

<p>Over the years there have been many shoddy slide duplicators, including the Bower device (q.v., BH Photo) which uses a simple 10 diopter lens which attaches directly to a camera with the proper adapter (Nikon, Canon, Olympus, etc). The best devices attached directly to a macro lens, often with a separate focusing tube or bellows. These seem to have all but vanished from the marketplace.</p>

<p>I'm scratching my head, thinking of what I might assemble from detrius accumulated over the last 50 years or so, including bellows, various focusing rails and light boxes and a selection of macro lenses. The next problem is finding a 35mm slide. The last ones I shot were in 2002.</p>

[mauro_franic]

<p>Robin, you are confused.</p>

<p>A DSLR scan of 35mm is materially inferior than a simple scan of 35mm. </p>

<p>Your idea that the above is not true but also a DSLR scan is "at least as good as a MF dupe of a 35mm slide" is, no offense, an indication you don't have any experience or understanding of this topic.</p>

<p>After reading this thread and looking at the results of attempting to scan with a DSLR you quote "it is probably enough for all but the most minute pixel peepers" Strange.</p><div></div>

[mauro_franic]

<p>To help understand.</p>

<p>Here you can compare:<br /> (1) 35mm film scanned with a scanner<br /> (2) 35mm film scanned with a DSLR<br /> (3) 35mm film scanned with a scanner, then downsized to just 5MP, then upsized back to 20MP</p>

<p>You may observe the following:<br>

(3) still contains more detail than the DSLR attempt. The DSLR is not even good enough for 5MP worth of real detail.</p><div></div>

[erie_patsellis]

<p>Edward, slide duplicators pop up from time to time. I waited 2 years to find mine, but having a dichro light source and proper bellows with a macro lens certainly makes a difference.</p>

<p>I'm amazed at how many are ready to dismiss this without trying it. (or for that matter, have the temerity to tell me that my methods are wrong for my needs!) There is a slight difference when I scan, admittedly, but why the hell would I even bother shooting 35mm for large prints? I have two MF outfits and shoot LF, both film and digitally, so it's the wrong tool for the job, period.</p>

[mauro_franic]

<p>Erie, could you please post a sample of the results you have been able to accomplish scanning with a DSLR?</p>

[mauro_franic]

<p>Also when you mention the difference is slight, it would be great if you shared what DSLR and what scanner are you referring to for the comparison. Thank you.</p>

[erie_patsellis]

<p>When I mention the slight difference, I am comparing scans made with a Polaroid Sprintscan 4000+ to the same image "scanned" with a Kodak SLR/N (shot at base ISO in RAW, no NR and conversion in Photodesk). As to the previous comments about the Kodak and NR induced artifacts, they are capable of stunning images, with the right person. Give me any DSLR and I bet I can make bad images from it in a matter of minutes. For those that are willing to use it in it's intended role (as a studio camera, with controlled lighting), it's just as capable as any other.<br>

From time to time, I will use a Leaf Micro Lumina F-mount scanning camera, the scans with it are as good (for my needs) as any scanned 35mm image I have used, as long as there aren't Dmax issues to contend with. It's not instant capture nor a DSLR and I won't muddy the waters with that comparison.<br>

The majority of my personal film work is C41 based, so Dmax isn't an issue. I process and print all my own images, with occasional scanning for online viewing or to share an image. Neither scanning method compares to optical printing done properly however. What I don't understand is this overwhelming desire to convince others that "they're wrong, you're right and just get over it." I have better things to do than compare minutiae between scan A and scan B. In my opinion, if quality were that important, you'd be using much larger negatives and cut the B.S. out totally. You have to try really bad to get a bad scan from a 4x5 negative or tranny. This reminds me too much of the "Leicanistas" who seem to thing sharpness is the overall indicator to image quality and will spend thousands just to get a 5% increase in perceived sharpness. Is it really that important that you get every last bit of potential detail out of a negative?</p>

<p>One of the projects I am involved in is (slowly!) digitising an archive of approximately 5,000 images for the local historical society, all done out of my own pocket, I do a box here and there as I have time. Is it really that important to wring every last bit out of an 8x10 glass plate? Doubtful in this application (like the one in the article referenced, btw) simply having an accounting and a general idea of what the image is of is more important. Anything determined to be important has a duplicate negative made of it and a contact print made. Should they want murals made, I can just as easily rescan an image as needed, no need for such demanding resolution for what amounts to a catalog of images.</p>

<p>I suppose if you need to justify in your own mind that the pile of money you spent was an "investment", by all means continue on, but I have far better things to do than continue in such pedantic drivel.</p>