Film scanning question

in iPod + iTunes + AppleTV edited January 2014
I've got a Minolta Dimage Scan Elite II film scanner (with IR channel). Using Dimage Scan Utility version 1.1.4. OS X 10.3.3 Scanning at maximum resolution (2840 dpi, or about 9 megapixels). Saving as 16-bit TIFF (around 60 MB). 8x sampling. Source medium is color 35mm slides about 30-40 years old but in good shape.

So when I scan, does it matter if I use ICE, ROC, GEM, or fiddle with color controls, e.g., contrast, RGB, brightness, etc. prior to scanning, or is all of that stuff alterable after the scan using various software programs, e.g., Photoshop?

If I can just as easily alter all of this after the scan with the same effect, than I'll spend a lot less time trying to find the perfect combination of settings to scan with (which is very time consuming since I scan the same image over and over after continually tweaking the settings). What settings are unalterable after the scan? Any advice?

I'm obviously not a pro at this, and am just looking to scan around 200-300 of my grandparents old slides from the 1960's.


  • Reply 1 of 5
    buonrottobuonrotto Posts: 6,368member
    Ideally, the scan comes in right the first time because each successive generation will be inherently more degraded than the previous one. Basically, my rule is that the first generation that I'm scanning should look like the original film is supposed to in its condition. I can then go back and later color and grain with a finer touch and not lose the original look in case I screw up. I use ICE since getting rid of scratches and dust is quite hard after the fact, and ICE technology does a good job at this. I adjust my levels/curves when scanning because I'm most afraid of losing some luminosity and color saturation if I go back later and try to do this. GEM does a better job than you at removing digital noise from the image, but I think it can be done post-process, and might want to be depending on how aggressive the scanner is at this. I prefer to tweak the images by hand than use ROC. It does a good job, but you might want more control over the process. I think it also depends o how skilled you are with Photoshop, whether you have access to the full version with masks, channels and such too.

    This page has excellent information about scanning. It's supposed to be the basics but it gets pretty in-depth about most of this stuff.
  • Reply 2 of 5

    Thanks. I'm applying ICE, GEM (light, or 33), and ROC at the time of scanning. I don't have access to Photoshop but will spring for Photoshop Elements if necessary. I can't see spending the money on the full version of Photoshop.

    My photo manipuation skills are pretty basic, so it's probably best that I apply ICE, GEM, and ROC as part of the scanning process and use as many of the auto features as I can.

    I didn't realize what a science scanning photos was until I actually tried it. It actually seems pretty complex if you really get into it.

    Any comment on TIFF 16-bit vs. TIFF 8-bit vs. JPEG? I assume that TIFF 16-bit would be the prefered setting of the three. I can always export from iPhoto as a JPEG when required.

    And any comment on oversampling? The Minolta software goes up to 16x, but that seemed excessive. Even 8x is probably too much. Plus it takes forever to scan when it goes over the image so many times.

    Part of my problem is that the Minolta software documentation is so poor that it doesn't help guide the user, e.g., explaining really how many times I need oversampling, for example, or is TIFF 16-bit really that much better than TIFF 8-bit. I'll check that website you mentioned.
  • Reply 3 of 5
    buonrottobuonrotto Posts: 6,368member
    If you want as best an electronic archive as possible, scan at 16 bit and back up the originals and fixed-up versions to gold archival CDs. Years later, when technology permits, you can probably take better advantage of these high bit-depth files. I scan at full bit depth, back up the CD and lower the bit depth of the copies on the hard drive for my printer or to publish to the web. no sense in wasting disk space.

    If you want to fix them up and publish to the web or print them, you can downgrade the bit depth to 8 bits because you'll never take advantage of it on home devices. Consumer printers can only print thousands of colors, only pro printers print in the low millions and cost a fortune to get those prints. Obviously, that's going to get better over time both for pro printers and home ones, but its a ways off. Our monitors can display the 8-bit files well enough, but they have to interpret on 16-bit colors. Publishing to web also means no real color management/calibration so the depth would be wasted IMO.

    TIFF is always preferred because it's lossless, no change of shifting color or losing data from compression that you can't get back. I'm not sure if JPEG format can save 16-bit files, can it? Another thing I do once I've backed up my original scans is I save the 8-bit ones on my drive as JPEGs, 95% quality. They don't miss the data that's thrown out and they're far smaller. Be aware that if you bring this stuff to a pro print house, it's best to leave them as TIFFs though.

    Oversampling is overdone on these things. Use 2x oversampling. Everything else is a waste, unless you're scanning line art, then go to 4x. The 16x thing won't improve the accuracy or quality of the scan beyond what the 2x sample will do.
  • Reply 4 of 5
    Okay here's what I'm using for scan settings:

    35mm color positive slides

    2820 dpi

    TIFF 16-bit



    Oversample x2

    Use monitor ICC profile



    GEM (33, or light)

    Auto setting for image correction - tone curve and histogram; brightness, contrast and color balance correction; and hue, saturation and lightness correction

    Those image correction settings are the ones that seem to be the most time-consuming. I can see spending an hour a slide tweaking all of these things, but that's just too much time given the number of slides I have (plus my two newborn twin girls really eat into my free time). And since these slides were taken over a number of years, I don't think I could develop a generic image correction profile and apply to all.

    The Dimage manual says that JPEG can not save at 16-bits.

    What's your opinion of services that offer to transfer slides to digitial format? I understand that some Kodak outlets will do this, although it can be expensive.

    Thanks for all of your help. A few quick remarks from someone who knows what they're doing can save hours of puzzling over a manual or surfing the web.
  • Reply 5 of 5
    buonrottobuonrotto Posts: 6,368member
    Some photography services in the area will do a better job than sending them off to Wegman's or a Kodak outlet. They're prohibitively expensive though, usually like $1.50 per frame. I doubt they would spend a lot of time fine-tuning the colors per frame anyway.

    I'm wasting all my time on the color correction thing myself. I'm doing it with about, hm, let's see? about 2000 slides from my stay in Italy, and it takes all day to do a roll. Each one is different, though I must say that comparing what I get out of these negatives is a hell of a lot better than what the film developers gave me in the first place. Slides probably don't have that problem because there's no extra step involved -- it is the film, which was always good.

    It all sounds good. Now go have fun with those cut little girls! I got a soft spot for twins.
Sign In or Register to comment.