meloV8
Established
Unfortunately, SM5000 and QC cant save the file raw 16 bit, only 8bit.
Scan to TIFF 16-bit gives worse image quality than 8bit Raw.
Scan to TIFF 16-bit gives worse image quality than 8bit Raw.
ScanMate drum scanner DIY maintenance, troubleshooting, mods
- Thread starter Fernando2
- Start date
- Latest activity Latest activity:
- Replies 1K
- Views 265K
monkeyfist
Established
I had SM5000 and i did scan everything in 16bit.
What kind of image quality issues you get from 16bit files? There was some settings that outputted this weird striped messed up image... i don't remember the SM5000 that well anymore.
Was it that you could scan 16bit, but not in the raw mode?
meloV8
Established
Let's be precise. You save the file in TIFF 16 bit or 16 bit RAW? It's a big difference.
monkeyfist
Established
Had to check my earlier posts, so i used the regular 16bit. Not raw, as raw gave messed up output on the 5000.
I actually discovered the raw output when i got my 11000 and on it it works perfectly even on 16bits.
I would still use the 16bit tiff instead of 8bit raw on the 5000. As i did compare the quality, and the sharpness difference was negligible at 5000dpi on both machines when using 16bit tiff on 5000 and 16bit raw on 11000.
I had test scans up at one point, but have deleted them from my server.
meloV8
Established
When You use Tiff 8/16bit, CQ always add to scan blur and more color noise. Such is the effect on scanners sm3000/4000/5000. I have not tested this on sm11000 so I do not know. Even ColorTrio gives better quality scans than QC save to TIFF 16bit 🙂 This is a software problem, not a hardware. QC is simply underdeveloped.
monkeyfist
Established
When i tested 11000 against 5000, the 11000 was only tiny bit more sharper using the raw mode without the small blur @ 5000dpi.
I would opt in 16bit colors with the tiny sacrifice in sharpness. The only way to get perfect sharpness anyhow is to always manually focus doing multiple small test scans to find optimal focus elevation when scanning at high DPI.
I never tested ColorTrio, but i'm under the assumption it only did 8bit.
You will gain much more from 16bit colors than the small bit of sharpness from raw mode.
The slight softening wont show up in prints but colors will. And the only reason to ever scan at higher than 600dpi is for a print. As scanning huge & downscaling for web only results in a sharpening effect. You wont gain anything from it.
For web use 8bit is probably fine, i got this SM11000 for printing.. and for that it's awesome. For web scans, it's a overkill and a hassle. Flextight P2 that i had before was much better for that.
I reuploaded the test scans i made, these were made for maximum grain sharpness. SM5000 16bit tiff and SM11000 16bit raw, but using just the green channel (in b&w it will give you optimal sharpness). Both manually focused for perfect focus.
meloV8
Established
Your examples, perfectly illustrates the problem, about which I write. Scan the SM5000 has an ugly pattern that forms after writing to 16bit TIFF file. Complete lack of sharpness, otherwise just does not look like film grain. Scan the SM11000 looks different because it is reduced and this pattern is no longer so visible.
You can perform attempts of saving the file TIFF16bit and TIFF 8bit and you will not be seen any difference - both will be equally poor.
Please try the scan a slide and check the difference in the shadows. You'll see that record to TIFF degrade the quality of scans, regardless of whether the file is saved to 8bit or 16bit.
Me and my friend spent a lot of time on comparisons quality scans and both we have come to the same conclusion. For scanners SM3000 / 4000/5000 save files in a different mode than RAW 8bit degrades the quality. It is a great shame that the software does not allow CQ save files RAW16bit.
Also, other Members scanners Scanmate in this forum, have a similar opinion.
Give yourself some time to test and analysis, and of course scan only wet, if you do not want to have problems with focus.
You can perform attempts of saving the file TIFF16bit and TIFF 8bit and you will not be seen any difference - both will be equally poor.
Please try the scan a slide and check the difference in the shadows. You'll see that record to TIFF degrade the quality of scans, regardless of whether the file is saved to 8bit or 16bit.
Me and my friend spent a lot of time on comparisons quality scans and both we have come to the same conclusion. For scanners SM3000 / 4000/5000 save files in a different mode than RAW 8bit degrades the quality. It is a great shame that the software does not allow CQ save files RAW16bit.
Also, other Members scanners Scanmate in this forum, have a similar opinion.
Give yourself some time to test and analysis, and of course scan only wet, if you do not want to have problems with focus.
monkeyfist
Established
Yes, the tiff file does introduce slight softening. The SM11000 file does not have it, as it's raw scan. I think i was the first to actually discover this thing, at least on this forum.
I would still think the 12bit color would be much more important than the slight loss in sharpness. But each for his own, i don't know what purpose do you scan your files.
Have you actually done test prints to see can you really see the sharpness difference in a print? It's really small difference, and at least my monitor is about 95dpi... a print is 600/740dpi. So looking grain at 100% on a monitor, is like looking at a print with a 6x magnifier. This is why you should look at prints, to evaluate the actual end product.. not just files.
I dont wetmount, as it does not really benefit so much. It's mostly for scratches & dust, and maybe tiny increase in contrast as you have two air to air surfaces less. It would not solve the focus, it's just that the autofocus is not accurate enough for very critical work. It's pretty ok on the 5000, but when you scan at 11000 it is not.
Kamph
Established
It's not just the sligt blur - it's the overall image quality that suffers when scanning with other settings than 8bit RAW. Colours seems less accurate and shadows more blocked. It's frustrating that a proffessionel scanner is let down by dodgy software. I love my SM5000. It produces fantastic results even though I'm forced to only scan in 8bit. Yes, I do think I could gain some by scanning with 16bit instead, thats why, when money allows it, I'll be looking for a SM11000. The higher resolution dosen't matter too much to me, but if I could get a working 16bit setting I would be a happy camper.
monkeyfist
Established
When scanning TIFF, it uses a color profile.. this might have something to do with the "less accurate" colors. I assume you are scanning slides if you worry about shadows. For slides, you can even calibrate a color profile that will give out perfect colors.
I personally do not shoot slide, just colornegs and B&W. But i did do a color profile calibration for "fun" with the SM5000 i had.
I did make pretty nice prints from 8x10" colornegs that i scanned with my SM5000 in 16bit mode.. so i think you are worrying for something that does not really affect print quality.
jzagaja
Well-known
I can only confirm what meloV8 says and negatives should emphasize the problem.
https://drive.google.com/open?id=0ByP8kkW_h00uQUJkWGxTUmY3NmM
https://drive.google.com/open?id=0ByP8kkW_h00uQUJkWGxTUmY3NmM
Kamph
Established
I mostly scan negatives as well, and it's easier for me to get "accurate colours" (this will always be somewhat subjective when scanning negs of course) when using the 8bit RAW setting. It was my understanding that profiling really won't help all that much when scanning negs - I might be wrong though.
I have a Provia 100F calibration target, however, I haven't got around to actually profile the scanner yet as I rarely shoot slides these days. Maybe I should give it a go, though I must say that my scans look pretty damn close to my Provia 100F slides already when the red channel is corrected (there's always a slight red cast in the shadows).
monkeyfist
Established
Yes, profiling wont help get positives out of negatives. It will help matching the color scanner outputs to the material in the scanner. This would be done using some generic color target.
I think your problem about accurate colors is more in the photoshop process. You could skip this mostly by just doing a white balance calibration on the negative mask before scanning.
Also, if you have problems in the shadows of negative film.. you are doing something weird. This could be a problem in your white balance calibration, as white calibration sets the white point. And shadows are the white point in the negative.
monkeyfist
Established
Yes, but this is a problem of one file having a color profile and another does not. The scanner should not give perfect files right out the scanner, unless you are scanning calibrated positives. Even though the 16bit file might look worse, it has more information in it. All you have to do, is adjust the data so that it looks good to you.
Here is what i got out of my SM5000 using 16bit files:
And all the colors are just as i wanted them, as i adjusted them in photoshop the way i wanted them to be. It did not look like this straight from the scanner at all. I then printed it, and my monitor & printer is both calibrated & profiled to match, so i get pretty much what i see on the screen. This is a 61x76cm print.
If you are using the 16bit tif, i would highly recommend doing at least some type of color target calibration to have at least some type of color profile for the scanner. Instead of using something like sRGB or Adobe that might not match the scanner's output at all.
meloV8
Established
Can you show an example that shows the advantage of quality 16bit TIFF file over RAW 8bit?
monkeyfist
Established
If you don't understand the advantages of 16bit file compared to 8bit, i would be wasting time. There are plenty of information available on the internet, but it just comes down to having more information.
The most important thing is, that printers can output more than 8bit colorspace. So using 8bit files is limiting number of tones you can put out in a print. This will show up especially in B&W work, as you are working with only 256 tones.
You could also do the RAW vs JPG comparison in digital photography. Raw files are 16bit and JPG 8bit. You shoot RAW when you wish to edit files, and JPG when you want a picture straight out of camera. You don't want to edit 8bit JPG files much, as the tones in them will break.
8bit was the norm in scanning when computers could not handle huge 16bit files, so they adjusted everything (curves, tones etc) ready and scanned directly as 8bit file. But as you are scanning 8bit raw, you are also skipping this important phase of 8bit scanning.
I came across this picture:
8bit really struggles in gradients where you need lots of tones.
meloV8
Established
I fully understand your point of view. When I scan Epson and Nikon scanners, used only scanning 16 bits because I see differences in the scans, the software works exactly as it should. The theory works. However, in the case of Scanmate, the practice is completely different.
Can you show an example of your scan, showing a 16bit TIFF advantage over RAW 8bit?
Can you show an example of your scan, showing a 16bit TIFF advantage over RAW 8bit?
Kamph
Established
No, I don't have any problems with the shadows when scanning negatives, only when I scan positives - sorry if I wasen't clear on that.
Some time ago I tried to WB on the negative mask but it resulted in an error as mentioned before. On the resulting scan the negative mask wasen't eliminated either. I'll try again and see if I get a different result.
monkeyfist
Established
I did show the print? It's scanned 16bit tiff on SM5000. Seems fine? There is no way i could have done that print from 8bit scan, especially using colornegative that requires extensive post work.
monkeyfist
Established
Quickest way to see if the calibration worked, is to do a preview.. the new white balance should show up on the preview. Some masks are darker, so it might not work on every film. I dont know, has worked fine on all films i have tested it so far.
If you are scanning positives also, i would recommend calibrating the color profile even if the film target you have is not for the exact film. It's still better than nothing, or at least as a test.