'grain' on scanning toned prints

[tim rudman]

Hi everyone,
It's a while since I was here because I am a darkroom worker with little need for digital techniques, for the most part.
I recently started scanning 16 x 12 prints on my new Epson 10000 xl A3 flatbed, using the Silverfast Ai that comes bundled with it. I am not experienced in digital techique but found the software fun to play with and so far quite easy. However, I am surprised at the prominence of grain on the screen when scanning split bleach-toned prints.

They are lightly bleached before sepia toning, i.e. extending into but not through the mid tones, which therefore contain a mix of yellowish sepia and black silver. The scans appear very much more grainy in these mid tones than do the prints.

Using one of the 3 'grain & noise removal' settings does reduce it at the expense of sharpness, but even slight sharpening with unsharp mask increases it again.

Is this an unavoidable effect with scanning, or is there a way around it?

Tim

 

[sanking]

Tim,

Have you tried looking at the print in different channels in Photoshop, or in whatever image processing software you are using? You will often find that grain and sharpness are much better with one of the RGB channels than with all of them. Just click on channels and then visually evaluate grain by looking at the individual channels. In scanning B&W film most people find that the green channel gives the least grain and sharpest image. I don't know about prints but it is entirely possible that the grain is caused by one of the layers and not all three.

Sandy

 

[tim rudman]

Tim,

Have you tried looking at the print in different channels in Photoshop, or in whatever image processing software you are using? You will often find that grain and sharpness are much better with one of the RGB channels than with all of them. Just click on channels and then visually evaluate grain by looking at the individual channels. In scanning B&W film most people find that the green channel gives the least grain and sharpest image. I don't know about prints but it is entirely possible that the grain is caused by one of the layers and not all three.

Sandy

Good idea Sandy, I didn't think of that, but I can see the logic, especially with this black-brown mix in the midtones. I'm still pretty new to these digital skills that I will have to learn, so thanks a lot. Is it normal for fine detail like grain to be exaggerated by scanners?

BTW, I scanned at 300 dpi with output to equal original in size. This is nowhere near the max for this scanner of course but the file size gets massive on the highest settings. Is this a wise approach and would it affect the quality much unless I want to print larger than original?
Tim

 

[jd callow]

Tim,
It has been years since I've used silverfast, but it sounds like there may be one or two issues to check.

First off to critically evaluate the image file you need to use Photoshop (or similar),not in the scanning software, and the file should be viewed at100%. What appears to be grain at 27% (or what ever percentage allows you to view the entire image on the screen) may only be the arbitrary rendering of grain as the program tries to scale the image to the view pane.

In the scanning software turn off or tone down any sharpening. This may be the cause of your problem.

In addition see if the colour, exposure, curve optimization settings aren't set too aggressively and causing an increase in local contrast and giving the appearance of greater grain.

I tend to scan images 'raw', with no clipping to the histogram, no exposure adjustments, no colour adjustments and no sharpening. I open the file into photoshop where I adjust and sharpen. The image adjustment tools found within the scanning software help in a production environment, but are not as good as the tools found in a good image editing program.

 

[tim rudman]

Tim,
It has been years since I've used silverfast, but it sounds like there may be one or two issues to check.

First off to critically evaluate the image file you need to use Photoshop (or similar),not in the scanning software, and the file should be viewed at100%. What appears to be grain at 27% (or what ever percentage allows you to view the entire image on the screen) may only be the arbitrary rendering of grain as the program tries to scale the image to the view pane.

In the scanning software turn off or tone down any sharpening. This may be the cause of your problem.

In addition see if the colour, exposure, curve optimization settings aren't set too aggressively and causing an increase in local contrast and giving the appearance of greater grain.

I tend to scan images 'raw', with no clipping to the histogram, no exposure adjustments, no colour adjustments and no sharpening. I open the file into photoshop where I adjust and sharpen. The image adjustment tools found within the scanning software help in a production environment, but are not as good as the tools found in a good image editing program.

Thanks jd
I thought I recalled being told that it was always best to do everything at the scanning stage where possible, to reduce the editing required in Photoshop.
I will check at 100%. I didn't know that.
Re sharpening, I used it very sparingly. So it would be better to scan unsharpened?
"In addition see if the colour, exposure, curve optimization settings aren't set too aggressively and causing an increase in local contrast and giving the appearance of greater grain."I did experiment (on a totally ad hoc basis!) to see what effect this had and you are of course right, it makes quite a difference. Trying to keep the colour and tones faithful to the print though, still gave me an exaggerated grain effect. GAN filter helped a bit though.
Tim

 

[jd callow]

Adjustments done at the time of the scan, in my experience, are never as good as those done in image editing software. If you are careful to capture all of what is present in your print and if your scanner hasn't added things that aren't there (noise, sharpening artifacts, and uncorrectable colour casts are some of the things to look out for) then you should be able to get better results in photoshop. Photoshop is not the easiest tool to get a handle on and to get the curve tool to work as you'd like is a pain, but it is worth the effort. Once you conquer curves and sharpening on adjustment layers in RGB, learn to do the same in LAB and you'll have taken the tool about as far as it can go -- or at least as far as I've taken it to date...

 

[tim rudman]

Sandy, thinking about what you said re rgb channels, these prints are B&W but toned for colour, therefore I couldn't pick one channel without distorting the colour. Or am I missing something?
Tim

 

[sanking]

Sandy, thinking about what you said re rgb channels, these prints are B&W but toned for colour, therefore I couldn't pick one channel without distorting the colour. Or am I missing something?
Tim

Tim,

If you find out that one of the RGB layers is the culprit for the excess grain you might be able to minimize the grain on that channel alone, with blur for example, without reducing overall sharpness.

But for sure you want to make sure that no sharpening is turned on when you make the scan, as JD suggested. Sometimes sharpening is the default with software.


Sandy

 

[tim rudman]

Tim,

If you find out that one of the RGB layers is the culprit for the excess grain you might be able to minimize the grain on that channel alone, with blur for example, without reducing overall sharpness.

But for sure you want to make sure that no sharpening is turned on when you make the scan, as JD suggested. Sometimes sharpening is the default with software.


Sandy

Thanks Sandy, JD
I'll start again!
I thought it seemed too easy!
Tim

 

[tim rudman]

Thanks Sandy, JD
I'll start again!
I thought it seemed too easy!
Tim

Just before I start rescanning this set, can I check that my setting of 300 dpi is a good choice for a scan to equal original in size (16 x 12")? I used 48 bit depth.

The scanner has an optical resolution of 2400 x 4800 dpi but scanning say 100 prints at this takes a lot of space!
Tim

 

[donbga]

Just before I start rescanning this set, can I check that my setting of 300 dpi is a good choice for a scan to equal original in size (16 x 12")? I used 48 bit depth.

The scanner has an optical resolution of 2400 x 4800 dpi but scanning say 100 prints at this takes a lot of space!
Tim

Tom,

What is your final intent for the scans?

Don Bryant

 

[sanking]

Just before I start rescanning this set, can I check that my setting of 300 dpi is a good choice for a scan to equal original in size (16 x 12")? I used 48 bit depth.

The scanner has an optical resolution of 2400 x 4800 dpi but scanning say 100 prints at this takes a lot of space!
Tim

Tim,

If the final print size is the same as the original there is no need to scan at more than 360 dpi, assuming you plan to print with an inkjet printer.

Sandy

 

[Marco B]

Just before I start rescanning this set, can I check that my setting of 300 dpi is a good choice for a scan to equal original in size (16 x 12")? I used 48 bit depth.

The scanner has an optical resolution of 2400 x 4800 dpi but scanning say 100 prints at this takes a lot of space!
Tim

Never try to scan a print at the highest optical resolution, it's ludicrous. These scanning resolution, 1200 ppi - pixels per inch - and upwards, are only useful for scanning negatives, not prints (and see other threads here on Hybrid to realize that most flatbed scanners actually are not capable of more than 1200 - 2400 true optical resolution at best).

If you need to scan a print and want to capture every bit of detail from the original darkroom print, than 600 ppi is the setting to go for, but be aware that every little scratch and dust speck will be visible and may require more PS work to remove.

Higher ppi's are useless, especially for the large 16x20 inch original size you already have, lower will mean a slight loss in detail and information captured. For just a reproduction at the original size, as you seem to aim for, the 300-360 ppi suggestions above are OK.

 

[tim rudman]

Thanks everyone.
The purpose of the scans? To digitize the prints for general use, which will include downsizing for web use, sending for reproduction in magazine and book form, cards, possibly posters & etc. I do want to capture every detail on the print to the best quality I can, but I understand there is no point in exceeding a certain level just because it is possible to do so.

At present I don't particularly want to reprint them digitally as I prefer to wet print, but who knows. I may well make substantially larger digital print copies in time but would expect to rescan specifically for that if and when.

I was under the impression that 300 dpi was the figure to go for. Sandy, you suggest 360. Is that better for my purpose?

Thanks for you help everyone. I'm more at home in the darkroom
Tim

 

[jd callow]

360 is a resolution specific to epson printers. Epson printers print at multiples of 360: 360, 720, 1440, 2880 and so 360ppi images are more efficiently printed by an epson. The 360ppi resolution becomes meaningless as soon as you resize the image or change output from an epson printer to a different printer or device. Additionally some scanners have 'native' optical or hardware resolutions. These scanners may list these resolutions and are generally obvious. Such as a 1600 dpi scanner might work well at 1600, 800, 600, 400 and 200, but not as well at 500, 300 etc...

Many people, myself included feel it is better to over sample (scan at a resolution in excess of both the source and the intended use) and then nurse the file size down to the intended use in multiple resizing steps using either bicubic sharpen resampling or mild USM with plain bicubic resampling (no sharpening or smoothing). By doing it this way you achieve a couple important goals:1) you capture all the detail that is available; 2) you control what information is 'thrown out' as the image is scaled to its final use size.

 

[tim rudman]

Thanks JD, that's very helpful.
I've still a lot to learn. Would you recommend a good book on (flatbed) scanning? I have a couple of intro level books which I have read cover to cover, but they are fairly basic.
Tin

 

[sanking]

Thanks JD, that's very helpful.
I've still a lot to learn. Would you recommend a good book on (flatbed) scanning? I have a couple of intro level books which I have read cover to cover, but they are fairly basic.
Tin

Tim,

I would recommend Real World Scanning and Halftones, by Blatner, Chavez, Fleishman and Roth. Peachpit Press.

Sandy

 

[tim rudman]

Tim,

I would recommend Real World Scanning and Halftones, by Blatner, Chavez, Fleishman and Roth. Peachpit Press.

Sandy

Thank you Sandy, I had found this with a Google search but wasn't sure how relevant it was from the title (Halftones)?
Tim