I'm not sure about Silverfast's CM implementation and how it arrives at its HDR48bit RAW capture of color negatives since you can't create a profile for them anyway. I'm a bit puzzled why SF's CM setting only gives AdobeRGB and "None" as the choices for the internal profile but embeds/assigns the SF_T Epson Perfection canned profile to all my color neg RAW scans. This often produces inconsistant results with different film brands like oversaturated previews especially in shots of flowers for both Kodak UC and HD400 films but not 17 year old Agfa XRS 100.

I've supplied a screenshot sample to show what I mean. With the scan of the flower shot on Kodak HD400 all I did in PS was keep the assigned scanner profile then inverted, made three more copies and assigned different matrix based working spaces to each, adjusted color & density by setting only the highlite endpoint for all three RGB channels with a 5 level increase in green using the left bottom slider in Levels. Similar previews could be achieved by selecting Enhance Channel Contrast with Snap to Neutrals turned off in the Options dialog box in Levels.

Edited all four individually while Soft Proofing to a Noritsu minilab profile set to Relative Intent. The hue of the flower in each of the assigned matrix working space images could be made bluer by adjusting the blue channel highlite slider but not with the scanner profile which created bad posterization artifacts.

The shot of the old man had an additional simple bowed curve applied all done in the scanner space. No converting on any of the images before editing was performed because it didn't make any difference.

From the different previews shown of the new Kodak and old Agfa neg brands it appears Silverfast's SF_T scanner profile works only with a specific film and scene gamut. Not sure. The flower was shot with a Minolta Freedom Zoom P&S and the old man was captured with a Yashica SLR with a 55mm prime lens on Agfa XRS 100. Don't know if that makes a difference.

It seems assigning a wide gamut matrix type working space to the flower image gives smoother results with less posterized saturation hot spots than keeping the original Silverfast assigned table based scanner profile. Can't understand why it doesn't do this to the shot of the old man because assigning a matrix working space to it gives a dulled rusty brown hue to the preview.

Any wonder color negs are a mystery.

http://www.pixentral.com/show.php?pi...Nvc8swsFA9PSc1