| [Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] |
Sorry if I missed a post already covering this, but here goes. The printer driver cannot accept a 16 bit per channel file and could not put that level of subtlety onto paper anyway, but having >8 bits per channel in your input can be very useful. This is especially true for negative film, which contains a greater range of brightness from the original scene than transparency film, albeit compressed into a *smaller* range of brightness on the film itself. It's even more useful for high end scanning back electronic cameras, which can capture even more dynamic range than negative film. (Stephen Johnson, for instance, has been documenting the national parks with digital cameras that provide up to 6000x8000 pixels, 142MB files with more than 10 stops of exposure latitude.) If you only wanted to linearly compress this huge range into the restricted range that will end up on the print, there'd be no problem -- if you got a good medium-tone scan, you'd always be fine. But often you have a very low or high key image, or you have a high contrast image with critical detail in both the highlights and shadows. 16 bits per channel images let you see that detail; they allow more radical levels and curves adjustments without posterization. (Radical levels/curves moves, especially on 8 bit data, can also significantly emphasize the appearance of film grain and scanner noise). But what about scanner drivers that don't output 16 bit files? And what about the larger file size and restricted editing capabilities in Photoshop's 16 bit mode? MY RECOMMENDATION: 1. If you have a >24 bit scanner that only outputs 24 bit files, use the scanner's levels or curves adjustments to get the brightness range of the image into the ballpark you want. For instance, if you're going to bring out details in deep shadows, do that boost here. With a properly written driver, this means that the data will be manipulated with 10 or 12 or whatever bits of precision per channel and *after* the adjustments have been applied, it will be scaled to fit into 8 bits. Do the finer adjustments and all the editing that doesn't involve major levels/curves adjustments in your image editer. You'll get better results than taking an unadjusted scan into Photoshop or whatever and making those big levels/curves moves there. 2. If your scanner will output 16 bit / channel files, *don't* use the scanner's adjustments. Do a straight scan. If you have a profile for your scanner, use it (by starting with a profile to profile conversion on the 16 bit/channel file if your scanner, like most, doesn't embed profiles). Do your major levels/curves work at 16 bits/channel in Photoshop. Save this file if you have the inclination and disk space, and then convert to 8 bits/channel for the rest of your editing. This will generally allow you to print, then go back to the 8 bit/channel file for tweaks, and re-print, without seeing any artifacts from 8 bit editing and without paying the memory and computation time costs of always working on a 16 bit file. Russell Williams -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Please: Stay on topic. Trim quoted messages. http://www.leben.com/lists for list instructions.
[Photo] [Yosemite News] [Yosemite Photos] [Scanner] [Gimp] [Gimp] Users
![]() |