wrote in message
I’m currently editing my D70’s NEFs in Photoshop CS2 (including downsizing & adding canvas), saving as 16 Bit TIFs and printing on my dye sub printer. My last step before saving is sharpening. I’m thinking that since my dye-sub printer only prints in 8 Bits, that I might as well save the TIFs first as 8 Bit files since it will have absolutely no effect on quality (and save space which really isn’t an issue for me). Does my argument make sense? If it does, at what point should I convert to 8 Bit (immediately upon opening NEF, before resizing, after sharpening??? It would seem that after sharpening would make sense since 16 Bit TIFs are converted to 8 Bit during printing anyway, right? Thanks for suggestions.
The bad news: not only are 16 bit TIF’s converted to 8 bits before printing, but an additional conversion to your printer’s profile is performed on the converted 8 bit image. In addition to adding noise when 16 bits is converted to 8 bits, each of these two profile conversions results in addition of noise to the bottom bit. Effectively you are printing closer to 7 than 8 bits per channel, with the bottom bit being noise. As of CS2 at least some of the dithering is controlled by a setting in Color Settings, but in CS and earlier versions, the noise was added regardless.
The good news: for color work, even 7 bits per channel produces excellent printed output.
I would strongly advise you to experiment and come to your own conclusion based on your own images. My strong prediction is that you will see no difference in the appearance of the final printout, even if you convert to 8 bits out of ACR and perform extreme edits afterward. Prophoto RGB, and 1.0 gamma images are exceptions to this, and they do benefit from 16 bit per channel representation.
I realize this goes against the conventional wisdom. Most of that wisdom is based only on verbal theory. I’ll be happy to discuss this with anyone, provided they illustrate their statements with actual images. —
Mike Russell
www.curvemeister.com
