Digital BW, The Print

on 6/29/02 1:54 PM, tboleyyh wrote:

> However since gray 2.2 does not map precisely
> to it without moving some tones, I would use Adobe instead if it doesn't
> screw up your current results, and I'd certainly turn that stinkin dither
> off. Better yet, stick to 16 bit, sorry Todd.


Ouch. Double ouch!

Good post, as always, Tyler, though I'm not sure I agree with the bit depth
conclusion. (Surprise, surprise <g>)

Two things are missing from your experiment, A) that doing the conversion in
16-bit mode improves the situation, B) that the improvement shows in print.

First off, let me restate my position on 16-bit editing. I just think 8-bit
editing is too often looked at as the primary cause of posterized prints and
I think that's more attention than it deserves. I still believe that.
However, I don't walk a tightrope without a net, so I scan in 16-bits, work
my images globally in 16-bits leaving local edits with masks to 8-bit mode,
and always have a 16-bit backup close at hand.

Anyway, post encouraged me to see if it's time for me to rethink things, but
based on my tests (surely NOT exhaustive), I'm not compelled to.
Furthermore, and consistent with my recommended approach to making bit depth
assessments, I'd put a portion of what little money I have to bet that if
Jim does everything he's doing now in the same way--except for maintaining a
16-bit workflow--he will still get the same result in print. My bottom line
has always been the print, not the histogram.

Now, back to the 16 bit for conversions...Since I don't have your test
pattern I just opened a 16-bit image with a predominance of dark tones (as
though a broody guy like me has any other type of images ;-)


The image was 16-bit grayscale, gamma 2.2, which I then downsampled to
8-bit.

           

            Mean     St. Dev.    Median

Grayscale
g 2.2       

16-bit:     86.40    68.42       84
8-bit:      86.40    68.43       84

16-bit
to sRGB:    85.82    69.81       83

8-bit
to sRGB
dither:     85.83    69.80       83

8-bit
to sRGB
no dither   85.78    69.74       83

The conversion to sRGB is no better in 16-bit mode than 8-bit with dither.

16-bit
to Adobe:   86.40    68.43       84

8-bit
to Adobe
dither:     86.40    68.42       84

8-bit
to sRGB
no dither:  86.40    68.42       84

The conversion to AdobeRGB is no better in 16-bit mode than 8-bit with or
without dither.

I also tried all of the above with different rendering intents and color
engines, neither of which changed any of the above numbers.

So, what do I take from this? A) it matters more what spaces we convert
between than what bit mode we do it from. B) for my type of image the dither
seemed only to help. C) I don't think Jim's problem is due to his
conversion, I think it's in his print-driver/driver-settings.

Todd