Re: 8 verses 16 bit

[Kennedy McEwen <rkm@xxxxxxxxxxxxxxxxxxxx>]

In article <3C840108.B37A7E04@uci.edu>, David Chien <chiendh@uci.edu> 
writes
>
>  I prefer 16-32 bits per channel RGB, with a total of 48-96-bits total
>when you add up all three, but even something like 12-bits per channel
>with a total of 36bits is better than just 8-bits.

Would you mind pointing any of us to ANY electro-optical imaging sensor 
which actually is capable of producing 16-32 bits per channel?  Real 
bits, please, not a lower 3 or 4 bits of noise - we don't want to 
degenerate this to "why doesn't the Nikon scanner range achieve its 4.2 
dynamic range specification"!   I asked yesterday but you haven't 
produced any info.  Similarly with the "human vision studies" that you 
cited.

Without raw data with a dynamic range which exceeds 16 bits per channel 
there is absolutely NO REASON WHATSOEVER for any software package to 
provide more than that depth of processing.  Without a visual system 
that is capable of distinguishing more than 8-bits per channel there is 
NO REASON to increase output dynamic range beyond that.  What can be, 
and is, discussed is where the transition from hi to low dynamic range 
optimally occurs in any given workflow.

There may well be reasons for increasing the functionality of PS 
operations to 16bits per channel - as has been discussed many times - 
but, so far, anything other than the minimum of what you are asking for 
has not been justified.

I am interested in reading solid justifications for more than 16 bits, 
because the arguments that have been presented to date on this thread 
simply don't stand up to any level of examination.
-- 
Kennedy
Yes, Socrates himself is particularly missed;
A lovely little thinker, but a bugger when he's pissed.
Python Philosophers
-
Turn off HTML mail features. Keep quoted material short. Use accurate
subject lines. http://www.leben.com/lists for list instructions.