ðòïåëôù 


  áòèé÷ 


Apache-Talk @lexa.ru 

Inet-Admins @info.east.ru 

Filmscanners @halftone.co.uk 

Security-alerts @yandex-team.ru 

nginx-ru @sysoev.ru 

  óôáôøé 


  ðåòóïîáìøîïå 


  ðòïçòáííù 



ðéûéôå
ðéóøíá












     áòèé÷ :: Filmscanners
Filmscanners mailing list archive (filmscanners@halftone.co.uk)

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[filmscanners] RE: 8bits vs. 16bits/channel: cantheeyeseethedifference



> From: Austin Franklin
>
> 1) How does PS convert from 16 bits to 8 bits?

It truncates.

> 2) If in fact it does choose the LSB algorithmically (meaning more than
> simply lopping off the lower 8 bits), is what it does dithering
> or not.  If
> it is not random, it is not dithering.  If it is simply rounding,
> it is not
> dithering.  I don't know what it's doing, and Roy says he can prove it's
> dithering...so we'll see.

The term "dithering" doesn't only refer to random or pseudo-random
processes. You can dither a 1-D signal by adding a bit-reversed counter to
the high-resolution value before truncating. In imaging, this is equivalent
to that patterned dither used by the old Windows 16-color VGA driver. But
it's still called dithering. It isn't the randomness that makes it
dithering, it's the fact that the energy is concentrated into the high
frequency end of the spectrum, where it is less objectionable. That's true
of Floyd-Steinberg, delta-sigma, patterned dither, random dither. It's all
the same thing.

> 3) If it does in fact "dither", is it even significant?.  I
> contend not, as
> no image in the real world will have 127.5 across a significant
> area to make
> a visually appreciable difference.

Well, the first message in this thread was a claim that the posterization
you get from 8bpc is indeed significant. The second message was my response
saying that it isn't a problem in real-world images, because of the
dithering resulting from the inherent noise. So in that sense we agree: you
won't see an image that has 127.5, or for that matter 128, across a
significant area--what you'll see is values jumping around among 126-129,
but averaging out to 127.5, or something like that.

--

Ciao,               Paul D. DeRocco
Paul                mailto:pderocco@ix.netcom.com

----------------------------------------------------------------------------------------
Unsubscribe by mail to listserver@halftone.co.uk, with 'unsubscribe 
filmscanners'
or 'unsubscribe filmscanners_digest' (as appropriate) in the message title or 
body



 




Copyright © Lexa Software, 1996-2009.