POV-Ray : Newsgroups : povray.advanced-users : Gamma issues : Re: Gamma issues Server Time
16 May 2024 15:15:50 EDT (-0400)
  Re: Gamma issues  
From: clipka
Date: 15 Jul 2014 16:44:04
Message: <53c59294@news.povray.org>
Am 15.07.2014 21:20, schrieb posfan12:
> It's the wording of the article that confuses me.
>
> For instance it says, "The title of this page and part of its contents are
> somewhat outdated..." What is outdated in the article? Those are the steps you
> have to take in POV 3.7. The article is meaningless if all you're using is POV
> 3.6.

For instance, the article suggests to "wait for the next beta" for the 
"srgb" family of keywords, which has already appeared years ago.

It also suggests to "use '#version 3.7', and remove any 'assumed_gamma'" 
statement, which will provoke a warning by now; you should instead 
explicitly specify "assumed_gamma 1.0".

> Also, it says, "If you are still using version 3.6, or need more flexibility of
> the gamma parameter:" Again, if all you're using is POV 3.6, then you shouldn't
> be reading the article at all.

You indeed shouldn't use 3.6 any longer. However, some parts of the 3.7 
gamma handling were already available in 3.6, by settig "assumed_gamma 
1.0" - and using the gamma_color_adjust macro mentioned in the article 
instead of the "srgb" family of keywords. The "only" problem back then 
was that this gamma handling was implemented only half-heartedly, 
apparently under the (utterly wrong) presumtion that this would be the 
one and only gamma handling model used in the entire ecosystem in which 
POV-Ray would be living. As a result, nobody actually really understood 
how this gamma 1.0 mode worked (let alone how it was /supposed/ to 
work), and using it was a pain in the ass. Especially when using 3rd 
party input images for textures, people realized that /something/ was 
wrong about the mode; it was probably then when many people started to 
seriously dislike and even distrust this gamma mode, some even to this day.

"All" that 3.7 did was to establish a clear notion of how the gamma 1.0 
model was supposed to work internally; what gamma models were used in 
the outside world; draw a clear-cut line between POV-Ray's "inner world" 
and the outside ecosystem (even where the outside ecosystem might happen 
to use the same gamma model); and last not least implement conversion 
algorithms right at those very borders, with optional tweakables where 
appropriate.

Oh, and then, of course, identify (and in most cases address) stuff that 
just /happened/ to be working better with the old gamma 2.2 (or 1.8 or 
whatever) model, such as deciding /when/ to do anti-alias oversampling 
(solved), or interpolating pigment gradients with strong brightness 
contrast (still pending, as just reminded by Warp). Other - even closely 
related - stuff turned out to be working better with the gamma 1.0 model 
instead, such as actually /computing/ anti-aliasing, or interpolating 
pigment gradients with strong colour contrast.

> Lastly, the article says, "As of version 3.7.0.beta.41, POV-Ray natively
> supports gamma-adjustment of color literals, provided they conform to the sRGB
> standard..." But it does not provide any examples of the new technique.

It does, in the code block right after the colon after that sentence. 
It's what the "srgb" family of keywords does: They auto-convert the 
specified sRGB colour components to whatever internal working gamma is 
specified via "assumed_gamma". (When using "rgb" instead, it is assumed 
that the colour components are already specified in accordance with the 
assumed_gamma setting, and no conversion is applied whatsoever).


Post a reply to this message

Copyright 2003-2023 Persistence of Vision Raytracer Pty. Ltd.