Am 23.08.2015 um 23:52 schrieb "Jörg \"Yadgar\" Bleimann":
> Hi(gh)!
> 
> On 23.08.2015 04:56, clipka wrote:
> 
>> If you're not happy with the look of the scene, indeed use "srgb"
>> instead of "rgb" to combat washed-out colours, and tweak the lighting to
>> combat overly bright scenes (if that's still an issue then); also, you
>> may want to reduce some materials' "ambient" settings or reduce
>> "ambient_light" (or use radiosity instead, which is a better idea
>> anyway).
> 
> Whatever values I tried for "s", I got completely weird hues - a green
> "grass" texture changed to dark blue or magenta!

What do you mean by "s"? And can you provide examples with a bit of SDL
code and the resulting imagery?

Green (i.e. a colour with the "green" channel having the highest value
of all channels) should never turn magenta (i.e. a colour with the
"green" channel having the lowest value of all channels), no matter the
gamma handling (as long as you're not trying pathological stuff like use
negative gamma values or some such). Therefore, if you actually see such
symptoms, my next best bet would be that there's something much more
fundamentally wrong about your POV-Ray binary and/or your hardware.


>> As of POV-Ray 3.7, the #version directive is not only mandatory, but
>> /must/ be the /first/ non-comment statement in your scene to ensure
>> flawless operation. And you do get a warning about that (if you'd care
>> to read it ;)).
> 
> Yes, I added #version 3.7, but it didn't help either...

What do you mean by "didn't help"? If "help" = "fix the colour issues",
then you're approaching the situation with wrong expectations. If "help"
= "provoke the colour issues even without including 'screen.inc'", then
I'm pretty surprised it didn't.

Specifying the "#version" directive should do no more nor less than
"stabilize" POV-Ray's behaviour with regards to whether "screen.inc" is
included or not (and along the way suppress a warning you should get
otherwise). If the "#version" directive achieves that, it means that I'm
on the right track as to why the include file makes a difference (please
let me know whether that is the case), and also confirms that the colour
changes are probably related to gamma. If it doesn't, it means something
entirely different is at work.

Post a reply to this message