[codey (old)]

Hi,

I got a problem with r_overBrightBits "0" and shaders with partially reflective surfaces using alpha channels. Look at these screenshots
This is how it looks with 1



and this is with 0



You see how the texture surrounding the glass parts gets very much brighter. Here's the shader which I think is the one used in Casa for windows:

textures/codey1/storage_win
{
    {
        map textures/codey1/storage_win.tga
    }
    {
        map textures/codey1/oilglas1.tga
        tcGen environment
        blendFunc GL_ONE GL_ONE 
    }
    { 
        map textures/codey1/storage_win.tga 
        blendFunc GL_SRC_ALPHA GL_ONE_MINUS_SRC_ALPHA 
    }
    {
        map $lightmap
        blendFunc GL_DST_COLOR GL_SRC_COLOR
        rgbGen identity 
    }
}

As I never have gotten my heads round the fairly mathematical explanations in the q3map2 shader manual I thought I'd ask if somebody have a fix for this before I start long trial-n-error session.

BTW - I understand that having it at 0 gives a finer colour range but it also takes out some 'knack' from the picture. Look in the lower picture how the clouds get much less punch and there's hardly any zing at all to the painted high-light / reflection in the pipe on the cistern. And screenshots come out much much darker, I had to do a very course adjustment to gamma in Photoshop to get the 0 image to match how it looks in-game, this is how the original looks:



cheers

[Hybridesque (old)]

I think the default for ETUT is going to be the same as it is in Q3/Q3UT3; overbrightbits 1 & mapoverbrightbits 2. After some testing; RabidCow, Tub & BladeKiller found it was still better than having it set to zero.

[codey (old)]

Thanks Hybridesque, that means I can focus the afternoon on other stuff.

Last info I found was that it would be 0, would be good to have a definitive word on this though.

cheers

[Hybridesque (old)]

I got the info a good while ago. You can take it as a given I suppose, and I'll take the fall if it's wrong.

[codey (old)]

OK - I'll take your word for it. Just for your info here's the last comment I saw on the subject;

Quote

Do the lighting for mapoverbrightbits 2 and overbrightbits 0 in q3ut and you should be fine.

In this thread from May 2nd

cheers

[Hybridesque (old)]

I've got something in my IRC logs somewhere, where TwentySeven was telling me why they decided to stick with obb1, mobb2. I'll dig it out for the benefit of everybody.

BladeKiller & RabidCow did some testing a while back. With RabidCow being rather clued up on Photography, their reasons for keeping obb 1, mobb 2 as it stands do make sense once you get into it.

[NulL]

add rgbgen identity to all the stages, they may be defaulting to identity lighting.

http://shaderlab.com...ch6.html#rgbgen

edit:: That might just make it bright all the time, but let us know.

[codey (old)]

Thanks for the suggestion $NulL but alas, no difference. I tried both rgbGen identity and rgbGen identityLightning on all stages but it looked the same. Took a quick look on Casa with obb 0 and it looks like those shaders have the same problem. Guess I could guesstimate a gamma adjustment in Photoshop to get the textures to a near match but it's not a ideal solution.

cheers

[NulL]

rgbgen const ( 0.5 0.5 0.5 )

Might do the trick if applied to the 3rd stage, and maybe the first.

[codey (old)]

Nice one $NulL, that did work! That parameter isn't covered in the q3map2 manual is it?

Now of course with oBB 1 the area gets darker - which again leads back to the question - has it been set for ETUT and to what

cheers