[FS CMM|Johnnyenglish]

TwentySeven, on 08 July 2010 - 10:36 AM, said:

Are you saving the _B as a 32bit tga with alpha?

I have now... oops. Sorry to have wasted your time.

[TwentySeven]

No problems :)

As for the shading, maybe use patches for those columns?

Trying to get q3map to phong shade them is a drama unless they're ase meshes.

[FS CMM|Johnnyenglish]

TwentySeven, on 08 July 2010 - 02:55 PM, said:

No problems :)
As for the shading, maybe use patches for those columns?

I've found a nice md3 model column for now... but one last question.

What would you recommend for a top-grade, highest visual quality compile for maps using bumpy features? I don't care about FPS and I don't mind having it run over night. I'd just like to see it at its best.

I'm currently using

-fs_game q3ut4 -light -nocollapse -fast -samples 3 -bounce 99 -samplesize 1 -patchshadows -radbump

Which is based on one of Rambetters suggestions. I'm not totally happy with the results at the moment.

Thanks John

This post has been edited by Johnnyenglish: 08 July 2010 - 03:55 PM

[GlaD-slackin]

johnny, personally i like samples 8 and samplesize 8, also play with the -dirty option, i like it(some dont) and remove that -fast! (for a good compile) also phong shading wont work unless you use the -shade option too.

NOTE: phong shading and bump/normal maps dont play so nice, they create wierd, undesired effects when used together in the same shader.

ALSO: Don't know what -nocollapse does, try compiling with then without it and see what you like better.

This post has been edited by Slackin: 08 July 2010 - 06:52 PM

[FS CMM|Johnnyenglish]

Thank you slackin, I've decided to try :

-light -samples 8 -bounce 99 -samplesize 8 -dirty -dirtdepth 130 -dirtmode 1 -shade -radbump

which I thought looked nice on a small sample map.

Unfortunately it's still compiling and has been all night so I can't judge if the results are good on my real map yet.

lol. John

[FS CMM|v3nd3tta]

LUL bounce 99?
you fail. there is no need for 99 bounces (reflections)
just try some less, like 5 to 10.
99 may be overkill in some peole's Opinions, because the bounce will stop at some point maybe and the other bounces will just be checks if there could be more bounces.

This post has been edited by .v3nd3tta.: 27 October 2010 - 02:20 PM

[NulL]

.v3nd3tta., on 10 July 2010 - 09:57 PM, said:

LUL bounce 99?
you fail. there is no need for 99 bounces (reflections)
just try some less, like 5 to 10.

99 is a perfectly valid number. I recommend learning about things before you try to mock people with your half baked thoughts on the subject.

[GlaD-slackin]

yes, i almost always bounce 99 for smoothness in lighting, unless i have a specific reason im not. (certian effect im looking for)

[TwentySeven]

Yeah it'll stop long before it gets to 99. The compilier stops 'bouncing' when there is no visible effect caused by the last bounce.

[FS CMM|Johnnyenglish]

TwentySeven, on 11 July 2010 - 12:01 AM, said:

Yeah it'll stop long before it gets to 99. The compilier stops 'bouncing' when there is no visible effect caused by the last bounce.

Which it does and I'm getting really nice results and it doesn't take too long.

I'm still unsure about the rad building process, what I now do (don't laugh) is run a very simple compile with this as the light stage.

-light -nocollapse -fast -samples 3 -bounce 2 -patchshadows -radbump

That gives me a subtle .rad file.

I then run the compile again with

-light -samples 8 -bounce 4 -samplesize 8 -dirty -dirtdepth 130 -dirtmode 1

and this gives me quite nice results.

But this doesn't seem to be right to me, can't help feeling I'm missing something that will give me better results... anyone?

Cheers John

PS. -fast according to the manual is perfectly acceptable for final builds?