Fog in register combiners using nvparse

Can someone tell me how to express the fog equation in nvparse?
I came up with this (please don’t laugh, I’m still struggling with rc’s in nvparse, mainly because I can’t find any docs on it!):-
final_product = tex3 * col0;
out.rgb = fog.rgb * unsigned_invert(fog.a) + final_product.rgb * fog.a;
out.a = unsigned_invert(zero);

This doesn’t work - it comes up with an error.
Please help.
The reason I’m doing this is because I’m doing reflecting bumpmapping in vertex programs/texture shaders, but I can’t get the final output to be tex3*col0 with fog.

out.rgb = lerp(fog.a,fog.rgb,final_product.rgb);

i guess its that, or fog and final_product swapped…

docs are out there enough on nvidias page, just take the time to download all the .pdf files and read them…

Thanks for the answer Dave, but…
No, that doesn’t work (swapping fog & final still comes up with an error).
This is my entire reg combiner script:-

final_product = tex0*col0;
out.rgb = lerp(fog.a, final_product.rgb, fog.rgb);
out.a = unsigned_invert(zero);

I get an error on line 1.
I really can’t find any documents on nvparse with register combiners on the nvidia website or in the sdk. There’s a “nvparse and texture shaders” presentation, but that’s it.

In the OpenGL SDK, the presentation ProgrammableTextureBlending.ppt has some infomation on using nvparse.

I haven’t used nvparse but from you post, using just the final combiner might be the error. Try inserting a stage before the final combiner stage ( that just discards results ).

No, I’ve seen a few examples where just the final combiner is used, and it works fine. If I did this:-
out.rgb = tex0.rgb * col0.rgb;
out.a = unsigned_invert(zero);

… I get the expected result…it’s just when it comes to fog (which I believe is only available in the final combiner) I can’t get it to work.

So that isn’t the problem, thanks anyway.

PH, I could kiss you!
Thank you for pointing out the “ProgrammableTextureBlending.ppt” document, it held the answer within!

The combiner should look like this:-

final_product = tex3*col0;
out.rgb = lerp(fog.a, final_product, fog);
out.a = unsigned_invert(zero);

That’s it, it works!

You’re welcome. I just checked the rc1.0_grammar.y file and as you noted, using just the final combiner is allowed ( the parser inserts the blank stage automatically ).