
Too much extra stuff in the reset section of the faders all work ok now.
[remove
mymix.amount=0;
mymix2.amount=0;
mymix.type =0;
mymix2.type =0;
i'll upload new copies]
again at http://forums.winamp.com/showthread.php ... did=157647 i think they are the correct versions now.
So all fine apart from when blending to certain scenes.
These scenes all have one thing in common
BUFLOAD bload();
BUFSAVE bsave(m);
in them.
all other scenes fades to these scenes work fine.
More details:
Now the last fade with my fader
-----------------
else
{if (mytime<5.1) {mymix4.amount=(mytime-4);}
---------------------
where
MIX mymix4(mymix3,SceneTo);
and
mymix4.type = 0; //normal alpha blending.
so it should blend from mymix3 TO SceneTo on alpha blend from 0 to 1.
but what actually happens when the SceneTo contains the BUFLOAD stuff is a strange blend from the mymix3 (which i belive is now the same as inverted default texture for teh fade) and alpha blends from 1 to 0. the alpha blend is the strange bit.
Why is it alpha blending backwards?
there is nothing in my code suggesting this. It should never perform thsi direction of alpha blend from my code.
It is unlikely to be video memory etc on my machine as it exhibits this behaviour on really low pbuffer values like 128x128, etc.
So somewhere I expect there is a bug.
Now what makes this really interesting is that i beleive the ATi radeon card doesn't work for some/many scenes. Now I beleive the scenes that failing the same ones that fail for this blend type.
Scenes with
BUFLOAD bload();
BUFSAVE bsave(m);
in them. All the ones on teh website that say nt working with ATi have this in.
But as I have no ATI card I cannot check this. But if it is the case there maybe something amiss with buffer load implementation in R4. And it can be tested now with these faders on a working nVIda card.
It could just be a coincidence but they appear to be linked together. MAybe we are closing to resolving the ATi problem also.