Man, did you ever read my first reply? Repeating myself, for each blue pixel set it’s new color to the average of it’s adjacent, non-blue pixels. If there aren’t any, make the pixel black. Guarenteed no blue borders whatsoever. It’s been a real pleasure helping you.
-Ilkka
Sorry but i’ve been so busy with the bull****s that WE’VE been talking so i’ve mist it!But now we’re talking like we shoud.You’re way is really good.10x for the help!
This thread has been quite comical to me. 
Don’t you just love it when you tell someone exactly how to do something and they act like you didn’t say anything? I know this just tickles everyone’s fancy here. 
And Mihail, I don’t know if you realize it, but it was your own fault this thread turned out the way it did. If all of these people here are jerks and meanies then every thread on here would be full of bashing and scornfull remarks. I have been around on this board for quite a while so I know how the regualrs are and they are NOT stupid by any means. But when 95% of your posts are off-topic and your replies to people trying to help you are scornfull, then what in the hell do you expect to happen? You sir are acting quite childish. Nuff said.
-SirKnight
[This message has been edited by SirKnight (edited 03-18-2003).]
Let’s sink this load of ****. Everybody answer one question below this topic, and it should be out of the front page in no time.
-Ilkka
Oh and don’t, under any conditions, reply to this thread anymore. (oops, I just did…)
ps. That’s some fast censorship!
Micka…
If you are using mip maps you will perhaps sample the “alpha” fixed image after you fix have been applied. This will possibly result in sampling in “blue” contents into the mip maps where the alpha is not fully zero.
Solutions…
You need to adjust your sampling algo to take alpha values into account for non “bleeding” sampling
or… You can apply you alpha fix for each sub mipmap after you have created the mipmaps.