der_ton wrote:
Nice job!
Thanks!

der_ton wrote:
From one single test I noticed there is something wrong with the model's orientation, but apart from that it worked.
I'll take a look at it.
der_ton wrote:
I took a look at the code, and some of that looks familiar, but that's probably because some stuff has to be done the same way. But if you took some inspiration from my exporter, credits would be nice
I have to admit that I took a look at your exporter when I came to a point the bone orientations were not correctly. I found the solution in the:
if (q.w < 0) do ....switch the sign of the quat components
and that was before I read the post of bozo where he exposed this solution so I guess I should credit you. And well, if you see something more, similar to your code let me know, please. Besides that, I don't remember more things.
I have credited people in the help file, you included but only for the md5 viewer so I will credit you for your exporter. I'm not here to make enemies.
Said that. I want to comment something about your code. You have a function that converts a left handed matrix to a right handed one. I thought that MAX stored rotations in a right handed way. So is that for something special or to assure that the matrices are correct?. I say that because I don't convert the matrices to right handed and all seems to work. But who know, always exists a case that destroys all your theories.
One thing more. Your exporter takes less time to export the geometry than mine. As I said before, that's because, to avoid duplicated vertices, I have to, for each face index, loop through the final list of vertices and compare vertex and texture coordinates for each vertex in the list. Do you do something like that or do you export duplicated vertices?.
That's all for now. And sorry for any inconvenience.
Greetings.