Hi all,
I’m building a system for jumping through cues in a long video source. I thought it would be good to build thumbnails out of a texture3D TOP and a geo instancer holding a single 16x9 rectangle.
A table DAT feeds the instancer with one row for every thumbnail, and columns holding position data and which index of the tex3D to use as a texture.
Everything works as expected except that the index has to be expressed in W coordinates normalized to 0-1 - so instead of nice clean integers I have to calculate some fractional value and I end up getting a blend between two frames.
Is there a better way to do all this?
This is part of a larger project I can’t post, but if my question need clarification I can make a sample project.
Thanks,
–David