Using the Photoshop plugin (and I’m assuming the exporter), its possible to save high-def uncompressed textures as 16fx4 or 32fx4, but when I try to generate a DX10 header (
nvtt::Container::Container_DDS10) using format
nvtt::Format::Format_RGBA, I get an error from the library that the container does not support the format.
How would one go about saving floating point textures with the nvtt library?
Really appreciate any advice!
I was actually able to figure this out by scanning the web. Here is an example of how to export a 16-bit red+green floating point image:
compressionOptions.setFormat( nvtt::Format_RGBA );
compressionOptions.setPixelFormat( 16, 16, 0, 0 );
I’m still having some strange issues working with floating point textures using the library. One thing I’ve noticed is that I have to avoid using the DX10 header to save them without errors. But even then, the resulting image appears mangled when I try to save floating point textures with anything less than 4 channels, as if the pixel information is crossed. For some reason, 4 channels (16f,16f,16f,16f) seems to work correctly, but not 1 or 2 channels. If anyone has any experience with this, please let me know. I will edit this again if I find a solution.