Original Post
Hey,
Does anyone know the exact algorithm used by directx to compress dxt1 textures? The algorithm i''m using is the one specified in the DX8.1 documentation, but produces different output to that of DXTexTool. The differences seem to be in how the DXTexTool chooses the extreme colours. eg. i have a full white pixel in a 4x4 block, but the high colour that DXTexTool chooses is a light green colour (there are some green pixels in the block). According to the DX doxs, the high colour should be white (as it is the highest in the original image).
Alternatively, is it possible to use D3DX functions to convert between various texture formats without initialising a DX Device?
Thanks,
CYer, Blitz