Hi guys, I keep running into problems when working with images with more than 8 bit channels, like float textures. I was thinking of simple file format which contains the following: width, height, num channels, data_type=8,16,32 int, float, double, Raw Data Would it be useful, or are there any better free alternatives available.

What kind of problems are you encountering, exactly? Why do you think introducing a new file format will solve those problems? What format is the data in that's giving you trouble?

Well I have been dealing with some floating point textures generated in glsl and obviously my code had some bugs in it. I can export the data back but its difficult to visualize since I wasnt using any image library like free image. I then ended up export this float file to disk with the width and height and wrote a program in C# to scale the values from 0 to 255 for display purposes. That did give me some insight. I then thought extending it to other datatypes which are bigger than byte could be useful.