Image / batch processing

[aaron1138]

I have a use case for batch processing large numbers of high-ish resolution images with shaders to perform 2.5D manipulation of voxel data stored as 2D PNG image stacks.

As far as a feature request for ShaderGlass, I would like to be able to use the CLI to call ShaderGlass with a shader preset (e.g. a slangp assembled in RetroArch), specify a source image with image format information and an output location / filename, size, image type, and format. It would also be helpful to be able to specify no window / offscreen processing, but this falls under "nice to have".

The format portion is also important as I am working with single channel grayscale images. I've noticed the way these PNGs are read by ShaderGlass and then shaders applied leads to a misinterpretation of the channel assignment (e.g. output is RGB as an all red image). RetroArch's image loading does not do this, so I assume it is a difference in image loading libraries or configuration.

It would be great to be able to feed a whole folder in and ShaderGlass to reuse resources to iterate through the images, but it would be more than enough to simply directly process and capture output one image at a time using external scripting.

If there is another tool better suited for this job, please let me know. I've looked a bit and not found anything quite suitable.