Ability to post-processing colormap for all images

[yslan]

Often, we are not satisfied with the colormap afterward. For me, I typically pick cmin and cmax based on initial samples and regret it later due to a further developed solution profile.

With rgb in png, we should be able to store up to 3 different fields (suppose they are not overlapping) and store them separately into each color using linear color map and integers between 0-255. Maybe it also needs to store the depth into transparency or something.

With that info, one should be able to reconstruct the image with any colormap and maybe some rescale based on min/max of the solution printed somewhere.

I wonder if anyone has explored this.
If so, what's the proper (fast and portable) tool to reconstruct images? python or imagemagick?

Also, can this speed up the rendering? given that each field can be save into grayscale images?

Thanks.