Ok, just making this task because my brain isn't here 100%
So, we thought we had figured out how to get images translated, but it seems we interpreted wrong.
We currently have images being pulled into the translation files with
gettext_additional_targets = ['image']
We need to keep this because it also allows translation alt-texts. But it also allows changing the path. @Abella informed me that changing the path didn't work for getting the translated image for him.
So, I did some investigation, and it turns out we need to set this config value:
figure_language_filename = "{path}{language}/{basename}{ext}"
but, this config value will translate a path like /images/file.png to /images/ca/file.png, without any modification of paths required, the stupid thing is that we've been using /images/en/file.png, which sphinx then tries to convert to /images/en/ca/file.png and reverts back to /images/en/file.png as it cannot find any.
So, I think I am going to need to figure out how to mass change all these paths first before we can have translation of images :D