Topaz does an amazing jobs on my old DSLR pics, particularly if i have the RAW files. It does however, struggle a lot with cell phone pics. I have read that a lot of recent cell phone camera “advancement” has come from in camera “AI” processing.
I suspect that this process is destroying a lot of the useful info that Topaz needs to reconstruct low quality photos. Would it be possible / is it sensible to train Topaz “cell phone specific” AI models ?