Gigapixel AI 7 Suggestions

  • When processing a queue, indicate how many items are in the queue, how many are “Done” and how many remain “Queued”
  • Improve speed with better multithreading. Currently the images are processed one by one, leaving the CPU and GPU with lots of idle time between images.
  • Allow option to output to a different parent folder, but preserve the folder structure of the original folders when images in the queue come from multiple folders. In other words, create the subfolder in the target parent so that images from multiple source folders will end up in their separate target subfolder in the target parent.
  • Improve app responsiveness while processing a queue. Currently toggling into and out of the app takes a few seconds and the “Done” links do not readily accept clicks.
  • Allow resizing by the “Long” or “Short” side instead of just height or width. Example: Setting a “Long” side resize to 1920px would ensure both portrait and landscape images end up with consistent upscaling ratios - a 800x600 image would become 1920x1440, and a 600x800 image would become 1440x1920.

Thank you for these suggestions! It would be very helpful if you made a new idea post for each idea so that we can review each one and allow other Gigapixel users to vote for the features that they are most excited about.

Personally, my favorite surprise would be if uncompressed models were included in one of the upcoming versions, specifically the models we tested in the 6.3.0 beta (early beginnings) - 256x256 dimensions. These are the precursors to the standard models and ghq. These models truly have enormous potential. They are much more robust. I sincerely believe that the Topaz team is developing and improving these models further, but simply, since the rest of the market does not offer anything more advanced, Topaz is releasing what is in line with the circumstances. Perhaps they should introduce an additional subscription for premium models or cloud processing, as was once the case for Video Enhance in the alpha stage.