Primary aim of the model is to improve human faces.Please take a look at the Video Roadmap Update for more context on what we’re focusing on right now. Frame number preview length may shorten on app restart.Videos with mismatched metadata and streams will display incorrect duration.Fixes issue with estimate not working for certain image sequences.Maps all audio streams with metadata to output.Adds Recover Original Detail option for Proteus, Iris, and Artemis models.Adds a new enhancement model Iris for face enhancement and/or improving low-to-medium quality progressive/interlaced videos.The technology is 100% good but the interface is assumning too much situations, and not let to avoid a lot of problems.A new release of Topaz Video AI is now available. Those are my thoughts after maybe 4000 videos processed. Because maybe the picture is clean… but with deinterlacing artifacts, or maybe is well upscaled, but the aspect ratio is not the right and the picture is squeezed. If we mix concepts, 50% of times the results are not improving the original. The deinterlace have to be only deinterlace) I think the problem is the interface should show different stepsġ, input pixel aspect (if not 50% of times we are defforming)Ģ deinterlace (independent of enhancing method. If we have to deinterlace and upscale a video with grain we have to use Dione alghoryhtm that is so bad with the grain… And if we selet proteus or Artemis, the upscale is ok and the degrain is ok, but the deinterlace is destroying the results You are assuming that any video have one problem.īut 90% times the videos have a lot of problems. I want to explain better the issue with the deinterlace.
0 Comments
Leave a Reply.AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |