The face situation with Iris is quite interesting: faces have been a weak point for AI upscaling for a while now and I wasn't entirely convinced that we were going to see much progress on that front in the short term.įor my test I used JYH-0001 (starring Yuma Asami). Thanks for bringing this to my attention - the more recent iterations of Topaz Video AI (the 3.0 series) was quite cutting edge and not something I was planning to use until it became a little more stable and it looks like now is a good time to jump in! It's still probably worth keeping 2.6.4 installed as well for backup situations where the current version doesn't work as well. Screenshot is not enf/cmnf related but just to give you an idea. In fact it works great on all lower quality content that needs enhancement. With this AI Model they finally got rid of those creepy faces and that typical "AI enhanced" look.Īlso works great on interlaced and heavily compressed or noisy content. I tried it and it works surprisingly well on this type of content. This model is trained to work on all three categories - progressive, interlaced, interlaced progressive.Note, the model tends to lose details on high-quality videos This model is also designed to enhance low- to medium-quality video, particularly, with noise and compression artifacts.
Real memory used by Lightroom: 1248.2 MB (15.2%) Real memory available to Lightroom: 8175.4 MB Operating system: Windows 10 - Business Edition
0 kommentar(er)
