Just a quick question for the TVAI development team…What is the difference between the Recover Details setting in Proteus and Iris versus the new Recover Original Detail setting?
I’m not on their team, but I can explain:
‘Recover Details’ Uses AI to “guess” what the smoothed out details are supposed to look like. In my testing, it most noticeably darkens small detail outlines.
‘Recover Original Detail’ Blends the original source with the enhanced output.
There is no difference. For all models where it is available, it blends the input into the model’s output.
Thanks! This maybe explains why jacking up Recover Original Detail helped so much in countering Iris’s tendency to reduce and soften details. When I tried the same thing with Recover Details, it started reintroducing noise (or so it seems).
After update to [Topaz Video AI v3.3] we have the option to add detail and recoeer detail.
My question is about the Proteus Model, How can i get the same results as prev to the update. Now I am having very bad outcome using the Auto Parameters in 4.0.3, comparing to y results while using 3.0.5.
Any idea if the parameters Add Noise and Add Details matter? Even changing here I am not able to replicates same results as Auto.
I am having the same on Dione TV (v4). up to v3.0.11, everything was good. Since v3.1+ things got worse. it cleans up less noise compared to previous versions. that is why I keep a copy of v3.0.11 installed as well along side the latest one
I cannot reinstall my older version, I have a MAC.
I use Recover Details sparingly with close-ups of people (and of course with 0 noise reduction). I use Gaia a lot, because it protects the existing details and seems to add some new details while sharpening the image, esp. when its run as first and second enhancement.
