I have a video that is slightly out of focus. Setting focus fix to normal doesn’t seem to do anything. Setting it to strong fixes the focus, but the resulting video is very blurry. What is the correct way to use focus fix? Do I need to run it back through and recover details?
All fix focus does is downscale the video before running the AI on it. This can help make the video more sharp by reducing the amount of pixel information in the video. It’s a way of tricking the AI into treating the video as a lower quality source.
Fix focus is ideal for videos that were upscaled with more common filters, or no filters at all. Otherwise, there is that trade-off of losing pixels from downscaling. If your video is focused on the wrong thing, this could help that, but the wrong focused thing will lose detail.
Yes, you need to treat the video as being a smaller resolution source. So like if your video is FHD, you need to treat it like it’s a DVD, and apply whatever settings you would use for a DVD of that quality.
1 Like
It does downscale and then upscale to the same resolution. I don’t quite understand why they lock the resolution selection. So for me often this makes no sense, I can do dowscales with other tools where no upscale again happens, then import the downscaled one into TVAI, this also tricks the AI into lower quality source, but then I can upscale it with a model to any resolution of my choice.
1 Like
