I meant FFV1 8K output YUV 4:2:0 10b, it produces errors either “Resolution error” or simply “Error”
This is an issue with a constant that has been set 10 years a ago
I sent some explanations here
I know ther’s a .json file to configure the encoder, but thing the patch should be done by topaz in some definitive ways, so the info hasn’t to be grabbed here and there from the forum
Since the latest GUI changes I can not figure out what the original pixel width and height of the source video is. Except in the Preview info line. But it is nowhere shown in the resize output resolution dialog. Why does the Enhancement dialog always revert back to Proteus? What is this icon for:
Isn’t it recommended model for upgrading video stream ?
I did notice that it changes depending of the resolution and quality of my file (suggested Nyx and Gaia)
Is this the same issue I noticed? I set a video to Rhea, set all the sliders, ran a preview, and when I clicked the original video again after having clicked the preview, the enhancement was back on Proteus but the sliders were where I had set them with Rhea.
My base is taken from a recent movie where I found these settings almost by chance.
So it became a base for all my “upgrades”. For the current test, much older, the movie being 95% of the time in dark scene, the real lack of details due to age and compression does not give such a good result, but, it is still clearly better than any other model, and I repeat the quality is constant, there are no rotten teeth or disgusting eyes.
So this time I used Rhea
The hair is also excellently restored. We can even see the eyelashes and eyebrows!
EDIT : Just noticed that 1080p is max pic on forum …
i know there’s very few mosquitos here and there, but they are very small, like 5 pixels, in a 8K frame, with a downscale and some SSIM compression, it will be done.
Moreover when animated your much rather take care of sharp edges than tracking a mosquito, I can’t see them, and I am not very tolerant toward this ^^
Is it a new behaviour with the beta ? if not I suspect bad handling of metas, metas that does not exist in such old gen video.
I did realize it with the patch I’m coding, in fact there were almost nothing ratified as a norms in MPEG1/2 gen.
No, but I have a video projector, SXRD matrix with an X1 processor forming an image of 300x170cm, on which we see the slightest defect, so yes 8K is very useful to me. If it’s not 8K, I can still go up to 4096x2160p higher than UHD 3840x2160p, so I can “pick” pixels from the movies by removing the black bars
As I said, I’m currently testing either 8K downscaled by the X1, or 8K that I’m going to re-encode in 4K in Lanczos/Callmut
Only a “last gen” console is able to generate an optimal image for me, either RGB (detected in BT.2020) 4K HDR 60fps YUV 4:2:2 or 30fps YUV 4:4:4 and again the visual difference is clear and sharp compared to the “classic”
Sadly Sony removed the 4096x2160p in order to be fully compatible with UHD, but it does not matter as I don’t use it to play videos
@tony.topazlabs : I’m aware that there is a dedicated Linux build thread, but it’s going ignored for a long time now, despite various friendly nudges and attempts to establish proof of life. I feel sufficiently frustrated as a result to ask here - is there any plan to resume Linux builds, or is that exercise now paused/abandoned?
