Hello,
Is there plans to include more control over film grain? The previous ver 2.6.4 GRAIN: Amount & Size options. I enjoyed that feature to better match the style of film I am upresing.
Thank you!
“Encoder error” can now be made when “H265 Main (NVIDIA)” or “H265 Main10 (NVIDIA)” is selected and processing is started.
NVIDIA drivers have been updated to the latest version. Imgur: The magic of the Internet
And there is still a problem with the screen display size being incorrect during playback. I could not tell the difference between the original video file that causes this problem and the one that does not.
Thank you for the great advice !
H265 Main (NVIDIA) does not work stii with the latest NVIDIA
drivers(ver.526.98 2022-11-16).
but AI model runs with “H264 High (NVIDIA)”.
Error occured at
VIDEO TYPE : Interlaced Progressive
AI MODEL : Dione:Robust
GPU : GTX1080
Tried out my tests with Theia again:
The ‘compression’ parameter still makes no change at any value.
The ‘noise’ parameter is working different now. It seems to be enhancing details at 100. I tried negative values too, and they make the output more blurry.
The ‘blur’ parameter does nothing at any value.
The results I got with Theia ‘Reduce Noise’ 39 at 100% scale then Proteus Auto to FHD were really good looking. I’m pretty sure that ‘Reduce Noise’ is mapped to the what the ‘Sharpen’ slider should be. And the other two sliders are mistakenly disabled, still.
I bought it from the 3.0 official release
When I found out in the past few days that 2.6.4 does look more natural and the colors are less distorted under the “proteus relative to auto” I have been using, it made me a little nervous
Did a lot of testing to make sure the really old version is better haha
Just after the test, 3.05 finally became natural! More details now, awesome
*google translate
Hi! I miss custom windows scaling from 2.6.4 It follows what you set in “change the size” of apps. Now one have to be Adobe Premier-close to the screen At least in output tab.
Nothing has changed for me.
There is still this weird deformation of the video in all windows.
Still when I click Export, the input and output windows are black, the program doesn’t show the currently processed file.
Hi
After test all 3.0 versions, 3.1,2,3,4,5. The experience is worse than 2.6
For example. Dione deinterlace is not working well. With 2.6 I load a video interlaced and the output is progressive. With 3.5 the output is progressive with a lot of deinterlacing artifacts.
The preview is painful. since before you can see how the program is creating the frames. Now the preview runs suddenly after created some frames. This have not any advantage, and a lot of inconvenience.
The stability is 100% worse. The program is closing when he wants. 2.6 much more stable.
In my oppinion you are changing the interface too much and each time we have to change our mind, but there ara any advantage,
The real problems still there. The deinterlace tool, should be the first option. The pixel aspect ratio input change, should be the second option. The program is assuming a pixel aspect ratio that in a lot of cases is defforming the original picture.
Thanks
I want to explain better the issue with the deinterlace. You are assuming that any video have one problem.
But 90% times the videos have a lot of problems.
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
I think the problem is the interface should show different steps
1, input pixel aspect (if not 50% of times we are defforming)
2 deinterlace (independent of enhancing method. The deinterlace have to be only deinterlace)
3 denoise method
4 upscale method
If we mix concepts, 50% of times the results are not improving the original. 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.
Those are my thoughts after maybe 4000 videos processed.
The technology is 100% good but the interface is assumning too much situations, and not let to avoid a lot of problems.
Best
At least in output tab.
