every time you try to use the RHEA model you will find an error
logsForSupport.zip (583.7 KB)
Looks like you have a 1050 Ti with 4GB of VRAM that is below the System Requirements for Video AI. Rhea needs a lot of VRAM to run and it looks to be overwhelming your card.
I reported on the depth of field issues here too:
I presume Rhea requires 8GB+ vram to run as it eats up all 6GB on my card before giving an error.
That was with a 360p input video.
Yeah I recall. I don’t expect them to address it though. It’ll likely be released as-is.
I noticed these areas, circled in blue, was also semi-brought into focus:
The ‘bobblehead effect’ is a deal breaker for me - if they can’t change it I doubt I’ll be using Rhea.
it does recover the details much better but for old VHS to MP4 progressive converted MP4s Iris was not good in handling noisy unsharp parts or had artificial irisses then I had to protheus then I used both but this does NOT read NOT improve the output. I am happy that ths model is sufficient enough to renew my subscription I can make any movie from Interlaced to Progressive with ther tools. It is a bit softening, but that will it make more appealing. My 30 years marriage Video 25 P looks good my 50 p from my son as a baby now 30 especially in the darker phases much better than the original
The 5.2.0.2b with the edit properties is SOOOOO MUCH better than 5.2. even better than the alpha version in speedI set it to FULL no Version has this it used to be MPEG
- Fixed missing swap views button.
- Restored colorspace selection.
- Ensures colorspace correctness for input videos with complete colorspace tags.
in the latest Beta it says this. Does that mean its going to stop giving everyone yellow skin tints? If so, how do you use this tag feature? (assuming they dont mean LUTs)
They’re talking about metadata tags embedded in the input file.
I like the new Rhea model. There are better results with old VHS videos with a lot of noise than using Nyx 3.
Videoproc is 45€ LIFETIME
EXTRAVIP code 45€ → 30€
Refund because of bugs 30€ → 15€
5.2 Beta results are beautiful, but I still need it on interlaced sources, please.
@nipun.nath - can anything be done with the model training to mitigate the depth of field issues I reported?
Can someone please explain to me the difference between Improve Detail and Sharpen?
A very general summary: Improve details tends to make dark lines more pronounced. Sharpen does more what the name implies, but turning it up too much will start to add dot patterns to things.
Good point. Even better would be a toggle switch!
5.1.4 is a little unusable for me since a lot of my process is comparing models and settings for new clips before sending them to render overnight. However, it is impossible to compare in 5.1.4 since it deletes a preview as soon as I create another. I cannot even queue up more than one preview at a time. It is such a bad experience over the previous version I had.
I do not see this addressed for 5.2 - is this still an issue? Was it silently fixed? Is it now a new feature we have to live with (I won’t)?
You can have multiple previews in v5.2
@tony.topazlabs For teams You might want to consider Proxy Generators H265 H264 422Proxy.Than Process to Beter Quality .VideoAI Team Could also Work with Google Drive ,DRopbox,ect.
- Beta is only has an effect (maybe stability issues) on what was changed in the change log, the rest is exactly like the release version
- Beta does not touch your release/regular installation nor it’s presets, they are installed side by side. you can launch them simultaneously.
- as of now, very slow as it was not yet optimized, it renders at x4 regardless to your settings, then bumps it down to your selected size (e.g. x2).
- Answer No. 2.
