Hey, I just posted the same thing then read your post, gonna delete it and have only this reply to you.
RTX 3080ti, 527.56 Studio Driver, Window 10 22H2, VEAI 3.0.7.
I also noticed “model failed” when it first started and then it started showing other models (I don’t know .ox meaning though - a newbie here).
It ran for 10 hours and the file didn’t look any different than the input.
Enhancing to same size, parameters based on 30 averages via spreadsheet, +1 noise, +1 grain, H264.
When I first start I don’t do a preview, but rather set to Input, thinking it might be less intensive if it doesn’t have to do that and run a tad faster - don’t know if that’s the case.
Gonna run it again and might even uninstall & re-install 3.06 since this is the first time I’ve ever noticed and the output file looks the same as the input file to me.
With Ampere, VEAI 3.0.6 and .7 tries to load models, that use tensor cores (I think), rt806-8500 ones, but, apparently, fails, and falls back to ox ones.
Still no tender loving care for the Theia models.
Only the Reduce Noise slider does anything, and the wrong thing at that—It sharpens the image.
Went through all my tests: DVD source at 100% scale just trying to clean it up for the Proteus model. On the last two versions I only tested Theia Fidelity, but I also tried Theia Detail this time. It has the exact same issue.
Kind of sad that we have only been able to use a third of Theia since 3.0.4. (That’s when I started testing it. Maybe it’s been broken this while time.)
I’m not entirely certain about this, but from what I’ve observed, FFmpeg appears to put a zero-length output file in the source/destination folder. Later, it fills the completed video file into it. At that point, (I believe,) it hands the job back to the GIU which renames it to the name you originally gave in the export dialog. - That is why it doesn’t get renamed when the GUI dies and FFmpeg continues and completes the job.
Yes, it starts with zero length and get filled during the ongoing process. One can open this temporary file at any time with VLC in order to check what finally will be delivered.
Are you sure that’s not just Windows NTFS behavior? Whether it be .hevc, .mov, or whatever, all these files (seemingly) start at 0 length, until you press the refresh button (T5) inside the folder; and sometimes several times.
Interesting. - Since I mainly Export long-length video, I don’t spend a lot of time watching the screen; it’s like trying to watch grass grow. As such, I don’t get to see much happen on the screen unless I’m fortunate enough to be sitting there as a job completes.
Also, because a lot of my latest work is being attempted on Beta releases, much of it never gets done.
In fact, I’m going to put a lot more of my work using the newer “official” releases.
Is there a way to see a preview of the video being processed as it is processing, like in the past? I want to make sure everything is looking good while I export and I don’t see any live preview for the export.
Also, I see that AV1 is supported. Has anyone tested that? What kind of card do you need to export in av1? From what I am reading av1 is supposed to be better than h264 or h265? Same quality, but lower file size?
Getting strange behavior from the Enhancements section. Several time attempting a preview resulted in a “model failed” message. In the case of using Theia, i started, flickered failed several times and then started creating the preview which appears to have completed correctly. - My default GPU is an RTX 3090, it it possible it failed using Nvidia and retried using an alternate driver?
Updating this comment… In this release Proteus has become VERY fussy about what it will actually process. - There are so many variations in what ‘standard’ video formats are currently made up of, including what was inherited during their evolution. - The odd thing is that it gets picky about processing files that were originally created by TVAI itself.
i have a question, how does somebody contact topaz with their issues via emai, or even dm on here, as i cant find it anywhere.
also the only way im ever able to locate the community on here is via google or via the replies on the topics ive responded to. as the community tab isnt on the main topaz site.
