when i set the stabilization passes to “5” I get an error after the first step of processing. I don’t know if it was the reason here but the processing failed with these error messages:
- Error message from AI engine: model failed.
- Error message from AI engine: download failed.
- Error message from AI engine: model failed.
- Error message from AI engine: download failed.
- Error message from AI engine: model failed.
- Error message from AI engine: download failed.
- Error message from AI engine: model failed.
- Error message from AI engine: download failed.
- Error message from AI engine: model failed.
- Error message from AI engine: download failed.
- Error message from AI engine: model failed.
- Error message from AI engine: download failed.
- Error message from AI engine: model failed.
- Error message from AI engine: download failed.
- Error message from AI engine: model failed.
- Error message from AI engine: download failed.
- Error message from AI engine: model failed.
- Error message from AI engine: download failed.
- Error message from AI engine: model failed.
- Error message from AI engine: download failed.
- Error message from AI engine: model failed.
- Error message from AI engine: download failed.
- Error message from AI engine: model failed.
- Error message from AI engine: download failed.
- Error message from AI engine: model failed.
- Error message from AI engine: download failed.
- Error message from AI engine: model failed.
- Error message from AI engine: download failed.
- Error message from AI engine: model failed.
- Error message from AI engine: download failed.
the video export called “keepheaders” export a video which can’t be loaded in Vegas Pro. the 3 other export option seems to be ok.
on my old hardware mediaplayer PopcornHour A200 (can’t test on something more recent), the one which has no issue at all is the “temp directory” video export.
one other video can’t be played, and 2 others shutdown the video with a small freeze on the last frame, one is a bit longer than the other one but they play ok and the whole video
on windows with MPC-BE :
→ keepheaders stop before the end with a freeze on the last picture during maybe 0.5 ms.
→ with preview can’t be played at all, but i did a second test (maybe a mistake on my side) and the second test work ok !
→ withoutpreview is played ok
→ temp directory work ok.
I spent $119 the other day on a year of upgrades. V3.1.0 and V3.1.1 are worthless to me. What was taking 1 and a half days on V3.0.12 is taking 6 days on these new ones. Why does it seem to be working fine for others but not for me at all? I don’t want to be stuck at V3.0.12 especially with all this buzz of the dramatic speed improvements for others.
maybe stay with the official release versions rather than with the beta version. you can always revert back to an older verison if it worked better for you.
The devs are already looking into the performance discrepancy some are having between v3.0.12 and the v3.1.x releases. Since you aren’t a beta tester, I will paste what they have asked for so they can investigate (rather than just complaints, they need details):
Anyone having TVAI 3.1.x running slower than 3.0.x or 2.6.4 please post the following information below.
CPU, RAM, GPU, VRAM, input resolution, model used, output resolution, fps or slow motion amount, spf/total time taken in both versions.
If the amount is significant please try reducing the Max Memory setting to around 75% and try again.
So there you go. Provide what they are asking for and it will help them look into this.
I don’t think anyone can answer your question, unless you provide the spec. of your computer.
CPU ? GPU ? OS ? RAM ?
I recommend everyone input their computer specification in the forum
→ “Preferences” → “Profile” → “About Me”
So everyone can see your system configuration by clicking your icon.
Conclusion:
Both examples are x2 interpolation.
In RIFE example Frames 1 & 3 are original. So they are intact (model do not output them).
In Apollo model frames 1 & 3 are changed (even thou should not?).
And frame 2 in Apollo (interpolated) is very blurred/lost details.
As I saw Chronos have same issue.
Ofcourse maybe Apollo & Chronos were not trained on animation…
Can it be fixed?
I think interpolation model should not add motion blur effect to interpolated frames.
Looks like its a result of wrong training data (blurred fast movement scenes should not be used as training data?).
v3.1.0 was working fine for me, but this latest update v3.1.1 has really increased processing time. How do I revert back to 3.1.0 until the bugs are fixed?
Hi David - Would changing the Max Memory Usage in Preferences make a difference and is the AI Processor set to your GPU? If you can let us know your GPU & CPU type and send us your logs, that’d be appreciated. You can get the logs from Help > Logging > Gather logs for Support, and attach the log file here. Thank you.
Hi Jim - what GPU does your machine have? Can you possibly send us your logs processing the same video with the same settings from v3.1.1 & v3.1.0? You can get the logs from Help > Logging > Gather logs for support. You can revert to v3.1.0 from the changelogs page.
It seems to have become a bad tradition to ignore any reports of parts of the program not working. No matter what value ADD NOISE has, the result is exactly the same.