-
Notifications
You must be signed in to change notification settings - Fork 80
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
AGC code is affected by wow/flutter #493
Comments
Yeah, I think I got one too, if you need a further example. Testing the new build now. |
If it's not fixed, that'd be nice, thanks! |
Ah, yeah. I'm afraid I'm still having an issue: https://drive.google.com/open?id=1KO0f3-ysRatvxDON7b-_B0EsSWdErvk_ Clip is a little longer than I intended, see issue #495. I meant to just show the cloud section after the "Garrrr Video" screen. |
frame 393, timecode 00:47:48.01 does appear affected. be sure the trailing _ is in the URL if you have trouble downloading toynbeeidea's sample. |
Thanks for the tip about the _, that got me just now... Going to look at two changes: Changing the AGC on first fields only, and keeping a running average (which should cover wow/flutter on CLV disks as well) |
Add some notebook enhancements and work on #493
I ran the entire video through last night. Seems better now as of 9ed93a0. There's a bit of a luma spike or something during the cloud sequence, but I'm inclined to write it off as a flaw in the source. Going to leave this open for the moment, as I gather that this is a partial fix? |
Sounds like it's due to the wow/flutter affecting the first field on one frame, causing the AGC to adjust both fields just for that frame. In cases where there's a lot of frames like that, might be best to use the --noAGC option, until AGC is a little smarter. (maybe a rolling average, ramping up if it changes rather than spiking on single errant frames) This is also something that might be fixed with a TBC-based AGC in once VITS is better worked out. #326 #490 |
Ran it through with --noAGC, and yeah, no more luma spike - thanks! |
Frame 393 seems to be due to a dropout in the hsync, more examples on issue 503. |
Here's another interesting case. A frame goes bright with no obvious dropouts. A decode in April had an audible pop here, but no problem in the audio in the current commit. Wow/flutter related I'm guessing? |
Can you recheck this? It should be better now. |
Checked pop.ldf.zip - doesn't sound great, but no pops, so closing |
@Gamnn found an issue where the two fields luma didn't match up after the AGC change. Oops.
The text was updated successfully, but these errors were encountered: