Skip to content
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

1080p > videos are choppy, hiccup every 5 secs #552

Open
Schizo opened this issue Mar 10, 2020 · 5 comments
Open

1080p > videos are choppy, hiccup every 5 secs #552

Schizo opened this issue Mar 10, 2020 · 5 comments

Comments

@Schizo
Copy link

Schizo commented Mar 10, 2020

Some videos (given in the example) seem to be choppy, every 4-5 seconds there is a little hiccup. To rule out network IO, I've run it on a local server but still the same effect. How can we figure out where the bottle neck is? Is there a profiling doc?

@Schizo Schizo changed the title 1080p > videos are choppy hickup every 5 secs 1080p > videos are choppy, hiccup every 5 secs Mar 10, 2020
@bvibber
Copy link
Owner

bvibber commented Mar 10, 2020

Could be related to the 1 MiB read chunk size; at higher resolutions you might be running through that data sooner and it's not providing enough of a buffer. Can you repro this with files from the demo page, or else do you have a publicly-available file to test with?

@Schizo
Copy link
Author

Schizo commented Mar 11, 2020

Yes, I've first reproduced it on the demo-page and then served that file locally and had the same issue. Is there a way to raze the 1 MiB read chunk size? I could see if that would solve the issue.

image

@Schizo
Copy link
Author

Schizo commented Mar 16, 2020

Any updates on this? Or a hint were I could change the cache size?

@yafesdev
Copy link

yafesdev commented Apr 3, 2020

@Brion I'm experiencing similar issues, I've tried to change OGVPlayer.js:2467, adding chunkSize: 16 * 1024 * 1024, (default is 1 * 1024 * 1024). However, that didn't seem to affect the streamed chunk size. Any Ideas?

@yafesdev
Copy link

yafesdev commented Apr 6, 2020

Not sure how the Example VP8 Video files were encoded but after re-encoding Tears_of_Steel_in_4k_-_Official_Blender_Foundation_release.webm.1440p.webm my cpu load wen't down a lot and no more lagging on my notebook.

ffmpeg -i ~/Downloads/Tears_of_Steel_in_4k_-_Official_Blender_Foundation_release.webm.1440p.webm  -c:v libvpx -qmin 0 -qmax 50 -b:v 9M  -slices 4 -g 1  ./demo/media/RED1440pReencoded9M.vp8.webm 

I just added a b:v 9M for high quality and -g 1 to have a keyframe on every frame (for quick seeking purposes), I've tested it both in threaded: true/false, and the playback normalized, without drop frames. hosted from a local server.

However:

On my Iphone 11 Pro and Ubuntu Desktop (i7) it was very choppy, so I've played around with a greater chunk size (16MB, 32MB). 32 Seemed better, but at some point it started dropping frames again. I event wen't higher to 512MB to see if the streamed data is the bottleneck and that actually proofed to be the case.

However for some reason the sound was getting more off by modifying the chunkSize.

Cheers

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants