-
Notifications
You must be signed in to change notification settings - Fork 9.4k
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
Perceptual speed index not working well for animated content #2514
Comments
Is this a flaw of the SI metric? I've seen similar issues on sites that have a continuous animation in the background (or a video). The I/O 2016 is an example where SI gets pushed out b/c the animation keeps changing. WPT never sees the page as "visually complete". |
Yes, I think it's a flaw in the module that is being used. I think it screenshots the page at intervals and waits for the difference between consecutive screenshots to be small. |
Yes this is a critical flaw of all metrics (SI and PSI) using a single target frame as the "ideal" or "idle" state and judging progress toward that frame. There are some other critical problems with PSI discussed over in speedline like paulirish/speedline#52 and paulirish/speedline#50. Plenty of work to do here :) |
Should we just close this issue? The problem lies in the metric and not LH :) |
I know @patrickhulce wants to keep investigating improvements, though he can comment on timescale for that. It would be nice to have a tracking bug over here, though it may make sense to open a new one with clearer title and summary of current state and what could be improved. |
Yeah I'm not sure if any immediate progress will be made on this specific problem though since speed index is so well established as a thing at this point, at best we'll come up with a variant in a new metric. It might be fine to leave this open to coalesce anyone filing issues like #2607 :) |
SG :)
Sounds terrible, who writes things like this. |
I've written an app that, on first load, shows a full screen video that is being fed by getUserMedia.
If I stick something over my camera I get a good score for perceptual speed index (1.3s), whereas if I run LH while the camera is uncovered and I'm waving my hands about, I get bad scores (4.7s is my worst)
Clearly the test is not measuring anything useful in this case and I wonder if it can be improved.
The text was updated successfully, but these errors were encountered: