You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I don't know enough about Android to say much for sure, but I think this is an "ANR".
I've noticed some slowness with longer (3+ hours) files before, but it's so severe on Ian Cutress' 9:30 Hot Chips livestream that playback dies after a few seconds. Youtube-hosted files aren't exactly stable, but this ffmpeg command generates a file that exhibits the problem:
Attempting to play that results in a failure after 5-ish seconds. Replace libopus with libmp3lame, and .opus with .mp3, and it plays just fine. Attached is a logcat dump started just before pressing play, and ^C'd just after playback stopped. The interesting part begins at timestamp 08-24 21:23:12.424. AudioAnchor is uid=10249, pid=32752.
It seems like there's maybe some computation/memory cost that scales with file size or length. top shows 70-ish % CPU usage from mediaserver and media.extractor aextractor, but their RSS is < 15 MiB. I see no spike in memory or CPU footprint from AudioAnchor itself.
(Although the baseline of ~16% CPU just paused on the playback screen with any file is somewhat concerning. surfaceflinger is ~20%. Is it redrawing continuously?)
The text was updated successfully, but these errors were encountered:
I don't know enough about Android to say much for sure, but I think this is an "ANR".
I've noticed some slowness with longer (3+ hours) files before, but it's so severe on Ian Cutress' 9:30 Hot Chips livestream that playback dies after a few seconds. Youtube-hosted files aren't exactly stable, but this ffmpeg command generates a file that exhibits the problem:
Attempting to play that results in a failure after 5-ish seconds. Replace libopus with libmp3lame, and .opus with .mp3, and it plays just fine. Attached is a logcat dump started just before pressing play, and ^C'd just after playback stopped. The interesting part begins at timestamp 08-24 21:23:12.424. AudioAnchor is uid=10249, pid=32752.
attempt_to_play_opus_everything.log
It seems like there's maybe some computation/memory cost that scales with file size or length.
top
shows 70-ish % CPU usage frommediaserver
andmedia.extractor aextractor
, but their RSS is < 15 MiB. I see no spike in memory or CPU footprint from AudioAnchor itself.(Although the baseline of ~16% CPU just paused on the playback screen with any file is somewhat concerning.
surfaceflinger
is ~20%. Is it redrawing continuously?)The text was updated successfully, but these errors were encountered: