js_mse_eme is an externally-published tool that is aimed to test the validity of a browser's HTML5 Media Source Extension and Encrypted Media Extension implementations
Apache License 2.0
91
stars
58
forks
source link
This test does not take into account devices with long pre-roll. #35
This line, on devices using Chromium with long video pre-roll, will result in couple of timeupdate events synthesized before any current time update actually comes from underlying native player. This results in baseTimeDiff with little connection to actual time update reporting. And since the test looks for worst delta against wall-clock, all actual changes in PTS may appear to have it rather big, no matter how stable reporting actually is.
One way to account for pre-roll might be to ignore for the synthesized events with
if (video.currentTime === 0.0)
return;
baseTimeDiff = util.ElapsedTimeInS() - video.currentTime;
and getting the baseline at the first non-zero time reported by the player.
jiaqzhao
commented
6 years ago
https://github.com/youtube/js_mse_eme/blob/7e0b921d3f6c4c311679d31cce13f2e3dd69abb2/js/tests/2019/conformanceTest.js#L535
This line, on devices using Chromium with long video pre-roll, will result in couple of
timeupdateevents synthesized before any current time update actually comes from underlying native player. This results inbaseTimeDiffwith little connection to actual time update reporting. And since the test looks for worst delta against wall-clock, all actual changes in PTS may appear to have it rather big, no matter how stable reporting actually is.One way to account for pre-roll might be to ignore for the synthesized events with
and getting the baseline at the first non-zero time reported by the player.