Closed
Bug 1275596
Opened 8 years ago
Closed 8 years ago
createMediaElementSource on YouTube video turns off the audio of any video played thereafter
Categories
(Core :: Audio/Video: MediaStreamGraph, defect, P1)
Tracking
()
VERIFIED
FIXED
mozilla49
Tracking | Status | |
---|---|---|
firefox47 | --- | unaffected |
firefox48 | + | verified |
firefox49 | + | verified |
People
(Reporter: epinal99-bugzilla2, Assigned: pehrsons)
References
Details
(Keywords: regression)
Attachments
(1 file)
58 bytes,
text/x-review-board-request
|
padenot
:
review+
Sylvestre
:
approval-mozilla-aurora+
|
Details |
Bug found by testing bug 1275393. STR: 1) Open YT video https://www.youtube.com/watch?v=jI-kpVh6e1U 2) Run in web console: context = new window.AudioContext(); source = context.createMediaElementSource($('video')); source.connect(context.destination); 3) Clik on any video link on the page Result: audio of 2nd video (and next ones) is off. Regression range: https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=f5fe9e219441413c6fc8429050ca19c910b7c62a&tochange=da8d6c4eab6142c98b04706255498750076fee9d
Blocks: 1208371
Has Regression Range: --- → yes
Has STR: --- → yes
status-firefox47:
--- → unaffected
status-firefox48:
--- → affected
status-firefox49:
--- → affected
tracking-firefox48:
--- → ?
tracking-firefox49:
--- → ?
Flags: needinfo?(pehrsons)
Keywords: regression
Comment 1•8 years ago
|
||
Andreas -- I believe this was introduced by your landing and that you're planning to take this (based on irc conversations). If I'm wrong, please needinfo me back so I can find an owner. We want to fix this (if we can) before Fx48 goes to Beta. Thanks.
Assignee: nobody → pehrsons
Rank: 15
Priority: -- → P1
Assignee | ||
Comment 3•8 years ago
|
||
I'll investigate now.
Status: NEW → ASSIGNED
Flags: needinfo?(pehrsons)
Assignee | ||
Comment 4•8 years ago
|
||
I'm pretty sure this revision is causing the problems we see: https://hg.mozilla.org/mozilla-central/rev/1e531196e085 MediaElementAudioSourceNode is a sub class of MediaStreamAudioSourceNode: https://dxr.mozilla.org/mozilla-central/rev/8d0aadfe7da782d415363880008b4ca027686137/dom/media/webaudio/MediaElementAudioSourceNode.h#15 I changed MediaStreamAudioSourceNode to lock on to the first track. This was fine for MediaElementAudioSourceNode as well since we don't support multiple AudioTracks in a single HTMLMediaElement yet. The only problem is -- if the first track has ended there will be no more data, so we can skip to the next track automatically. This is not really defined in the spec, but automatically switching to the next track is more in line with how we behaved before my refactor.
Assignee | ||
Comment 5•8 years ago
|
||
This is how it was meant to work when the refactor landed in Bug 1208371. We have no test coverage of seeking apparently. Review commit: https://reviewboard.mozilla.org/r/55446/diff/#index_header See other reviews: https://reviewboard.mozilla.org/r/55446/
Attachment #8756852 -
Flags: review?(padenot)
Comment 6•8 years ago
|
||
Comment on attachment 8756852 [details] MozReview Request: Bug 1275596 - Ignore ended tracks when selecting new track in MediaStreamAudioSourceNode and MediaElementSourceNode. r?padenot https://reviewboard.mozilla.org/r/55446/#review52190
Attachment #8756852 -
Flags: review?(padenot) → review+
Assignee | ||
Comment 7•8 years ago
|
||
try push: https://treeherder.mozilla.org/#/jobs?repo=try&revision=002c5a2989c5
Comment 8•8 years ago
|
||
Tracking this regression for 48/49 based on the fact that the audio is turned off.
Assignee | ||
Comment 9•8 years ago
|
||
That try looks bad. Here's a new one rebased: https://treeherder.mozilla.org/#/jobs?repo=try&revision=fd792d2f9daa Maire, if that looks equally bad, could you decide how to handle those intermittents while I'm off for the night? I don't see any relation at all between my changes and the oranges that show up on try.
Flags: needinfo?(mreavy)
Comment 10•8 years ago
|
||
I see the same failure in try pushes that don't include your patch; I think it's a timing issue related to the speed of the test VM. Perhaps due to the track cloning, but not your current patch, so you can land. See https://treeherder.mozilla.org/#/jobs?repo=try&revision=bd4b1ee34e18b7691193ee6d0c29da8de9941690 (and I've seen it on other Try pushes, but not on inbound)
Flags: needinfo?(mreavy) → needinfo?(pehrsons)
Assignee | ||
Comment 13•8 years ago
|
||
Comment on attachment 8756852 [details] MozReview Request: Bug 1275596 - Ignore ended tracks when selecting new track in MediaStreamAudioSourceNode and MediaElementSourceNode. r?padenot Approval Request Comment [Feature/regressing bug #]: Bug 1208371 [User impact if declined]: Using MediaStreams or media elements as input to WebAudio might make it go silent when it shouldn't. [Describe test coverage new/current, TreeHerder]: Tested manually. [Risks and why]: Low. Change is simple. [String/UUID change made/needed]: None.
Attachment #8756852 -
Flags: approval-mozilla-aurora?
Comment 14•8 years ago
|
||
bugherder |
https://hg.mozilla.org/mozilla-central/rev/0fa7d5fb96af
Status: ASSIGNED → RESOLVED
Closed: 8 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla49
Comment 15•8 years ago
|
||
Comment on attachment 8756852 [details] MozReview Request: Bug 1275596 - Ignore ended tracks when selecting new track in MediaStreamAudioSourceNode and MediaElementSourceNode. r?padenot Fix a recent regression, taking it
Attachment #8756852 -
Flags: approval-mozilla-aurora? → approval-mozilla-aurora+
Comment 16•8 years ago
|
||
bugherder uplift |
https://hg.mozilla.org/releases/mozilla-aurora/rev/d0bf93e1ad2e
Updated•8 years ago
|
Flags: qe-verify+
Comment 17•8 years ago
|
||
Verified as Fixed in Firefox Beta 48.0b2 (20160620091522) and Firefox DevEdition 49.0a2 (20160622004014) on Windows 7 x64 and Windows 10 x64.
Status: RESOLVED → VERIFIED
Comment 18•8 years ago
|
||
Updating tracking flags, based on the above comment.
You need to log in
before you can comment on or make changes to this bug.
Description
•