927474 - "probably" can play "audio/webm"

Status: RESOLVED DUPLICATE of bug 884275

(Core :: Audio/Video, defect)

Description

[Gian-Carlo Pascutto [:gcp]]

canPlayType("audio/webm") currently returns "probably"  (the same thing applies to video/webm)

http://www.whatwg.org/specs/web-apps/current-work/multipage/the-video-element.html#dom-navigator-canplaytype
"Generally, a user agent should never return "probably" for a type that allows the codecs parameter if that parameter is not present."

It's not clear to me whether webm allows it right now, but note that plans to add Opus/VP9 are in the works, which would seem to require adding codecs to get sane behavior:
https://groups.google.com/a/webmproject.org/forum/?fromgroups=#!topic/webm-discuss/UzoX7owhwB0
"No WebM2: the existing WebM container will be extended to allow VP9
and Opus streams, and to accommodate any of their special features.
Content-type remains video/webm, and the codec strings are being
worked out with Matroska."

Then there's a few resources like this:
http://www.w3schools.com/tags/av_met_canplaytype.asp
Note: This method can only return "probably" if codecs are included.

We should probably not return probably, but return maybe, maybe?  :-)

FWIW Chrome claims it can "maybe" play "audio/webm".

Comment 1

[Ralph Giles (:rillian)]

Background (also discussed on irc):

We have argued vp9+opus should be 'video/webm2' to avoid confusion over whether a given file can be played back. I haven't seen a proper discussion of this, but the two obstacles to this outcome:

a) Google is proceeding to implement webm with any combination of vp8, vp9, vorbis, and opus.
b) We ourselves would like to be able to do vp8+opus for recording WebRTC streams and general MediaEncoder API performance on mobile.

I think (b) makes it hard to push a clean case against (a) so we should just accept any combination of the above codecs for playback in the a 'webm' container, and start returning 'maybe' instead of 'probably' for {audio|video}/webm without a codecs parameter.

> http://www.w3schools.com/tags/av_met_canplaytype.asp

vp8.0, eh? Well, at least our parser already handles that.

Comment 2

[Chris Pearce [:cpearce (Not reading bugmail)]]

We could push WebM2 as any combination of {vp8,vp8,vorbis,opus}, leaving WebM1 as purely {vp8,vorbis}. Anthony suggested this when I spoke to him the other day about this.

Comment 3

[Ralph Giles (:rillian)]

I've pushed back. https://groups.google.com/a/webmproject.org/d/msg/webm-discuss/UzoX7owhwB0/kgbK7pzYR0UJ

Comment 4

[Gian-Carlo Pascutto [:gcp]]

That looks like a real dupe to me.