Closed Bug 1051566 Opened 9 years ago Closed 9 years ago

Add H.264 mode 0 support


(Core :: WebRTC: Audio/Video, defect)

Not set



Tracking Status
firefox32 --- wontfix
firefox33 --- fixed
firefox34 --- fixed
b2g-v2.0 --- wontfix
b2g-v2.1 --- fixed


(Reporter: jesup, Assigned: jesup)




(2 files)

We should support H.264 mode 0, though we prefer mode 1.

This requires offering/negotiating it, and keying off it for the packet size (override the value passed into codec init)

Note: will not work on OMX HW 8x10 currently (need to add to dependency list for that).
Note: I updated the webrtc_landing self-call page ( to have a Mode 0 checkbox.
Attachment #8470501 - Flags: review?(ethanhugg)
We should add some tests for this
Comment on attachment 8470501 [details] [diff] [review]
Support H.264 Mode 0 in GMP OpenH264

Review of attachment 8470501 [details] [diff] [review]:

::: media/webrtc/signaling/src/media-conduit/WebrtcGmpVideoCodec.cpp
@@ +204,5 @@
>    // Pass dummy codecSpecific data for now...
>    nsTArray<uint8_t> codecSpecific;
>    // H.264 mode 1 only supported so far

I think we can remove this comment now.

::: media/webrtc/signaling/src/media/VcmSIPCCBinding.cpp
@@ +2865,3 @@
>    return VCM_H264_MODE_1;
> +#else
> +  return VCM_H264_MODE_0|VCM_H264_MODE_1;

I changed this to do only MODE_0 and ran some tests and it seems to work.
Attachment #8470501 - Flags: review?(ethanhugg) → review+
Whiteboard: [openh264-uplift]
Closed: 9 years ago
Resolution: --- → FIXED
Comment on attachment 8470501 [details] [diff] [review]
Support H.264 Mode 0 in GMP OpenH264

Note: should be paired with bug 1050461

Approval Request Comment
[Feature/regressing bug #]:N/A

[User impact if declined]: No support for H.264 mode 0, which is required by IETF 6184.  Most equipment doesn't use mode 0, but some equipment people care about does.

[Describe test coverage new/current, TBPL]: On m-c since the 12th.  Considerable manual testing by Cisco, and manual testing.  I added a checkbox to force Mode 0 to our primary test page ( We can't test H.264 OMX using a real codec in TBPL.  

[Risks and why]: Quite low risk (very small changes, and only a few code paths that aren't shared with Mode 0); almost all risk is from the other patch.

[String/UUID change made/needed]: none
Attachment #8470501 - Flags: approval-mozilla-aurora?
Attachment #8470501 - Flags: approval-mozilla-aurora? → approval-mozilla-aurora+
Whiteboard: [openh264-uplift]
You need to log in before you can comment on or make changes to this bug.