New issue
Advanced search Search tips

Issue 896010 link

Starred by 3 users

Issue metadata

Status: Started
Owner:
Components:
EstimatedDays: ----
NextAction: ----
OS: Linux
Pri: 2
Type: Bug



Sign in to add a comment

Add P010 raw video format support

Project Member Reported by sreerenj...@intel.com, Oct 16

Issue description

UserAgent: Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Ubuntu Chromium/69.0.3497.81 Chrome/69.0.3497.81 Safari/537.36

Example URL:

Steps to reproduce the problem:
Add P010 raw video format support

P010 is the two-plane 10-bits per channel YUV 4:2:0 video format where each sample is a two-byte little-endian value with the lower six bits ignored. So far this is the only format supported by many of the Intel platforms for 10 bit hardware decode and encode and this is already implemented
in meida/gpu/vaapi.

What is the expected behavior?

What went wrong?
There is no mapping in chrome for the similar ffmpeg pixel format AV_PIX_FMT_P010LE.

Did this work before? N/A 

Is it a problem with Flash or HTML5? N/A

Does this work in other browsers? N/A

Chrome version:   Channel: n/a
OS Version: 
Flash Version: 

Contents of chrome://gpu: 

I have a quick fix patch for this. No unit tests are added yet. Also no libyuv support for the same. I will upload it for review after the clean up.
 
Labels: Needs-Triage-M69
Labels: Needs-Feedback
@sreerenj.balachandran What is the current state of the patch?  Is this now fixed?
@dougman: the patch is here https://chromium-review.googlesource.com/c/chromium/src/+/1286934

I'm waiting for Dale to make a final decision. Technically, I'm not agreeing with one of the reviewer but I can make the change proposed in the review (no p010, only p016 will be added then) if Dale is also proposing the same.
Dongseong had some comments that you never replied to? He seemed in favor of our recommended approach, so I suggest y'all talk.
IIUC Dongseong is saying that we need P010 :), not the other way around. He says, buffertype can be P016, thats it. right?

Also I believe, DS's question was for Hubinette. isn't it?
Hubbe@ has left the team. BufferType and VPT are the same thing just for different classes, so that's an argument in favor of P016 only.

Comment 7 by dbbrooks@chromium.org, Today (13 hours ago)

Owner: sreerenj...@intel.com
Status: Started (was: Unconfirmed)
Don't think this fits the "Unconfirmed" state it is currently set to. Changing to "started" since the patch is in and passed review. Also assigning it to the patch owner sreerenj.balachandran@. 

Comment 8 by dbbrooks@chromium.org, Today (13 hours ago)

Labels: -Needs-Feedback

Sign in to add a comment