Hardware accelerated playback on RPI with OpenGL

[Sergey Borovkov]

Sorry. Pressed send too early.
2. I tried the following pipeline with gst-launch:
gst-launch-1.0 filesrc location=... ! decodebin ! glupload ! glimagesink
But I am getting the same insane amount of frames dropped.

Any ideas what could be performance bottleneck? How can I make sure that
context is being shared and nothing is copied?

Thanks in advance.

On 6 January 2016 at 22:22, Sergey Borovkov <serge.borovkov at gmail.com>
wrote:

> Hi. I've been trying to use qmlglsink on Raspberry Pi. (It works fine if
> Android codepaths are used since it's also OpenGL ES). Unfortunately I
> encountered some severe performance issues when playing 1080p videos. (I am
> using omxdecoder that is set up with broadcom OMX library).
> I tried few pipelines:
> 1. In my application I have the following one filesrc -> decodebin ->
> glupload -> qmlglsink.
> As I said it worked almost without any changes. Just using Android
> codepaths (and adding markDirty to updatePaintNode since Qt did not redraw
> the scene with reused texture). Problem I am getting is that omxdecoder
> reports a great number of dropped frames. For Big Buck Bunny trailer I am
> getting around 600 messages about frame dropped. Qt itself reports that
> time spent in rendere is 40-80 ms. And while it could cause some frames
> being dropped 600 frames looks really crazy.
> 2. I tried the following pipeline with gst-launch:
> gst-launch-1.0 filesrc location=... ! decodebin ! glupload
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/gstreamer-devel/attachments/20160106/73e135e2/attachment.html>