pygobject performance optimizations posted

2013-08-27 Thread Daniel Drake
Hi,

http://rpmdropbox.laptop.org/f18/ now includes
pygobject3-3.4.2-6.fc18.olpc2 (on 13.2.0 developers can install with
yum update pygobject3)

This new package includes some optimizations from pygobject upstream.

I can't say the improvement is as visible as I was hoping for, but
some progress is being made, and there is hopefully more to come.

Any testing appreciated. It's also important to know that nothing has
broken with these changes.

Daniel
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


Re: gst-vmeta testing

2013-08-27 Thread Daniel Drake
On Sun, Aug 25, 2013 at 5:30 PM, Carlos Rafael Giani
d...@pseudoterminal.org wrote:
 Now for human observation: there is no flicker problem.

 Interesting. Try again with this video please:
 http://download.openbricks.org/sample/H264/SampleClip.m4v
 With this one, I could definitely see flickering in past tests.

No flickering with that video.


 However, the
 video is noticably choppier when played back in the gstreamer-1.0
 environment. For example, look at the smoothness of the front door
 opening in that video.

 This looks like an issue with the parser. I am not sure what to make of it.
 I find it rather unlikely that this is caused by the decoder. I see warnings
 like these often with somewhat broken videos. Marvell's decoder does the
 parsing on its own, while mine doesn't ; instead, it leaves the parsing up
 to the existing GStreamer MPEG parsers. It is possible then that the custom
 Marvell parser code simply swallows broken timestamps etc. without saying
 anything.
 I wasn't able to try the video myself, since I have been very busy the last
 few days with something else. I will get to try it out tomorrow hopefully.

SampleClip.m4v plays back really smoothly with none of the frame drop
errors. And I just checked another sample,
http://techslides.com/demos/sample-videos/small.mp4 and that works
too.

So it does seem like you are right in suspecting something else at
play with the hellmans_mayonnaise_commercial.mpeg file.

Either way it would be nice if you could confirm/deny the findings
there, next time you have your device booted.

Thanks
Daniel
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel