statistics about received H264 NAL units
Ottavio Campana
o.campana at videotec.com
Tue Feb 5 07:38:16 PST 2013
On 02/05/2013 04:16 PM, Tim-Philipp Müller wrote:
>> I need to create a program that allows me to extract a statistic about
>> the packets of an H.264 received video. Particularly, I would like to
>> count the total NAL units received and the number of the units that
>> were corrupted and that were therefore not decoded.
>>
>> I'd like to know if it possibile to do it with gstreamer. I don't need
>> to play the video, I just need a simple output that for each NAL unit
>> signals whether it was correctly decoded or not, so that I could
>> execute the program with a pipe from another program and evaluate all
>> the statistics that I need.
>>
>> Do you think it is possibile to do it? Can you please help me make it?
> I think you would want/need decoder support for what you want to do. You
> could probably detect whole frames that don't get decoded at all, but
> not if individual NALs or slices weren't decoded correctly and you still
> get a frame with some artefacts.
Ok, but I'd like to know is there a way to achieve that or something
similar with gstreamer.
I am thinking about something like
|gst-launch rtspsrc location=rtsp://192.168.0.20/live1.sdp !||rtph264depay !||h264parse
|
||but I don't know how to finish the command to be able to get
information about what had a correct syntax.
If you're wondering why I am interested in NAL units and not frames or
bit-error rate, it is because I am trying to characterize a device that
I do not control and I'd like to have measure that is not falsed by the
GOP size. Streaming only I frames reduces the propagation error, but I
would like to have a measure that can be independent of this kind of
"tricks".
Thank you in advance if you can help me,
Ottavio
||
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/gstreamer-devel/attachments/20130205/016d20ac/attachment.html>
More information about the gstreamer-devel
mailing list