I've been trying to get this to work for a couple of days now, without
any success. Since this is the basis for a university project, I'm
afraid I cannot do any real work until I can get the DSP to decode and
display video.
So once again, I hope you guys could share your expertise in solving
this issue. Currently it seems that GStreamer TI is able to
communicate with the DSP, but the problem lies in displaying the
output to the screen using one of the sinks.
I have tried two different pipelines, and here's how they differ:
1. gst-launch -v filesrc location="sample.m4v" ! TIViddec2 !
omapdmaifbsink
This pipeline simply starts the process, but quickly returning with an
ERROR: "Pipeline doesn't want to pause.". Following it, the process
exits.
2. gst-launch -v filesrc location="sample.m4v" ! TIViddec2 !
TIDmaiVideoSink videoStd=VGA videoOutput=LCD sync=false
The second pipeline uses a different sink. This sink seems to work a
little better since the process does begin. I'm getting the file's
header information, and it seems as is the DSP is started to decode
some of the first frames. However, after a second or two I'm getting a
message saying:
ERROR: from element /GstPipeline:pipeline0/
GstTIDmaiVideoSink:tidmaivideosink0: Unable to initialize display
gst_tidmaivideosink_render (): /GstPipeline:pipeline0/
GstTIDmaiVideoSink:tidmaivideosink0
So essentially, it seems like the second pipe would work, if the
display is set correctly. I've tried doing 'export DISPLAY=:0" and
also setting the sink to point to any one of /dev/fb* devices, but
without any luck (same exact error message).
Hope you guys can help..
Regards,
Guy