|
|
First of all, hi everybody and thanks for your time.
I'm trying to setup a scenario when an user just can play seamlessly video over multiple AP all connected to the same RTSP server.
I've set up a VLC RTSP server to play mp4 video, and the client connects using gstreamer. My pipeline it's as follows:
pipeline=gst.parse_launch('rtspsrc name=source source. ! rtph264depay ! decodebin2 ! autovideosink source. ! rtpmp4gdepay ! ffdec_aac ! autoaudiosink')
I then iterate the elements inside rtspsrc and fetch the rtpbin inside it.
Once I have access to the rtpbin, I set buffer-mode to High/low watermark buffering and latency to a time (for example, 4000 msec). That latency it's set in the jitterbuffers inside rtpbin.
Now I play the stream and, if the network does a handover, it continues playing during the time of the buffer, but when it ends, even if the network connection is resumed, it stops for severlas seconds (much more than the buffer time) and then there is some corruption and QoS messages posted over the bus (acceptable when using UDP) and then it resumes.
As I understand, if the handover between AP's it's done in a time shorter than the latency from rtpbin, it shouldn't affect the stream as to force it to freeze several seconds. What am I doing wrong?
I'm totally stuck with this, so if anyone can give me some hints, it would be really nice.
Thanks a lot!
|
Locked
