[gst-devel] gstrtpjitterbuffer and sequence number restart?

[Tero S. <ter...@gm...>]

Hi,

When RTP stream has discontinuity in sequence numbers the source is
considered valid again if the next packet carries the next higher
sequence number.  I can see this behavior in rtpsource but for some
reason gstjitterbuffer never recovers.  How should the buffer get
flushed so that the stream would continue to be processed again?

rtpsource rtpsource.c:402:rtp_source_process_rtp: seq 17443, PC: 96, OC: 24960
rtpsource rtpsource.c:270:calculate_jitter: rtparrival 1405963725,
rtptime 3361221841, clock-rate 24000, diff 1017, jitter: 34584
rtpsource rtpsource.c:197:push_packet: pushing new packet
rtpjitterbuffer
gstrtpjitterbuffer.c:803:gst_rtp_jitter_buffer_chain:<rtpjitterbuffer0>
Received packet #17443
rtpjitterbuffer
gstrtpjitterbuffer.c:848:gst_rtp_jitter_buffer_chain:<rtpjitterbuffer0>
Pushed packet #17443 on queue 17
rtpsource rtpsource.c:416:rtp_source_process_rtp: unacceptable seqnum received
rtpsource rtpsource.c:300:init_seq: base_seq 16338
rtpsource rtpsource.c:402:rtp_source_process_rtp: seq 16338, PC: 1, OC: 280
rtpsource rtpsource.c:270:calculate_jitter: rtparrival 1405976144,
rtptime 3360090335, clock-rate 24000, diff 1143925, jitter: 1176347
rtpsource rtpsource.c:197:push_packet: pushing new packet
rtpjitterbuffer
gstrtpjitterbuffer.c:803:gst_rtp_jitter_buffer_chain:<rtpjitterbuffer0>
Received packet #16338
rtpjitterbuffer
gstrtpjitterbuffer.c:883:gst_rtp_jitter_buffer_chain:<rtpjitterbuffer0>
Packet #16338 too late as #17426 was already popped, dropping

All packets are dropped after this.

-- 
Tero