[gst-devel] timestamping pipeline problem / payloader / filesrc / (fakesink/udpsink)

SP GLE spglegle at yahoo.fr
Wed Dec 5 13:40:34 CET 2007


Hi,

When executing the following cmd :

GST_DEBUG_NO_COLOR=1 GST_DEBUG="*identity*:5,*sink*:5" gst-launch
filesrc location=/tmp/file.alaw blocksize=4000 ! audio/x-alaw,
rate=8000, channels=1 ! rtppcmapay min-ptime=100000000
max-ptime=100000000 ! identity check-perfect=true sync=true ! fakesink
sync=true 

We observe a timestamp discontinuity (see below), the output buffers
(to fakesink) from rtppcmapay obtained from the first 4000 Bytes buffer
from 
filesrc seems to have a monotic timestamp and so properly synchronised
by fakesink, the second buffer from filesrc is splitted by rtppcmapay
(min-ptime) and these buffers are not time contiguous with previous
ones.
So when replacing fakesink by an udpsink the first buffers are properly
sent, because of the buffer timestamp but next ones are sent in burst,
saturating receiver jitter buffer.

What's the right way to solve this problem ?

Regards.

Laurent



0:00:00.150360000 10201 0x8157a78 DEBUG             basesink
gstbasesink.c:1682:gst_base_sink_do_sync:<fakesink0> possibly waiting
for clock to reach 0:00:00.000000000
0:00:00.150426000 10201 0x8157a78 DEBUG             basesink
gstbasesink.c:1688:gst_base_sink_do_sync:<fakesink0> clock returned 1
0:00:00.150786000 10201 0x8157a78 DEBUG             basesink
gstbasesink.c:1682:gst_base_sink_do_sync:<fakesink0> possibly waiting
for clock to reach 0:00:00.100000000
0:00:00.252075000 10201 0x8157a78 DEBUG             basesink
gstbasesink.c:1688:gst_base_sink_do_sync:<fakesink0> clock returned 0
0:00:00.252405000 10201 0x8157a78 DEBUG             basesink
gstbasesink.c:1682:gst_base_sink_do_sync:<fakesink0> possibly waiting
for clock to reach 0:00:00.200000000
0:00:00.352063000 10201 0x8157a78 DEBUG             basesink
gstbasesink.c:1688:gst_base_sink_do_sync:<fakesink0> clock returned 0
0:00:00.352378000 10201 0x8157a78 DEBUG             basesink
gstbasesink.c:1682:gst_base_sink_do_sync:<fakesink0> possibly waiting
for clock to reach 0:00:00.300000000
0:00:00.452047000 10201 0x8157a78 DEBUG             basesink
gstbasesink.c:1688:gst_base_sink_do_sync:<fakesink0> clock returned 0
0:00:00.452351000 10201 0x8157a78 DEBUG             basesink
gstbasesink.c:1682:gst_base_sink_do_sync:<fakesink0> possibly waiting
for clock to reach 0:00:00.400000000
0:00:00.552032000 10201 0x8157a78 DEBUG             basesink
gstbasesink.c:1688:gst_base_sink_do_sync:<fakesink0> clock returned 0
0:00:00.552857000 10201 0x8157a78 DEBUG             basesink
gstbasesink.c:1682:gst_base_sink_do_sync:<fakesink0> possibly waiting
for clock to reach 0:00:00.099999999
0:00:00.552913000 10201 0x8157a78 DEBUG             basesink
gstbasesink.c:1688:gst_base_sink_do_sync:<fakesink0> clock returned 1
0:00:00.553305000 10201 0x8157a78 DEBUG             basesink
gstbasesink.c:1682:gst_base_sink_do_sync:<fakesink0> possibly waiting
for clock to reach 0:00:00.199999999
0:00:00.553368000 10201 0x8157a78 DEBUG             basesink
gstbasesink.c:1688:gst_base_sink_do_sync:<fakesink0> clock returned 1
0:00:00.553730000 10201 0x8157a78 DEBUG             basesink
gstbasesink.c:1682:gst_base_sink_do_sync:<fakesink0> possibly waiting
for clock to reach 0:00:00.299999999
0:00:00.553781000 10201 0x8157a78 DEBUG             basesink
gstbasesink.c:1688:gst_base_sink_do_sync:<fakesink0> clock returned 1
0:00:00.554220000 10201 0x8157a78 DEBUG             basesink
gstbasesink.c:1682:gst_base_sink_do_sync:<fakesink0> possibly waiting
for clock to reach 0:00:00.399999999
0:00:00.554308000 10201 0x8157a78 DEBUG             basesink
gstbasesink.c:1688:gst_base_sink_do_sync:<fakesink0> clock returned 1
0:00:00.555075000 10201 0x8157a78 DEBUG             basesink
gstbasesink.c:1682:gst_base_sink_do_sync:<fakesink0> possibly waiting
for clock to reach 0:00:00.099999999



      _____________________________________________________________________________ 
Ne gardez plus qu'une seule adresse mail ! Copiez vos mails vers Yahoo! Mail 




More information about the gstreamer-devel mailing list