Original commit message from CVS:
patch by: Alessandro Decina
* gst/mpegtsparse/mpegtspacketizer.c:
* gst/mpegtsparse/mpegtsparse.c:
* gst/mpegtsparse/mpegtsparse.h:
Add request pad for getting the full transport stream coming in.
Original commit message from CVS:
* configure.ac:
Require core CVS. This is implicit in the -base CVS
requirement already, so we might just well spell it
out. Also, we do need at least 0.10.14 for
gst_element_class_set_details_simple(). Make check
for gmyth a bit more restrictive so things don't break
if the next version changes API.
* ext/alsaspdif/alsaspdifsink.c:
Work around alsa alloca macros triggering 'always evaluates to
true' warnings with gcc-4.2 and fix compilation with gcc-4.2.
Also don't leak the device string.
* ext/mpeg2enc/gstmpeg2enc.cc:
* ext/soundtouch/gstpitch.cc:
* gst/modplug/gstmodplug.cc:
Fix compilation with g++4.2 and -Wall -Werror (also needs plugin
define fix from core CVS). Fixes#462737.
Original commit message from CVS:
* gst/rtpmanager/rtpsession.c: (rtp_session_next_timeout),
When reconsidering RTCP timeouts, set the next timeout against the last
report time instead of the current clock time so that we don't end up
reconsidering forever.
Original commit message from CVS:
* gst/rtpmanager/gstrtpjitterbuffer.c:
(gst_rtp_jitter_buffer_chain), (gst_rtp_jitter_buffer_loop):
Only peek at the tail element instead of popping it off, which allows
us to greatly simplify things when the tail element changes.
* gst/rtpmanager/gstrtpsession.c:
(gst_rtp_session_event_recv_rtp_sink):
* gst/rtpmanager/gstrtpssrcdemux.c:
(gst_rtp_ssrc_demux_sink_event):
Forward FLUSH events instead of leaking them.
* gst/rtpmanager/rtpjitterbuffer.c: (rtp_jitter_buffer_reset_skew),
(calculate_skew), (rtp_jitter_buffer_insert):
* gst/rtpmanager/rtpjitterbuffer.h:
Remove the tail-changed callback in favour of a simple boolean when we
insert a buffer in the queue.
Add method to peek the tail of the buffer.
Original commit message from CVS:
Patch by: Gautier Portet <kassoulet at gmail dot com>
* gst/xingheader/gstxingmux.c:
The size of the Xing header is actually 417 as it's rounded to the
next smaller integer. Fixes#397759.
* gst/xingheader/gstxingmux.c: (xing_generate_header),
(xing_push_header):
Some random cleanup, add FIXMEs and TODOs and check if the newsegment
event to the beginning was successful before pushing the header again.
Original commit message from CVS:
Patch by: Wai-Ming Ho <webregbox at yahoo dot co dot uk>
* gst/mpegtsparse/mpegtspacketizer.c:
(mpegts_packetizer_stream_new):
Don't skip PAT with version number 0. Fixes#483400.
* gst/mpegtsparse/mpegtsparse.c: (mpegts_parse_apply_pat):
Make all values above 0 mark a referenced program as they can be
incremented and only 1 had marked a referenced program before, causing
actually referenced programs to be unreferenced.
Original commit message from CVS:
* gst/rtpmanager/gstrtpjitterbuffer.c:
(gst_rtp_jitter_buffer_flush_start),
(gst_rtp_jitter_buffer_flush_stop),
(gst_rtp_jitter_buffer_change_state), (apply_offset),
(gst_rtp_jitter_buffer_loop):
Remove some old unused variables.
Don't add the latency to the skew corrected timestamp, latency is only
used to sync against the clock.
Improve debugging.
* gst/rtpmanager/rtpjitterbuffer.c: (rtp_jitter_buffer_init),
(rtp_jitter_buffer_reset_skew), (calculate_skew):
* gst/rtpmanager/rtpjitterbuffer.h:
Handle case where server timestamp goes backwards or wildly jumps by
temporarily pausing the skew correction.
Improve debugging.
Original commit message from CVS:
Patch by: mutex at runbox dot com
* gst/mpegtsparse/mpegtspacketizer.c:
(mpegts_packetizer_parse_adaptation_field_control):
* gst/mpegtsparse/mpegtsparse.c: (mpegts_parse_base_init),
(mpegts_parse_init), (mpegts_parse_push):
* gst/mpegtsparse/mpegtsparse.h:
Remove useless src pad that only results in not linked errors,
fix a broken pointer dereference and make MAX_CONTINUITY constant
conform to the standard to stop outputting corrupted data.
Fixes#481276, #481279.
Original commit message from CVS:
* gst/rtpmanager/gstrtpbin.c: (free_client):
Fix crasher in dispose.
* gst/rtpmanager/rtpjitterbuffer.c: (calculate_skew):
Handle cases where input buffers have no timestamps so that no clock
skew can be calculated, in this case interpollate timestamps based on
rtp timestamp and assume a 0 clock skew.
Original commit message from CVS:
* gst/rtpmanager/gstrtpjitterbuffer.c: (apply_latency),
(gst_rtp_jitter_buffer_loop), (gst_rtp_jitter_buffer_query):
Remove jitter correction code, it's now in the lower level object.
Use new -core method for doing a peer query.
* gst/rtpmanager/rtpjitterbuffer.c: (rtp_jitter_buffer_init),
(calculate_skew), (rtp_jitter_buffer_insert):
* gst/rtpmanager/rtpjitterbuffer.h:
Move jitter correction to the lowlevel jitterbuffer.
Increase the max window size.
When filling the window, already start estimating the skew using a
parabolic weighting factor so that we have a much better startup
behaviour that gets more accurate with the more samples we have.
Increase the default weighting factor for the steady state to get
smoother timestamps.
Original commit message from CVS:
2007-09-27 Julien MOUTTE <julien@moutte.net>
* gst/flv/gstflvparse.c: (gst_flv_parse_tag_audio),
(gst_flv_parse_tag_video): I got it wrong again, audio rate
was not detected correctly in all cases.
Original commit message from CVS:
* gst/rtpmanager/gstrtpbin.c: (gst_rtp_bin_dispose),
(gst_rtp_bin_finalize):
Fix cleanup crasher.
* gst/rtpmanager/rtpjitterbuffer.c: (rtp_jitter_buffer_init),
(calculate_skew):
* gst/rtpmanager/rtpjitterbuffer.h:
Dynamically adjust the skew calculation window so that we calculate it
over a period of around 2 seconds.
Original commit message from CVS:
2007-09-26 Julien MOUTTE <julien@moutte.net>
* gst/flv/gstflvparse.c: (gst_flv_parse_tag_audio),
(gst_flv_parse_tag_video): codec_data is needed for every tag
not just the first one. (Fix a stupid bug i introduced without
testing)
Original commit message from CVS:
2007-09-26 Julien MOUTTE <julien@moutte.net>
* gst/flv/gstflvparse.c: (gst_flv_parse_tag_audio),
(gst_flv_parse_tag_video): Fix bit masks operations to be
sure we detect the codec_tags and sample rates correctly.
Fix raw audio caps generation.
Original commit message from CVS:
* gst/librfb/gstrfbsrc.c:
* gst/librfb/rfbdecoder.c:
* gst/librfb/rfbdecoder.h:
Added offset-x, offset-y, width and height property
for selecting a region from the screen
Original commit message from CVS:
* gst/librfb/gstrfbsrc.c:
Minimum raw encoding is working now
* gst/librfb/rfbdecoder.c:
fix address while reading from stream
Original commit message from CVS:
* gst/librfb/gstrfbsrc.c:
raw encoding is working, but it looks like the
ffmpegcolorspace plugin can't handle high resolutions
Original commit message from CVS:
* ext/alsaspdif/alsaspdifsink.c:
* ext/timidity/gsttimidity.c:
* ext/timidity/gstwildmidi.c:
* gst/mpegvideoparse/mpegvideoparse.c:
Fix memory leaks. More to come.
* tests/check/Makefile.am:
* tests/check/generic/states.c:
Improved state change unit test.
Original commit message from CVS:
* gst/librfb/gstrfbsrc.c:
* gst/librfb/rfbdecoder.c:
* gst/librfb/rfbdecoder.h:
It is now possible to connect to a vncserver.
there are still some issues with the ouput of
the screen. Looks like some lines are confused
Original commit message from CVS:
* gst/real/gstrealvideodec.c: (gst_real_video_dec_chain),
(open_library), (gst_real_video_dec_init),
(gst_real_video_dec_set_property),
(gst_real_video_dec_get_property), (gst_real_video_dec_class_init):
* gst/real/gstrealvideodec.h:
Don't generate an error for occasional decoding errors.
Add max-errors property.
Error out when we receive max-errors in a row. Fixes#478159.
Original commit message from CVS:
* gst/librfb/gstrfbsrc.c:
Add password property (write only)
* gst/librfb/rfbdecoder.c:
Read the reason on failure
Use the password property for authentication
* gst/librfb/rfbdecoder.h:
Add defines for version checking
Original commit message from CVS:
* gst/librfb/Makefile.am:
* gst/librfb/d3des.c:
* gst/librfb/d3des.h:
* gst/librfb/rfbdecoder.c:
* gst/librfb/vncauth.c:
* gst/librfb/vncauth.h:
VNC Authentication should be working now
temperaly with fake password 'testtest'
Original commit message from CVS:
* gst/librfb/rfbdecoder.c:
* gst/librfb/rfbdecoder.h:
Added some documentation about security handling
start implementing security handling for rfb 3.3
Original commit message from CVS:
* ChangeLog:
Add missing newline.
* gst/librfb/rfbdecoder.c:
Fix the build (missing stdlib.h).
* gst/spectrum/gstspectrum.c:
* gst/spectrum/gstspectrum.h:
Use basetransform segment so that it is correctly managed on flushes
and start/stop. Report message timestamp as stream time, which is what
an application can understand. (Yes these are adapted from wim recent
level element changes)
Original commit message from CVS:
* gst/rtpmanager/gstrtpbin.c: (new_ssrc_pad_found):
Link to the right pads regardless of which one was created first in the
ssrc demuxer.
* gst/rtpmanager/gstrtpjitterbuffer.c:
(gst_rtp_jitter_buffer_chain), (gst_rtp_jitter_buffer_loop):
* gst/rtpmanager/gstrtpsession.c: (gst_rtp_session_process_rtp),
(gst_rtp_session_chain_recv_rtp), (gst_rtp_session_chain_send_rtp):
* gst/rtpmanager/rtpsource.c: (calculate_jitter):
Improve debugging.
* gst/rtpmanager/gstrtpssrcdemux.c: (create_demux_pad_for_ssrc),
(gst_rtp_ssrc_demux_init), (gst_rtp_ssrc_demux_finalize),
(gst_rtp_ssrc_demux_sink_event),
(gst_rtp_ssrc_demux_rtcp_sink_event), (gst_rtp_ssrc_demux_chain),
(gst_rtp_ssrc_demux_rtcp_chain),
(gst_rtp_ssrc_demux_internal_links):
* gst/rtpmanager/gstrtpssrcdemux.h:
Fix race in creating the RTP and RTCP pads when a new SSRC is detected.
Original commit message from CVS:
* gst/rtpmanager/gstrtpbin.c: (gst_rtp_bin_set_property),
(gst_rtp_bin_get_property):
Use lock to protect variable.
* gst/rtpmanager/gstrtpjitterbuffer.c:
(gst_rtp_jitter_buffer_class_init),
(gst_jitter_buffer_sink_parse_caps), (gst_rtp_jitter_buffer_chain),
(convert_rtptime_to_gsttime), (gst_rtp_jitter_buffer_loop):
Reconstruct GST timestamp from RTP timestamps based on measured clock
skew and sync offset.
* gst/rtpmanager/rtpjitterbuffer.c: (rtp_jitter_buffer_init),
(rtp_jitter_buffer_set_tail_changed),
(rtp_jitter_buffer_set_clock_rate),
(rtp_jitter_buffer_get_clock_rate), (calculate_skew),
(rtp_jitter_buffer_insert), (rtp_jitter_buffer_peek):
* gst/rtpmanager/rtpjitterbuffer.h:
Measure clock skew.
Add callback to be notfied when a new packet was inserted at the tail.
* gst/rtpmanager/rtpsource.c: (rtp_source_init),
(calculate_jitter), (rtp_source_send_rtp):
* gst/rtpmanager/rtpsource.h:
Remove clock skew detection, it's move to the jitterbuffer now.
Original commit message from CVS:
* gst/rtpmanager/gstrtpbin.c: (create_session):
Also set NTP base time on new sessions.
* gst/rtpmanager/gstrtpjitterbuffer.c:
(gst_rtp_jitter_buffer_loop), (gst_rtp_jitter_buffer_query),
(gst_rtp_jitter_buffer_set_property),
(gst_rtp_jitter_buffer_get_property):
Use the right lock to protect our variables.
Fix some comment.
* gst/rtpmanager/gstrtpsession.c:
(gst_rtp_session_getcaps_send_rtp),
(gst_rtp_session_chain_send_rtp), (create_send_rtp_sink):
Implement getcaps on the sender sinkpad so that payloaders can negotiate
the right SSRC.
Original commit message from CVS:
* gst/rtpmanager/gstrtpbin.c: (calc_ntp_ns_base),
(gst_rtp_bin_change_state), (new_payload_found), (create_send_rtp):
Calculate and configure the NTP base time so that we can generate better
NTP times in SR packets.
Set caps on new ghostpad.
* gst/rtpmanager/gstrtpjitterbuffer.c:
(gst_rtp_jitter_buffer_loop):
Clean debug statement.
* gst/rtpmanager/gstrtpsession.c: (gst_rtp_session_class_init),
(gst_rtp_session_init), (gst_rtp_session_set_property),
(gst_rtp_session_get_property), (get_current_ntp_ns_time),
(rtcp_thread), (gst_rtp_session_event_recv_rtp_sink),
(gst_rtp_session_internal_links), (gst_rtp_session_chain_recv_rtp),
(gst_rtp_session_event_send_rtp_sink),
(gst_rtp_session_chain_send_rtp), (create_recv_rtp_sink),
(create_send_rtp_sink):
* gst/rtpmanager/gstrtpsession.h:
Add ntp-ns-base property to convert running_time to NTP time.
Handle NEWSEGMENT events on send and recv RTP pads so that we can
calculate the running time and thus NTP time of the packets.
Simplify getting the current NTP time using the pipeline clock.
Implement internal links functions.
Use the buffer timestamp to calculate the NTP time instead of the clock.
* gst/rtpmanager/gstrtpssrcdemux.c: (create_demux_pad_for_ssrc),
(gst_rtp_ssrc_demux_init), (gst_rtp_ssrc_demux_sink_event),
(gst_rtp_ssrc_demux_chain), (gst_rtp_ssrc_demux_rtcp_chain),
(gst_rtp_ssrc_demux_internal_links),
(gst_rtp_ssrc_demux_src_query):
* gst/rtpmanager/gstrtpssrcdemux.h:
Implement internal links function.
Calculate the diff between different streams, this might be used later
to get the inter stream latency.
* gst/rtpmanager/rtpsession.c: (rtp_session_send_rtp):
Simple cleanup.
* gst/rtpmanager/rtpsource.c: (rtp_source_init),
(calculate_jitter), (rtp_source_send_rtp), (rtp_source_get_new_sr):
Make the clock skew window a little bigger.
Apply the clock skew to all buffers, not just one with a new timestamp.
Calculate and debug sender clock drift.
Use extended last timestamp to interpollate for SR reports.
Original commit message from CVS:
* gst/spectrum/gstspectrum.c: (gst_spectrum_start),
(gst_spectrum_transform_ip):
Use the correct parameter order for the memset calls.
Thanks to Christian Schaller for noticing.
Original commit message from CVS:
* configure.ac:
* gst/spectrum/Makefile.am:
* gst/spectrum/demo-audiotest.c: (draw_spectrum),
(message_handler), (main):
* gst/spectrum/demo-osssrc.c: (draw_spectrum), (message_handler):
* gst/spectrum/gstspectrum.c: (gst_spectrum_base_init),
(gst_spectrum_class_init), (gst_spectrum_init),
(gst_spectrum_dispose), (gst_spectrum_set_property),
(gst_spectrum_get_property), (gst_spectrum_start),
(gst_spectrum_setup), (gst_spectrum_message_new),
(gst_spectrum_transform_ip):
* gst/spectrum/gstspectrum.h:
Port GstSpectrum to GstAudioFilter and libgstfft, add support
for int32, float and double, use floats for the message contents,
average all FFTs done in one interval for better results, use
a better windowing function, allow posting the phase in the message
and actually do an FFT with the requested number of bands instead
of interpolating.
* tests/check/elements/spectrum.c: (GST_START_TEST),
(spectrum_suite):
Improve the units tests by checking for a 11025Hz sine wave
and add unit tests for all 4 supported sample types.
Original commit message from CVS:
* gst/real/gstrealvideodec.c: (gst_real_video_dec_chain),
(gst_real_video_dec_setcaps):
Add some more debugging.
Don't set LONG for width/height in caps.
Set correct output buffer size when caps changed.
The custom message sent to the decoder should not include the format and
subformat. Fixes#471554.
