Some traks might contain a redirect rtsp uri inside
hndl atom (which is a dref atom entry). This commit makes qtdemux
post a message when it finds one of these traks and there are
no other traks.
Fixes#597497
Some misinterpreted data could result in posting redirect messages
with empty redirect strings. It is better not to post them.
An example is the file on bug #597497
This avoids the "bad register name `%dil'" compilation errors on 32bit where
because of 'r' gcc puts the value in a general purpose register and then tries
to access the lower part as %dil/%sil which is not existing on 32bit. 'q' requests
a-d registers
This drops frames if they're too late anyway before blending and all
that starts but QoS events are not forwarded upstream. In the future
the QoS events should be transformed somehow and forwarded upstream.
Greatly simplify the IDIT chunk handling by using sscanf
instead of 'manually' parsing. Also replaces strncasecmp and
is_alpha/is_digit with glib versions.
Currently this only works if the kernel size doesn't change, in the future
it will be possible to change the kernel size too without draining
the complete history and without loosing anything.
Partially based on a patch by
Thiago Santos <thiago.sousa.santos@collabora.co.uk>
Make the QtDemuxSample struct smaller by keeping the duration and the pts_offset
as their 32 bit values.
Make some macros to calculate PTS, DTS and duration of a sample.
Deref the sample index less often by keeping a ref to the sample we're dealing
with.
Move sample position checks into qtdemux_parse_samples where we can protect it
with a lock.
Refactor and make an qtdemux_ensure_index function.
Rename qtdemux_do_push_seek to qtdemux_seek_offset in order to avoid confusion
with gst_qtdemux_do_push_seek.
When nsamples_out is larger than nsamples_in, using unsigned
ints lead to a overflow and the resulting value is wrong and
way too large for allocating a buffer. Use signed integers
and returning immediatelly when that happens.
Use more efficient formula that uses less multiplies.
Reduce the amount of scalar code, use MMX to calculate the desired
alpha value.
Unroll and handle 2 pixels in one iteration for improved pairing.
Convert the alpha value to 0->255 when setting and to 0->256 when using as
a scaling factor. This makes sure we can reach the full opacity value of 0xff in
all cases.
Fix some comments, clarify some FIXMEs
Remove the on-ntp-stop signal check now that the jitterbuffer is in
-good and we know that it supports this signal.
Use GstRTPBaseAudioPayload as the base class. This saves a lot of code and fixes
a bunch of problems that were already solved in the base class.
Fixes#853367
Don't make copied in the getter and setter for SDES in the RTPSource. This
avoids a couple of copies of the SDES structure when generating RTCP
packets.
Add a new spspps-interval property to instruct the payloader to insert
SPS and PPS at periodic intervals in the stream.
Rework the SPS/PPS handling so that bytestream and AVC sample code both use the
same code paths to handle sprop-parameter-sets. This also allows to have the AVC
code to insert SPS/PPS like the bytestream code.
Fixes#604913
For some reason latest gcc/binutils accept movzxb here while
movzbl would be correct and is the only thing accepted by older
gcc/binutils.
Fixes bug #604679.
This provides another 7% speedup for the time domain convolution and 1.5%
speedup for the FFT convolution on Mono input.
This optimization assumes that the compiler simplifies calculations
and conditions on constant numbers and unrolls loops with a constant
number of repeats.
This will always use time-domain convolution, which lowers the latency.
With FFT convolution it's always a multiple of the kernel length,
with time domain convolution it's only the pre-latency of the filter kernel.
This provides a great speedup, especially the relationship between kernel
length and processing size is now logarithmic instead of linear. Below a
kernel size of 32 it's a bit slower, afterwards it's much faster:
17 0.788000 -> 0.950000
33 1.208000 -> 1.146000
65 2.166000 -> 1.146000
...
4097 107.444000 -> 1.508000
For sizes smaller 32 the normal time-domain convolution is chosen,
for larger sizes the FFT convolution is automatically used.
Fixes bug #594381.
Remove some redundant calculations, move comparisions out of
inner loops, etc.
This makes the convolution about 3 (!) times faster but
processing time is of course still proportional to the
filter size.
Quicktime uses ISO 639-2 for language codes, but GST_TAG_LANGUAGE
is supposed to hold a ISO 639-1 code, so convert as needed using
the new API from -base.
See #602126.
Matroska uses three-letter ISO 639-2B codes, but GST_TAG_LANGUAGE is
supposed to contain two-letter ISO 639-1 codes, so use new language
code mapping functions in -base to convert between those two as
needed.
Fixes#505823.
When an RTSP extension returns NULL or an empty transport string, just ignore it
and try to get the next possible transport. Fixes playback of RealMedia streams.
Set the current_entry to 0 (instead of -1) in push mode so that we correctly
calculate the current frame number and timestamp.
Add some more debug info and fic the duration debug.
Send pending tags only from the streaming thread, just after we've sent
the newsegment event, not with e.g. flush-start. This not only does the
right thing, but also makes sure we're not trampling over variables set
up in the streaming thread from the seeking thread in case someone tries
to issue a seek just as the demuxer is parsing the headers.
Fixes#601617. Spotted by Ognyan Tonchev.
We ref the buffer before pushing it downstream in order to get the CSRCs of it
after pushing. This causes performance problems when downstream elements want to
change the metadata because the buffer needs to be subbuffered.
Instead, read and store the CSRCs of the buffer in an array before pushing it
and process the array after pushing the buffer. This allows us to remove the
ref/unref pair.
Fixes#603376
We were only setting the suggested buf size for video,
we can set it for audio as well.
This and 195e14529d80ef318ce3a778c1995efb11f266cd
fix an issue that prevented seeking on large avi files
on WMP (non-recent versions).
qtdemux only added the whole stsd atom as 'codec_data'
in its output caps for SVQ3. This patch makes it add
the SEQH (inside a SMI atom) and a gamma field (taken
from the gama atom) if available.
Fixes#587922
Release the jbuf lock before emiting the request-pt-map signal to avoid
deadlocks. We also need to catch the shutdown case when locking again.
Fixes#593354
This allows per band configuration of a peak, low shelf or
high shelf filter, which can be very useful if the band frequencies
and widths are manually configured.
celtdepay : added default framesize(480) channels(1) and clockrate(32000)
depay_setcaps : now gets channels and framesize from string with default value
depay_process : now adds timestamp to outbuf
Added frame_size to GstRtpCeltDepay
Changed some GST_DEBUG to GST_DEBUG_OBJECT or GST_LOG_OBJECT
celtpay : getcaps : gets channel and framesize and sets caps
Added frame-size to static caps for audio/x-celt
* gst/qtdemux/qtdemux_fourcc.h: Add new fourccs for use by the mj2
unpacker.
* gst/qtdemux/qtdemux.c (qtdemux_parse_trak): Unpack JPEG2000 component
mapping and channel definitions from the jp2h header. Will add
component-map and channel-definitions elements to the caps if the
component maps or channel definitions are nonstandard, where standard
order means RGB, 444 packed YUV, or greyscale, with no alpha channel.
Fixes#598915.
AVI format can handle VBR audio provided audio chunks are of fixed duration
(cfr fixed duration video frames). Apply this approach to (always) parsed
raw AAC and (if parsed) to MPEG-1/2 audio.
See #368681.
... but disable this for now. Although it basically works fine,
user experience might be shaky (depending on taste), since there
is no keyframe info in push mode.
Refactor the filter setup. Add two new filters with shelf characteristics for
first and last band. Change gain calculation as recommended in the quoted
document (no qrt needed). Rename variables to match the formulas in the
document.
No need to maintain our own genre table in qtdemux. The genres are
identical to the ID3 genres, so we can just use libgsttag's
gst_tag_id3_genre_get() to look them up.
The performance category is meant to be used to audit codepaths that lead to bad
performance (e.g. copies, conversion that can be avoided).
Remove the event category which is not used.
For vbr audio streams we need to use the number of blocks to calculate the
timestamps.
When the allocation of additional index memory fails, don't throw away what
we had before.
Various cleanups.
Implement scanning of the file when we can parse the index.
Some refactoring of common code.
Cleanups and comments.
Remove some reimplemented code.
Remove index massage code and put a FIXME where we should do something
equivalent later.
Remove some duplicate counters.
Be smarter when updateing the current the timestamp and offset in the stream
because we can reuse previously calculated values when simply go forward one
step.
Correctly set metadata on outgoing buffers.
Add a new function and datastructure to parse and hold the index entries on a
per stream base. Also avoid doing too much work trying to figure out the
timestamps and durations as we can trivially do that later.
Less information in the entries makes them 2 times smaller and not doing too
much work makes this code about 12 times faster than the regular case.
Hook in the new function alongside the existing function for comparison until
the rest of the code is updated to handle the new index datastructure.
For calculating the durations of each sample, we are supposed to add each
duration modulo 1<<32 so make the elapsed time counter a uint32.
Fixes#595942
If it looks like we would be allocating a silly size for our sample
index, just bail out instead of trying to allocate it. Helps with
broken or fuzzed files where we might end up trying to malloc a
couple of hundred MBs otherwise.
Make sure we don't read beyond the atom boundary. Note that the code
behaves slightly differently in the corner case where there is not
enough atom data for the specified number of samples (n_samples_time)
in the atom, but still enough data to fill the pre-allocated index of
n_samples entries: before we would just stop parsing the stts data
and continue, whereas now we will likely error out. This should not
be a problem in practice though. We could maintain the old behaviour
by doing reads with a size check inside the loop if needed.
Use GstByteReader to parse stsz and stsc chunks, and check size of
available data before parsing it, instead of blindly assuming there
will be enough data. Fixes crashes with some fuzzed/broken files.
At the end, Dirac streams have an EOS packet with 0 length.
Don't ever sit in an infinite loop when processing one. Allows
muxing Dirac into mkv to complete successfully.
Add a property to allow control over what event causes a file
to finish being written and a new file start. The default is
the same as before -- each buffer causes a new file to be
written. Added is a case where buffers are written to the
same file until a discontinuity in the stream.
Add a parameter 'ignore-pt' that disables creating a gstrtpptdemux module and
ghosts the pads of gstrtpjitterbuffer instead of the ones of gstrtpptdemux.
Fixes#594490
When we receive a reordered packet with the same timestamp as the previous one
(which can happen for fragmented packets) don't consider the packet as lost but
instead wait for the reordered packet to arrive.
Switch the warning-level, so that a reordering does not get a warning, only
an actual produced lost-packet.
Fixes#594251
In gst_rtspsrc_parse_digest_challenge(), rtspsrc does a g_strndup of the auth
header items and then passes them to gst_rtsp_connection_set_auth_param()
without freeing.
Fixes#594133
When receiving a sync-packet, all sessions with the same cname will be compared
and synced together. In this process, there could still be references to a
session that has been shut down in the meanwhile.
This patch makes sure that these references are removed when shutting down a
session, so that the syncing can be done safely.
Fixes#594283
The priv->clock_rate variable could become -1 between when its checked to not
be -1 and when its used, causing an assertion. Fixed by taking the mutex
earlier in the chain() function.
Fixes#593955
This reintroduces the fix for bug #593391. It also applies it in
gst_rtp_session_sync_rtcp() which has very similar code to
gst_rtp_session_send_rtcp().
When we construct a timestamp that would result in a timestamp that is earlier
than when the packet was received, reset the skew calculation as this is
probably a sign that the sender restarted or paused.
Fixes#593354
Recent changes in gst-plugins-good/gst/effectv prevents it from being compiled
with gcc 3. The problem is that the new code uses preprocessor conditionals
within a macro call which does not work with older versions of gcc.
Fixes bug #593688.
In case of non-interleaved (= sequentially payloaded) streams,
the AU-Index serves little purpose (that is not already covered by
RTP fields). (Broken) Payloaders might consider this field then
to be disregarded and have non spec compliant values, e.g. each
RTP packet having AU-Index 2 (rather than 0). As such, ensure/force
simple sequential sending of non-interleaved streams.
3GPP specs define a number of tags along with precise layout. While these
are normally expected to be found in a container whose major brand is a
3GPP brand, this may also happen when a 3GPP brand is only mentioned as a
compatible brand. Apply some checks, heuristic and fallbacks to extract
such tags as well.
Handle large, invalid or otherwise unusual chunk sizes.
Verify some chunk sizes to be at least the size they are
expected to be and round up some sizes to even number for
e.g. offset administration, which must also be properly
tracked in push mode.
After a BYE packet from a source, stop forwarding the SR packets for lip-sync
to rtpbin. Some senders don't update their SR packets correctly after sending a
BYE and then we break lip-sync. We prefer to let the jitterbuffers drain with
the current lip-sync instead.
Connect to the pad-removed signal of the ptdemux elements so that we remove the
ghostpads for them. Fixes cleanup when going to NULL as well as when releasing
the sinkpads.
Fixes#561752
Remove all individual SDES properties and use one sdes property that takes a
GstStructure instead. This will allow us to add more custom stuff to the SDES
messages later.
Remove all individual SDES properties and use one sdes property that takes a
GstStructure instead. This will allow us to add more custom stuff to the SDES
messages later.
Free the session when all the request pads are released.
Don't mess with the session list in free_session as it is called from a foreach
on that list.
Set the state of the upstream element to NULL first.
See #561752
We usually only get SR packets in our chain function but if an invalid packet
contains the SR packet after the RR packet, we must not fail but simply ignore
the malformed packet.
Fixes#581375
Since neither rtpmanager nor any of the payloaders properly implement
pad allocation, there is no way for the rtpmanager to inform downstream elements
of the new SSRC if there is an SSRC collision. So the warning is emitted all the
time and it is confusing.
Fixes#580144
No short-desc as we have them in the element details.
Also keep things (Makefile.am and sections.txt) sorted.
Reword ambigous returns. No text after since please.
When the number of participants is less than 50, the RFC allows for sending the
BYE packet immediatly instead of using the regular BYE timeout.
Fixes#567828.
Original commit message from CVS:
Patch by: Olivier Crete <tester at tester dot ca>
* gst/rtpmanager/gstrtpsession.c:
(gst_rtp_session_setcaps_send_rtp), (create_send_rtp_sink):
* gst/rtpmanager/rtpsession.c: (rtp_session_set_internal_ssrc):
When an SSRC is found on the caps of the sender RTP, use this as the
internal SSRC. Fixes#565910.
Original commit message from CVS:
* gst/rtpmanager/gstrtpsession.c:
(gst_rtp_session_event_send_rtp_sink),
(gst_rtp_session_getcaps_send_rtp):
* gst/rtpmanager/rtpsession.c: (check_collision),
(rtp_session_schedule_bye_locked), (rtp_session_schedule_bye):
* gst/rtpmanager/rtpsession.h:
Rename a method to better reflect what it really does.
Original commit message from CVS:
* gst/rtpmanager/gstrtpsession.c:
(gst_rtp_session_getcaps_send_rtp):
Use method to get the internal SSRC.
* gst/rtpmanager/rtpsession.c: (rtp_session_class_init),
(rtp_session_set_property), (rtp_session_get_property):
Add property to congiure the internal SSRC of the session.
Fixes#565910.
Original commit message from CVS:
* gst/rtpmanager/rtpsession.c: (rtp_session_set_internal_ssrc):
Only change the SSRC of the session and reset the internal source when
the SSRC actually changed. See #565910.
Original commit message from CVS:
* gst/rtpmanager/rtpsource.c: (rtp_source_init),
(rtp_source_update_caps), (get_clock_rate):
* gst/rtpmanager/rtpsource.h:
When no payload was specified on the caps but there was a clock-rate,
assume the clock-rate corresponds to the first payload type found in the
RTP packets. Fixes#565509.
Original commit message from CVS:
Patch by: Arnout Vandecappelle <arnout at mind dot be>
* gst/rtpmanager/rtpjitterbuffer.c: (rtp_jitter_buffer_reset_skew),
(calculate_skew):
* gst/rtpmanager/rtpjitterbuffer.h:
Keep track of the last outgoing timestamp and of the last sender-side
time. Timestamps can only go forward if they do at the sender
side, can only go back if they do at the sender side, and remain the
same if they remain the same at the sender side. Fixes#565319.
Original commit message from CVS:
* gst/rtpmanager/rtpsession.c: (obtain_source),
(rtp_session_create_source), (rtp_session_process_rtp),
(rtp_session_process_sr), (rtp_session_process_rr),
(rtp_session_process_sdes), (rtp_session_process_bye):
Make obtain_source return an aditional ref so that we don't lose our ref
to it when a session cleanup occurs when we are emiting a signal.
Emit the on_new_ssrc signal for the CSRC, not the SSRC.
Fixes#562319.
Original commit message from CVS:
* gst/rtpmanager/gstrtpbin.c: (gst_rtp_bin_reset_sync),
(gst_rtp_bin_clear_pt_map):
Reset the sync parameters when clearing the payload type map too.
Fixes#562312.
Original commit message from CVS:
* gst/rtpmanager/gstrtpbin.c: (get_client),
(gst_rtp_bin_reset_sync), (gst_rtp_bin_associate),
(gst_rtp_bin_handle_sync), (create_stream),
(gst_rtp_bin_class_init), (new_ssrc_pad_found):
* gst/rtpmanager/gstrtpbin.h:
Remove a lot of per stream state that is not needed and pass new info in
the method call.
Add signal to reset sync parameters.
Avoid parsing the caps to get a clock_base, we get this from the sync
signal now.
Original commit message from CVS:
* gst/rtpmanager/rtpsession.c: (rtp_session_class_init),
(rtp_session_init), (rtp_session_set_property),
(rtp_session_get_property):
Add property to configure the RTCP MTU.
Original commit message from CVS:
* gst/rtpmanager/rtpsession.c: (rtp_session_class_init),
(copy_source), (rtp_session_create_sources),
(rtp_session_get_property):
Add G_PARAM_STATIC_STRINGS.
Add property to return a GValueArray of all known RTPSources in the
session.
* gst/rtpmanager/rtpsource.c: (rtp_source_class_init),
(rtp_source_create_sdes), (rtp_source_set_property),
(rtp_source_get_property):
Remove properties to set the various SDES items, an application is never
supposed to change the RTPSource data.
Change the SDES getter properties to one SDES property that returns all
SDES items in a GstStructure.
Original commit message from CVS:
Patch by: Olivier Crete <tester at tester dot ca>
* gst/rtpmanager/gstrtpbin.c: (gst_rtp_bin_release_pad):
Release the right pads on rtpbin. Fixes#561752.
Original commit message from CVS:
* gst/rtpmanager/gstrtpsession.c: (get_current_times),
(rtcp_thread), (gst_rtp_session_chain_recv_rtp):
Pass the running time to the session when processing RTP packets.
Improve the time function to provide more info.
* gst/rtpmanager/rtpsession.c: (rtp_session_class_init),
(rtp_session_init), (update_arrival_stats),
(rtp_session_process_rtp), (rtp_session_process_sdes),
(rtp_session_process_rtcp), (session_start_rtcp),
(rtp_session_on_timeout):
* gst/rtpmanager/rtpsession.h:
Mark the internal source with a flag.
Use running_time instead of the more useless timestamp.
Validate a source when a valid SDES has been received.
Pass the current system time when processing SR packets.
* gst/rtpmanager/rtpsource.c: (rtp_source_class_init),
(rtp_source_init), (rtp_source_create_stats),
(rtp_source_get_property), (rtp_source_send_rtp),
(rtp_source_process_rb), (rtp_source_get_new_rb),
(rtp_source_get_last_rb):
* gst/rtpmanager/rtpsource.h:
Add property to get source stats.
Mark params as STATIC_STRINGS.
Calculate the bitrate at the sender SSRC.
Avoid negative values in the round trip time calculations.
* gst/rtpmanager/rtpstats.h:
Update some docs and change some variable name to more closely reflect
what it contains.
Original commit message from CVS:
* gst/rtpmanager/gstrtpjitterbuffer.c:
(gst_rtp_jitter_buffer_chain_rtcp):
Initialize return value to fix compiler warning about uninitialized
variable.
Original commit message from CVS:
* gst/rtpmanager/gstrtpbin.c: (gst_rtp_bin_associate),
(gst_rtp_bin_handle_sync), (create_stream), (free_stream),
(new_ssrc_pad_found):
Remove internal sync pad, use signals instead to get lip-sync
notifications.
* gst/rtpmanager/gstrtpjitterbuffer.c:
(gst_rtp_jitter_buffer_base_init),
(gst_rtp_jitter_buffer_class_init),
(gst_rtp_jitter_buffer_internal_links), (create_rtcp_sink),
(remove_rtcp_sink), (gst_rtp_jitter_buffer_request_new_pad),
(gst_rtp_jitter_buffer_release_pad),
(gst_rtp_jitter_buffer_sink_rtcp_event),
(gst_rtp_jitter_buffer_chain_rtcp),
(gst_rtp_jitter_buffer_get_property):
* gst/rtpmanager/gstrtpjitterbuffer.h:
Make it possible to send SR packets to the jitterbuffer.
Check if the SR timestamps are valid by comparing them to the RTP
timestamps.
Signal the SR packet and the timing information to listeners.
* gst/rtpmanager/gstrtpssrcdemux.c: (create_demux_pad_for_ssrc),
(gst_rtp_ssrc_demux_rtcp_chain), (gst_rtp_ssrc_demux_src_query):
Remove some unused code.
* gst/rtpmanager/rtpjitterbuffer.c: (rtp_jitter_buffer_reset_skew),
(calculate_skew), (rtp_jitter_buffer_get_sync):
* gst/rtpmanager/rtpjitterbuffer.h:
Keep track of the last seen RTP timestamp so that we can filter out
invalid SR packets.
Original commit message from CVS:
* gst/rtpmanager/rtpsource.c: (get_clock_rate):
Fix GST_DEBUG call to only have as many arguments as required
by the format string. Fixes a compiler warning.
Original commit message from CVS:
* gst/rtpmanager/gstrtpbin.c: (gst_rtp_bin_associate),
(gst_rtp_bin_sync_chain), (create_stream), (new_ssrc_pad_found):
Do not try to keep track of the clock-rate ourselves but simply get the
value from the jitterbuffer.
* gst/rtpmanager/gstrtpjitterbuffer.c:
(gst_jitter_buffer_sink_parse_caps), (gst_rtp_jitter_buffer_chain),
(gst_rtp_jitter_buffer_get_sync):
* gst/rtpmanager/gstrtpjitterbuffer.h:
Add some debug info.
Pass the clock-rate to the jitterbuffer.
Also pass the clock-rate along with the rtp timestamp when getting the
sync parameters.
* gst/rtpmanager/gstrtpptdemux.c: (gst_rtp_pt_demux_chain):
Fix some debug.
* gst/rtpmanager/rtpjitterbuffer.c: (rtp_jitter_buffer_reset_skew),
(calculate_skew), (rtp_jitter_buffer_get_sync):
* gst/rtpmanager/rtpjitterbuffer.h:
Keep track of clock-rate changes and return the clock-rate together with
the rtp timestamps used for sync.
Don't try to construct timestamps when we have no base_time.
* gst/rtpmanager/rtpsource.c: (get_clock_rate):
Request a new clock-rate when the payload type changes.
Reset the jitter calculation when the clock-rate changes.
