Let's not put every single mp3 frame in our index, a few frames per
second should be more than enough. For now use an index interval
of 100ms-500ms depending on the upstream size, to keep the index at
a reasonable size. Factor out the code that adds the index entry
into a separate function for better code readability.
While technically upstream may be seekable even if it doesn't know
the exact size, I can't think of a use case where this distincation
is relevant in practice, so for now just assume we're not seekable
if upstream doesn't provide us with a size. Makes sure we don't
build a seek index when streaming internet radio with sources that
pretend to be seekable until you try to actually seek.
Don't introduce glitches in the output by a) relaxing the threshold for
taking upstream timestamps in preference to our calculated timestamps and
b) only set the discont flag on outgoing buffers in response to an incoming
discont buffer.
Fixes: #575046
Original commit message from CVS:
* gst/mpegaudioparse/gstmpegaudioparse.c: (channel_mode_class),
(GST_TYPE_MP3_CHANNEL_MODE), (mp3_type_frame_length_from_header),
(gst_mp3parse_emit_frame), (mp3parse_get_query_types):
* gst/mpegaudioparse/gstmpegaudioparse.h:
Do an initial class_ref on an internal enum type from within the
class_init function so that there aren't any issues when multiple
mp3parse elements are started in separate threads at the same
time. (Why we use an enum type here if the tag is registered as
a string type, I don't know). Also remove custom UNUSED macro
and use GLib's instead.
Original commit message from CVS:
* ext/mad/gstmad.c: (gst_mad_dispose), (gst_mad_sink_event),
(gst_mad_chain):
* ext/mad/gstmad.h:
* gst/mpegaudioparse/gstmpegaudioparse.c: (gst_mp3parse_dispose),
(gst_mp3parse_sink_event), (gst_mp3parse_emit_frame):
* gst/mpegaudioparse/gstmpegaudioparse.h:
Cache all events except EOS if we still have to send a NEWSEGMENT
event. This will let TAG events be forwarded until after decodebin
to an encoder for example as decodebin only links the pads
after NEWSEGMENT. Fixes bug #518933.
Original commit message from CVS:
* gst/mpegaudioparse/gstmpegaudioparse.c: (gst_mp3parse_reset),
(gst_mp3parse_handle_first_frame), (mp3parse_time_to_bytepos),
(mp3parse_bytepos_to_time):
* gst/mpegaudioparse/gstmpegaudioparse.h:
Add support for seeking based on the VBRI seek table. Might make
sense to use interpolation in the table later to get hopefully a
bit more accurate values.
Original commit message from CVS:
* gst/mpegaudioparse/gstmpegaudioparse.c: (gst_mp3parse_reset),
(gst_mp3parse_emit_frame), (gst_mp3parse_handle_first_frame),
(mp3parse_total_bytes), (mp3parse_total_time):
* gst/mpegaudioparse/gstmpegaudioparse.h:
Add initial support for reading VBRI headers as found in VBR files
created by some Fraunhofer encoders. Currently we only read the
number of frames and bytes (and calculate duration, etc from this)
but there is also a seek table that we currently don't use.
Original commit message from CVS:
* gst/mpegaudioparse/gstmpegaudioparse.c:
(gst_mp3parse_sink_event), (gst_mp3parse_emit_frame):
* gst/mpegaudioparse/gstmpegaudioparse.h:
Queue segment event and push it after we know the caps on the pad or
else an autoplugger might not have plugged the element yet and the
segment is lost.
Original commit message from CVS:
* gst/mpegaudioparse/gstmpegaudioparse.c: (gst_mp3parse_reset),
(gst_mp3parse_emit_frame), (mp3parse_handle_seek):
* gst/mpegaudioparse/gstmpegaudioparse.h:
Save some memory for each frame by only saving the start timestamp
and start byte position instead of additionally the stop timestamp
and stop byte position. This requires us to use a doubly-linked list
but still saves 8-12 bytes per frame.
Original commit message from CVS:
* gst/mpegaudioparse/gstmpegaudioparse.c: (gst_mp3parse_base_init),
(gst_mp3parse_init):
Use GST_BOILERPLATE instead of manual GType magic.
* gst/mpegaudioparse/gstmpegaudioparse.c:
(gst_mp3parse_handle_first_frame), (mp3parse_time_to_bytepos),
(mp3parse_bytepos_to_time):
* gst/mpegaudioparse/gstmpegaudioparse.h:
Implement seeking, byte->time, time->byte conversions with the Xing
seek table if available. This allows better at least a bit more
accurate seeks and file position reporting.
Original commit message from CVS:
* gst/mpegaudioparse/gstmpegaudioparse.c: (gst_mp3parse_reset),
(gst_mp3parse_emit_frame), (gst_mp3parse_handle_first_frame):
* gst/mpegaudioparse/gstmpegaudioparse.h:
If the Xing header provides a total time, use it to calculate the
correct average bitrate immediately, instead of sending updates as
we parse the stream.
Original commit message from CVS:
* gst/mpegaudioparse/gstmpegaudioparse.c:
(mp3_type_frame_length_from_header), (gst_mp3parse_reset),
(gst_mp3parse_emit_frame), (gst_mp3parse_handle_first_frame),
(gst_mp3parse_chain), (mp3parse_total_bytes),
(mp3parse_total_time):
* gst/mpegaudioparse/gstmpegaudioparse.h:
Implement parsing of Xing headers from the first frame of the stream,
and use it to report duration correctly where possible.
Original commit message from CVS:
* gst/mpegaudioparse/gstmpegaudioparse.c:
(mp3_type_frame_length_from_header), (gst_mp3parse_reset),
(gst_mp3parse_init), (gst_mp3parse_sink_event),
(gst_mp3parse_emit_frame), (gst_mp3parse_chain),
(gst_mp3parse_change_state), (mp3parse_time_to_bytepos),
(mp3parse_bytepos_to_time), (mp3parse_total_bytes),
(mp3parse_total_time), (mp3parse_handle_seek),
(mp3parse_src_event), (mp3parse_src_query),
(mp3parse_get_query_types), (plugin_init):
* gst/mpegaudioparse/gstmpegaudioparse.h:
Implement seeking via average bitrate, and position+duration
querying in mp3parse. Later, it will support frame-accurate seeking by
building a seek table as it parses.
Add 'parsed=false' to the sink pad caps, and 'parsed=true' to the src
pad caps. Bump the priority to PRIMARY+1 so that it is autoplugged
before any extant MP3 decoder plugin. This allows us to remove framing
support from the decoders, if we want, and will provide them with
accurate seeking automatically once it is finished.
Fix the handling of MPEG-1 Layer 1 files.
Partially fix timestamping of packets arriving from a demuxer by
queueing the incoming timestamp until the next packet starts, rather
than applying it immediately to the next pushed buffer.
Original commit message from CVS:
* gst/mpegaudioparse/gstmpegaudioparse.c:
(mp3_type_frame_length_from_header), (gst_mp3parse_reset),
(gst_mp3parse_init), (gst_mp3parse_dispose),
(gst_mp3parse_sink_event), (gst_mp3parse_chain), (head_check),
(gst_mp3parse_change_state):
* gst/mpegaudioparse/gstmpegaudioparse.h:
Make timestamp handling in mp3parse saner; now works for at least
simple cases.
Original commit message from CVS:
* gst/mpegaudioparse/Makefile.am:
* gst/mpegaudioparse/gstmpegaudioparse.c:
(mp3_type_frame_length_from_header), (gst_mp3parse_class_init),
(gst_mp3parse_reset), (gst_mp3parse_init), (gst_mp3parse_dispose),
(gst_mp3parse_sink_event), (gst_mp3parse_chain), (head_check),
(gst_mp3parse_change_state), (plugin_init):
* gst/mpegaudioparse/gstmpegaudioparse.h:
Bring mp3parse into the 21st century.
Use its own debug category, use gstadapter, format nicely to 80
columns, and fix incorrect handling of 32 kHz and less files.
Original commit message from CVS:
* gst/mpegaudioparse/gstmpegaudioparse.c: (gst_mp3parse_init),
(gst_mp3parse_sink_event), (gst_mp3parse_chain):
* gst/mpegaudioparse/gstmpegaudioparse.h:
Put timestamps on buffers.
Original commit message from CVS:
Remove all config.h includes from header files, add it to each source file and remove duplicate config.h includes from several source files
Original commit message from CVS:
New mimetypes gone into effect today - this commit changes all old mimetypes over to the new mimetypes spec as described in the previous commit's document. Note: some plugins will break, some pipelines will break, expect HEAD to be broken or at least not 100% working for a few days, but don't forget to report bugs
Original commit message from CVS:
* removal of //-style comments
* don't link plugins to core libs -- the versioning is done internally to the plugins with the plugin_info struct,
and symbol resolution is lazy, so we can always know if a plugin can be loaded by the plugin_info data. in theory.
