-
Notifications
You must be signed in to change notification settings - Fork 345
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
mpegts-mythtv harmonization #927
base: master
Are you sure you want to change the base?
Conversation
@kmdewaal I am trying to reduce the changes regarding how the |
They are const in mpegts.c.
The only STREAM_TYPE_* define used is STREAM_TYPE_PRIVATE_DATA. STREAM_TYPE_AUDIO_DTS is already defined by FFmpeg to a different value. 0x8a is used in MISC_types[]. STREAM_TYPE_AUDIO_HDMV_* defines are from: MythTV@d15e482 They are now unused and FFmpeg has added the entries to HDMV_types[].
STREAM_TYPE_SUBTITLE_DVB is from: MythTV@aaa9372#diff-ae852538fd61b663231c8bdda1fdbade40c97e8305401fcc6f1ad9f395efdf18 STREAM_TYPE_VBI_DVB is from: MythTV@6da1854 However, the corresponding uses in MISC_types[] can never be used. in mpegts_set_stream_info() pes->stream_type = stream_type; ... mpegts_find_stream_type(st, pes->stream_type, MISC_types); mpegts_set_stream_info() is called either with stream_type = 0 or stream_type = get8(&p, p_end) 0x100 and 0x101 exceed UINT8_MAX and thus will never be found in the search.
Use the names given in ETSI 300 468 and list all of the values used, even if the defines aren't.
I'm not sure that the new name is correct since I don't have a copy of ISO/IEC 13818-6. However, it should be close.
Appears to be from 2006: MythTV@156914d MythTV@53d9721 Disabled in 2010: MythTV@3cc5fc8#diff-2fb8a33ba832f242d27f19c9f25755b57af9807343e99dac77263c1c5ebc383a
Neither removed section has done anything since 2010: MythTV@3cc5fc8#diff-2fb8a33ba832f242d27f19c9f25755b57af9807343e99dac77263c1c5ebc383a
Appears to have been from an FFmpeg merge in 2009: MythTV@6d62e66
…ptor This has no functional change since MythTV only uses the MythTV addition AVStream::carousel_id if AVStream::codecpar::codec_id == AV_CODEC_ID_DSMCC_B.
originally from: MythTV@a1d4d11 referencing: https://code.mythtv.org/trac/ticket/1887 ISO/IEC 13818-1:2021 specifies a valid range of [0x0010, 0x1FFE] in § 2.4.4.6 Semantic definition of fields in program association section and Table 2-3 – PID table
…ng, and pmt_scan_state req_sid is from an FFmpeg merge: MythTV@2a56a37#diff-2fb8a33ba832f242d27f19c9f25755b57af9807343e99dac77263c1c5ebc383a removed from FFmpeg in FFmpeg/FFmpeg@90d13e3 scanning and pmt_scan_state are from MythTV@084e3f8 MpegTSContext::req_sid is only set by mpegts_read_header(). First, in pmt_cb() always create PMT filters. Now we don't want to stop_parse early since that would prevent creating all the PMT filters, so remove those changes to pat_cb() and pmt_cb(). MpegTSContext::pmt_scan_state is only tested in mpegts_read_header() and its use is equivalent to `!ts->prg[i].pmt_found`. Now MpegTSContext::scanning and MpegTSContext::pmt_scan_state are unused, so remove them. MpegTSContext::req_sid is now set but otherwise unused, so remove it as well. Additionally, in mpegts_read_header() sid would cause an invalid index of -1 on the first iteration. The fallback code in mpegts_read_header() has been broken since MythTV@0b0068c sid != pmt_pid, so it was looking for the wrong filter in MpegTSContext::pids.
ca5a144
to
ed1722e
Compare
Nine new commits further reduce our customization and also improve handling of MPEG Transport Streams with multiple programs. |
Disabling CRC is from: MythTV@084e3f8 referencing https://code.mythtv.org/trac/ticket/328 However, the sample does not trigger that fallback. Deal with incomplete PMT streams in BBC iPlayer IPTV MythTV@c11ee69 references https://code.mythtv.org/trac/ticket/9926 The sample is unavailable and the issue was not in FFmpeg's unmodified demuxer, so, following subsequent harmonization, I assume it is no longer an issue. If it still is, MythTV's other modifications should be looked at first.
There is one PMT per program, so add the stream_type to the corresponding Program. This means we no longer need to look at an MpegTSFilter. The searching logic in is_pmt_equal() is equivalent to that in pmt_equal_streams(). However, we now look at the entire PMT not just is_desired_stream()s. Nothing now uses MpegTSContext::pid_cnt or MpegTSContext::pmt_pids, so remove them and a lot of other now unused code. See ISO/IEC 13818-1:2021 § 2.4.4.9 Program map table Perhaps we could use TS_program_map_section()'s version_number and possibly current_next_indicator instead of comparing the PMT ourselves? Maybe by modifying update_av_program_info()? export_pmt() would probably need if (old_version == -1 || old_version != version) also its buffer should probably be moved to AVProgram since each program has a different PMT. Similarly streams_changed() should have a program_number parameter, since we probably only care about one program at a time.
This can be a problem when there is more than one program in one file, e.g. a full TS capture with multiple digital subchannels. The `#if`ed out log in AutoSelectAudioTrack was broken; I replaced it with something that should be similar.
This is clearer than the infinite loop with breaks (effectively gotos).
ed1722e
to
4ce4353
Compare
This is a MythTV addition that exports the PMT from an MPEG-TS. However, each program has its own PMT so having a single pmt_section in AVFormatContext is not correct when there are multiple programs in the stream. Previously, each PMT overwrote the last one. Adapt AvFormatDecoder to look in AVProgram. In my limited testing, the PMT did not seem necessary for ATSC captions. MythAVBufferRef: Satisfy the rule of five; however, since nullptr is now allowed, the newly defined constructors and operators are unused. The compiler generated ones were causing use after free segmentation faults.
This is an int to match type with AVProgram. In mpegts.c, SectionHeader::id is a uint16_t as ISO/IEC 13818-1:2021 specifies for program_number. This allows us to ignore stream changes in programs that are not being watched when there is more than one program. Regarding the changes to mpeg.c: Originally from: dvb/ac3 patches 1, 2, and 3 from Mark Anderson MythTV@87795f6 changed in: Refs #8134. internal dvd player. resolve problem where pcm_s16be audi… MythTV@e64371a change reverted in: Refs #8134. revert most of [24239]. Anduin withers reported it makes … MythTV@a3d6e14 referencing: #8134 (internal dvd player: pcm_s16be improperly detected as pcm_dvd audio codec.) – MythTV https://code.mythtv.org/trac/ticket/8134 An MPEG Program Stream can only have one program, so FFmpeg does not create an AVProgram for it. Thus av_find_program_from_stream() (and therefore get_current_AVProgram()) will always return nullptr, so avprogram_id will be ignored. The calling conventions of functions with "C" and "C++" language linkage are not necessarily the same; however, in practice they are, so AvFormatDecoder::streams_changed() could be used directly. Doing it correctly still eliminates the use of a friend function and its double declaration at only the cost of an extra function call, which doesn't matter since streams_changed() is only called very rarely.
by using the PMT's version_number (and current_next_indicator) to determine when to call streams_changed(). (FFmpeg already checks the current_next_indicator in pmt_cb() before the call to update_av_program_info().) The only change is the first seen PMT for each program will now never call streams_changed(). is_pmt_equal() would return 0 for the first seen PMT for each program when opening the file, but streams_changed is null at that point. This may prevent a call to streams_changed() when a new PAT version adds a program. However, MythTV does not fully support multiple programs in a single transport stream, so this is acceptable. The new PAT may also remove the currently playing program, which the PMT code cannot detect.
b99b409
to
3e50172
Compare
I fixed a typo in commit message "used" was supposed to be "unused". The new commit changes when the |
It does nothing since pmt_cb() already closes the already existing filter before calling and the call in handle_packet() will only occur when `ts->pids[pid] == NULL`, i.e. there is no filter to close. It was originally from: PMT tracking in ffmpeg patch from danielk. MythTV@236872a
This has no functional change and reduces the changes to FFmpeg.
Checklist