Mythtranscode uses the wrong audio track in the output

[anonymous]

I'm having an issue that is similar to ticket #6626. A couple of times a week (especially for certain shows), mythtranscode using a Lossless profile generates a output file with the wrong audio track.

mythtranscode -v all --profile Lossless --video -i survivor.mpg -o survivor-transcode.mpg

The input file is an MPEG2-TS (from HRHomeRun) with 2 audio tracks. In the original, audio track #1 is 6 channel AC3 while audio track #2 is 2 channel. Audio track #2 is easy to identify since it plays back with a slight A/V sync issue. Audio track #1 is in sync (and 5.1 of course):

2010-05-09 06:25:17.348 Opening /data/recordings/RecordedTV/survivor.mpg
2010-05-09 06:25:17.389 Input #0, mpegts, from '/data/recordings/RecordedTV/survivor.mpg':
2010-05-09 06:25:17.390   Duration: 00:59:56.91, start: 65258.976767, bitrate: 11163 kb/s
2010-05-09 06:25:17.390     Stream #0.0[0x31]: Video: mpeg2video, yuv420p, 1920x1080 [PAR 1:1 DAR 16:9], 1001/60000, 65000 kb/s, 59.94 tbr, 90k tbn, 59.94 tbc
2010-05-09 06:25:17.390     Stream #0.1[0x34]: Audio: ac3, 48000 Hz, 5.1, s16, 384 kb/s
2010-05-09 06:25:17.390     Stream #0.2[0x35]: Audio: ac3, 48000 Hz, stereo, s16, 128 kb/s
2010-05-09 06:25:17.390     Stream #0.3[0x120]: Data: 0x0101
2010-05-09 06:25:17.390 Skipping unsupported codec 2 on stream 3
2010-05-09 06:25:17.400 1920x1088 P:900900 02:22:31:004 #0 fl:2IT_0 pos: 376
2010-05-09 06:25:17.400 Found 1st valid video frame

What makes this ticket different from #6626 is the log seems to show that audio track #1 is being selected properly:

2010-05-09 06:25:17.305 Using 1 CPUs for decoding
2010-05-09 06:25:17.305 AFD: InitVideoCodec() 0x8362960 id(MPEG2VIDEO) type (Video).
2010-05-09 06:25:17.305 detectInterlace(Detect Scan, Interlaced Scan, 29.97, 1080) ->Interlaced Scan
2010-05-09 06:25:17.305 AFD: EIA-608 caption 1 is in the English language.
2010-05-09 06:25:17.305 AFD: Using ffmpeg for video decoding
2010-05-09 06:25:17.305 AFD: Looking for decoder for MPEG2VIDEO
2010-05-09 06:25:17.305 AFD: Opened codec 0x8362960, id(MPEG2VIDEO) type(Video)
2010-05-09 06:25:17.305 AFD: Stream #1, has id 0x34 codec id AC3, type Audio, bitrate 384000 at 0x8362d20
2010-05-09 06:25:17.306 AFD: codec AC3 has 6 channels
2010-05-09 06:25:17.306 AFD: Looking for decoder for AC3
2010-05-09 06:25:17.306 AFD: Opened codec 0x8367400, id(AC3) type(Audio)
2010-05-09 06:25:17.306 AFD: Audio Track #1 is A/V stream #1 and has 6 channels in the Undetermined language(7695972).
2010-05-09 06:25:17.306 AFD: Stream #2, has id 0x35 codec id AC3, type Audio, bitrate 128000 at 0x83677c0
2010-05-09 06:25:17.306 AFD: codec AC3 has 2 channels
2010-05-09 06:25:17.306 AFD: Looking for decoder for AC3
2010-05-09 06:25:17.307 AFD: Opened codec 0x8367990, id(AC3) type(Audio)
2010-05-09 06:25:17.307 AFD: Audio Track #2 is A/V stream #2 and has 2 channels in the Undetermined language(7695972).
2010-05-09 06:25:17.307 AFD: Stream #3, has id 0x120 codec id DVB_VBI, type Data, bitrate 0 at 0x8367f30
2010-05-09 06:25:17.313 AFD: data codec (Data)
2010-05-09 06:25:17.313 RingBuf(/data/recordings/RecordedTV/survivor.mpg): CalcReadAheadThresh(0 KB)
			 -> threshhold(64 KB) min read(0 KB) blk size(32 KB)
2010-05-09 06:25:17.313 AFD: Trying to select audio track (w/lang)
2010-05-09 06:25:17.313 AFD: Trying to select audio track (wo/lang)
2010-05-09 06:25:17.313 AFD: Selected track 1: Undetermined AC3 5.1ch (A/V Stream #1)
2010-05-09 06:25:17.313 AFD: Initializing audio parms from audio track #1
2010-05-09 06:25:17.313 AFD: Audio format changed 
			from id(NONE)    -1Hz -1ch   0bps    ; id(NONE)    -1Hz -1ch   0bps   
			to   id( AC3) 48000Hz  6ch  16bps    ; id( AC3) 48000Hz  2ch  16bps   
2010-05-09 06:25:17.313 Dec: Selected track #1 in the English language(6647399)
2010-05-09 06:25:17.313 Dec: Resyncing position map. posmapStarted = 0 livetv(0) watchingRec(0)

The output file contains a single audio track. However, it's the 2 channel audio with the A/V sync issues. So, it appears that audio track #2 is actually being used.

I'll be happy to attach portions of the original file or output file if it would help.

Thank you for your help!

[robertm]

Dupe of #8180, #6626, and probably a half dozen others.