Windows XP
Still good:
avidemux_2.6.19_win32.exe
Regressed:
avidemux_2.6.19_r170410_win32.exe
(0. Fwiw, .ts.idx2 are identicals.)
1. (Select a range, then) Save (Copy/Copy/TS).
1r. Video is fine, but "no" sound.
Good log:
a few lines like
Quote
[adm_lavLogCallback] 14:24:14-892 [lavc] Application provided invalid, non monotonically increasing dts to muxer in stream 1: 30720 >= 14081
[saveLoop] 14:24:14-892 [FF]Error writing audio packet
where "A > b".
Bad log:
lots of lines like
Quote
[adm_lavLogCallback] 14:28:21-516 [lavc] Application provided invalid, non monotonically increasing dts to muxer in stream 1: 73841 >= 73841
[saveLoop] 14:28:21-516 [FF]Error writing audio packet
where "a = b".
NB:
I'm not sure I saved the exact same range(s). Then the quoted logs may not be an appropriate hint. (But the issue is real.)
A sample would help
10 MiB (start of file) sample: http://dl.free.fr/bewdxoyeI (http://dl.free.fr/bewdxoyeI)
Password: 2042
NB:
I assume the issue is not specific to this file.
[initFaad] 03:47:00-614 [FAAD] using 2 bytes of extradata
13 10
[initFaad] 03:47:00-614 [FAAD] Found :48000 rate 2 channels
[initFaad] 03:47:00-614 [FAAD] Frequency mismatch!!! 24000 to 48000 (SBR ?)
[initFaad] 03:47:00-614 Sbr detected
[setupAudio] 03:47:00-614 Setting up 1 audio track(s)
[setupAudio] 03:47:00-614 [audioTrack 0] Creating audio encoding stream, starttime 00:00:01,960(copy)
[audioCreateCopyStream] 03:47:00-614 Creating audio stream copy with compensation : startTime=00:00:01,960
[audioCreateCopyStream] 03:47:00-614 and shift =00:00:00,000
[ADM_audioStreamCopy] 03:47:00-614 Creating copy stream, startTime=00:00:01,960, shift=0
[goToTime] 03:47:00-614 go to time 1,96 secs
[goToTime] 03:47:00-614 => seg 0, rel time 1,96 secs
[resetAfterSeek] 03:47:00-614 Resetting faad
[initFaad] 03:47:00-614 [FAAD] using 2 bytes of extradata
13 10
[initFaad] 03:47:00-614 [FAAD] Found :48000 rate 2 channels
[initFaad] 03:47:00-614 [FAAD] Frequency mismatch!!! 24000 to 48000 (SBR ?)
[initFaad] 03:47:00-614 Sbr detected
[FF] Muxer opened
[FF] Using 0 bytes for video extradata
[FF] Video initialized
[rescaleFps] 03:47:00-614 TimeBase for video 1000/25000
[FF] Bitrate 128
[initAudio] 03:47:00-614 Language for track 0 is unknown
[FF] Audio initialized
with build 827b282 and above sample:
[FF] Saving
[saveLoop] 03:47:00-615 avg fps=25000
[initUI] 03:47:00-615 Muxer, creating UI, video duration is 00:00:02,560
[adm_lavLogCallback] 03:47:00-640 [lavc] Timestamps are unset in a packet for stream 1. This is deprecated and will stop working in the future. Fix your code to set the timestamps properly
[adm_lavLogCallback] 03:47:00-640 [lavc] Encoder did not produce proper pts, making some up.
[adm_lavLogCallback] 03:47:00-661 [lavc] Application provided invalid, non monotonically increasing dts to muxer in stream 1: 12641 >= 12641
[saveLoop] 03:47:00-661 [FF]Error writing audio packet
result: intermittent audio dropouts
Video
Frame rate : 25.000 fps
Audio
ID : 318 (0x13E)
Menu ID : 1882 (0x75A)
Format : AAC
Format/Info : Advanced Audio Codec
Format version : Version 2
Format profile : HE-AAC / LC
Muxing mode : ADTS
Codec ID : 15
Duration : 42s 240ms
Bit rate mode : Variable
Channel(s) : 2 channels
Channel positions : Front: L R
Sampling rate : 48.0 KHz / 24.0 KHz
Frame rate : 23.438 fps (1024 spf)
Compression mode : Lossy
Delay relative to video : -196ms
2 problems
- holes in timestamp : libavformat rejects it now
- AAC/HE not managed correcly
avidemux_2.6.99_r170411_win32.exe
Much better after
[muxer/ffmpeg] Try to fill in the missing timestamps, AAC/HE will not work well (https://github.com/mean00/avidemux2/commit/52f899f852cdeeac9b5affd55354cebd7f2d43ca)
But still not good:
long sound, brief blank, long sound, brief blank, ...
avidemux_2.7.0_r170412_win32.exe
(Fully) Fixed after
[core/AudioSTream] avoid diverging pts when dealing with AAC (https://github.com/mean00/avidemux2/commit/ba50f03114ad5d07b442959899bc2869e8fb6b0d)
:)
Ftr, I now get (only)
Quote
...
Paint event
[audioClock] Drift detected :0 vs 820466, delta=820466
[audioClock] Drift detected :1033799 vs 1076477, delta=42678
[audioClock] Drift detected :36703144 vs 36660488, delta=-42656
[audioClock] Drift detected :36873821 vs 36916488, delta=42667
Percent:93
...
Is this the best that can be expected, or would there be some additional tweak(s) to further improve?
avidemux_2.6.19_r170413_win32.exe
Includes
(support) [muxer/ffmpeg] Try to fill in the missing timestamps, AAC/HE will not work well (https://github.com/mean00/avidemux2/commit/41564514d90d08e8ab94ee3d9fd936215eb8557e)
But still misses the second part(s) of the fix.
development is only happening on 2.7.0. 2.6.19 is halted, as far as I understand.
https://github.com/mean00/avidemux2/branches
Quote from: Jan Gruuthuse on April 14, 2017, 04:43:07 AM
2.6.19 is halted, as far as I understand.
Hum, the opposite is written in
What is the new "support" branch? (2017.04.13) (http://avidemux.org/smif/index.php/topic,17679.0.html)
Confused...
Updating ffmpeg takes a long time to stabilize
2.7 = new ffmpeg
2.6 = old ffmpeg
avidemux_2.6.19_r170420_win32.exe
(Fully) Fixed after
(support) [core/AudioSTream] avoid diverging pts when dealing with AAC (https://github.com/mean00/avidemux2/commit/75c571ef43f934a0c4423897be81448daccf172e)
:)