--add-metadata causing errors in post processing

[vutang50]

Hi,
When running the program from a vevo source with --add-metadata parameter, in post processing phase, the following error is encountered:
[ffmpeg] Adding metadata to 'test.mp4'
ERROR: av_interleaved_write_frame(): Operation not permitted

This is true even with avconv installed. I tried to manually run ffmpeg and avconv
ffmpeg -i test.mp4 -codec copy -metadata title="Test Title" test1.mp4
eventually runs into this error:
[mp4 @ 0x2547260] malformated aac bitstream, use -absf aac_adtstoasc
av_interleaved_write_frame(): Operation not permitted

If i then execute ffmpeg -i test.mp4 -codec copy -metadata title="Test Title" -bsf aac_adtstoasc test1.mp4
That succeeds and
$ du -h *.mp4
149M test1.mp4
163M test.mp4

$ ffprobe -loglevel quiet -show_format test1.mp4 | grep title
TAG:title=Test Title

I did not research what would be the reason for the malformed bit stream, but it happens consistently from the same source. Don't know if a good solution is to somehow activate that flag or allowing us to add more parameters in the post processor. Other wise I suppose I could use -exec or a script is to interpret the --write-json-file out and then run the ffmpeg command externally to embed the meta data. I essentially would also like to embed thumbnail but have not tried that.

[jaimeMF]

Please post the full output of youtube-dl --verbose URL OTHER_OPTIONS, otherwise we can't debug the problem.

[jaimeMF]

I can reproduce it with the test video http://www.vevo.com/watch/hurts/somebody-to-die-for/GB1101300280.

Using the ffmpeg downloader instead of the native hls downloader works. Maybe we could use it instead.

[vutang50]

I am not sure if you still need the verbose output, but I have added it below. Is there a parameter to use the ffmpeg downloader or does that need to be added to the command line?

youtube-dl --add-metadata --verbose --write-info-json --write-thumbnail -o test.mp4 http://cache.vevo.com/m/html/embed.html?video=USCJY1431509
[debug] System config: []
[debug] User config: []
[debug] Command-line args: ['--add-metadata', '--verbose', '--write-info-json', '--write-thumbnail', '-o', 'test.mp4', 'http://cache.vevo.com/m/html/embed.html?video=USCJY1431509']
[debug] Encodings: locale UTF-8, fs UTF-8, out UTF-8, pref UTF-8
[debug] youtube-dl version 2014.10.25
[debug] Python version 2.7.6 - Linux-3.13.0-43-generic-x86_64-with-Ubuntu-14.04-trusty
[debug] Proxy map: {}
[Vevo] Retrieving oauth token
[Vevo] USCJY1431509: Downloading JSON metadata
[Vevo] USCJY1431509: Downloading HLS formats
[Vevo] USCJY1431509: Downloading webpage
[info] Writing video description metadata as JSON to: test.info.json
[Vevo] USCJY1431509: Downloading thumbnail ...
[Vevo] USCJY1431509: Writing thumbnail to: test.jpg
[debug] Invoking downloader on u'http://hls-aws.vevo.com/v3/hls/2014/11/USCJY1431509/a50b702a-4eb4-4f08-b6d0-a0070c4627e6/5200/uscjy1431509_5200k_1920x1080_h264_5200_aac_128.m3u8'
[download] Destination: test.mp4
[hlsnative] USCJY1431509: Downloading m3u8 manifest
[hlsnative] USCJY1431509: Downloading segment 1 / 28
[hlsnative] USCJY1431509: Downloading segment 2 / 28
[hlsnative] USCJY1431509: Downloading segment 3 / 28
[hlsnative] USCJY1431509: Downloading segment 4 / 28
[hlsnative] USCJY1431509: Downloading segment 5 / 28
[hlsnative] USCJY1431509: Downloading segment 6 / 28
[hlsnative] USCJY1431509: Downloading segment 7 / 28
[hlsnative] USCJY1431509: Downloading segment 8 / 28
[hlsnative] USCJY1431509: Downloading segment 9 / 28
[hlsnative] USCJY1431509: Downloading segment 10 / 28
[hlsnative] USCJY1431509: Downloading segment 11 / 28
[hlsnative] USCJY1431509: Downloading segment 12 / 28
[hlsnative] USCJY1431509: Downloading segment 13 / 28
[hlsnative] USCJY1431509: Downloading segment 14 / 28
[hlsnative] USCJY1431509: Downloading segment 15 / 28
[hlsnative] USCJY1431509: Downloading segment 16 / 28
[hlsnative] USCJY1431509: Downloading segment 17 / 28
[hlsnative] USCJY1431509: Downloading segment 18 / 28
[hlsnative] USCJY1431509: Downloading segment 19 / 28
[hlsnative] USCJY1431509: Downloading segment 20 / 28
[hlsnative] USCJY1431509: Downloading segment 21 / 28
[hlsnative] USCJY1431509: Downloading segment 22 / 28
[hlsnative] USCJY1431509: Downloading segment 23 / 28
[hlsnative] USCJY1431509: Downloading segment 24 / 28
[hlsnative] USCJY1431509: Downloading segment 25 / 28
[hlsnative] USCJY1431509: Downloading segment 26 / 28
[hlsnative] USCJY1431509: Downloading segment 27 / 28
[hlsnative] USCJY1431509: Downloading segment 28 / 28
[ffmpeg] Adding metadata to 'test.mp4'
[debug] ffmpeg command line: avconv -y -i test.mp4 -c copy -metadata date=20141110 -metadata 'artist=Taylor Swift' -metadata 'title=Blank Space' test.temp.mp4
ERROR: av_interleaved_write_frame(): Operation not permitted
Traceback (most recent call last):
File "/usr/local/bin/youtube-dl/youtube_dl/YoutubeDL.py", line 1114, in post_process
keep_video_wish, new_info = pp.run(info)
File "/usr/local/bin/youtube-dl/youtube_dl/postprocessor/ffmpeg.py", line 481, in run
self.run_ffmpeg(filename, temp_filename, options)
File "/usr/local/bin/youtube-dl/youtube_dl/postprocessor/ffmpeg.py", line 69, in run_ffmpeg
self.run_ffmpeg_multiple_files([path], out_path, opts)
File "/usr/local/bin/youtube-dl/youtube_dl/postprocessor/ffmpeg.py", line 63, in run_ffmpeg_multiple_files
raise FFmpegPostProcessorError(msg)
FFmpegPostProcessorError

[pricejn2]

@vutang50 There is a --prefer-ffmpeg parameter to force the use of ffmpeg instead of the default avconv.

[jaimeMF]

@pricejn2 I got the same error using latest ffmpeg.
Using the native downloader is harcoded in the extractor, we would need to change it.

[pricejn2]

Ah, I see. Thanks for clarifying.