Subtitles not embedded as expected

[forthrin]

youtube-dl -c -f hls-2364 --all-subs --embed-subs https://tv.nrk.no/serie/tara-og-vegard/MYNT16000115/sesong-1/episode-1

Subtitles are not embedded into the .mp4 file as expected. The following vague error message is given, though no obvious errors can be seen when examining the .ttml file.

ERROR: file:Bestegirlza-MYNT16000115.no.ttml: Invalid data found when processing input

[dstftw]

Post the full output of youtube-dl when run with --verbose option.

[yan12125]

ffmpeg does not support embedding TTML subtitles, so you need to convert them to srt or other ffmpeg-compatible formats first. However, TTML conversion in youtube-dl is broken for this case:

$ youtube-dl -v -f hls-2364 --write-sub --sub-format ttml --convert-subtitles srt --embed-subs https://tv.nrk.no/serie/tara-og-vegard/MYNT16000115/sesong-1/episode-1
[debug] System config: []
[debug] User config: []
[debug] Command-line args: ['-v', '-f', 'hls-2364', '--write-sub', '--sub-format', 'ttml', '--convert-subtitles', 'srt', '--embed-subs', 'https://tv.nrk.no/serie/tara-og-vegard/MYNT16000115/sesong-1/episode-1']
[debug] Encodings: locale UTF-8, fs utf-8, out UTF-8, pref UTF-8
[debug] youtube-dl version 2015.12.18
[debug] Git HEAD: 016dd82
[debug] Python version 3.5.1 - Linux-4.3.2-1-ARCH-x86_64-with-arch-Arch-Linux
[debug] exe versions: ffmpeg 2.8.3, ffprobe 2.8.3, rtmpdump 2.4
[debug] Proxy map: {}
[NRKTV] MYNT16000115: Downloading webpage
[NRKTV] MYNT16000115: Downloading f4m manifest
[NRKTV] MYNT16000115: Downloading m3u8 information
[info] Writing video subtitles to: Bestegirlza-MYNT16000115.no.ttml
[debug] Invoking downloader on 'http://nordond7b-f.akamaihd.net/i/wo/open/7d/7d4b00863c1c5230af1f364d92a9ed15db68b65a/dcbe7d41-010f-42c3-956e-94656a3141f2_,141,316,563,1266,2250,.mp4.csmil/index_4_av.m3u8?null='
[download] Bestegirlza-MYNT16000115.mp4 has already been downloaded
[download] 100% of 103.45MiB
[ffmpeg] Converting subtitles
WARNING: You have requested to convert dfxp (TTML) subtitles into another format, which results in style information loss
Traceback (most recent call last):
  File "<string>", line 27, in <module>
  File "/home/yen/Executables/Multimedia/youtube-dl/youtube_dl/__init__.py", line 410, in main
    _real_main(argv)
  File "/home/yen/Executables/Multimedia/youtube-dl/youtube_dl/__init__.py", line 400, in _real_main
    retcode = ydl.download(all_urls)
  File "/home/yen/Executables/Multimedia/youtube-dl/youtube_dl/YoutubeDL.py", line 1675, in download
    url, force_generic_extractor=self.params.get('force_generic_extractor', False))
  File "/home/yen/Executables/Multimedia/youtube-dl/youtube_dl/YoutubeDL.py", line 674, in extract_info
    return self.process_ie_result(ie_result, download, extra_info)
  File "/home/yen/Executables/Multimedia/youtube-dl/youtube_dl/YoutubeDL.py", line 720, in process_ie_result
    return self.process_video_result(ie_result, download=download)
  File "/home/yen/Executables/Multimedia/youtube-dl/youtube_dl/YoutubeDL.py", line 1345, in process_video_result
    self.process_info(new_info)
  File "/home/yen/Executables/Multimedia/youtube-dl/youtube_dl/YoutubeDL.py", line 1657, in process_info
    self.post_process(filename, info_dict)
  File "/home/yen/Executables/Multimedia/youtube-dl/youtube_dl/YoutubeDL.py", line 1721, in post_process
    files_to_delete, info = pp.run(info)
  File "/home/yen/Executables/Multimedia/youtube-dl/youtube_dl/postprocessor/ffmpeg.py", line 500, in run
    srt_data = dfxp2srt(f.read())
  File "/home/yen/Executables/Multimedia/youtube-dl/youtube_dl/utils.py", line 2026, in dfxp2srt
    end_time = begin_time + parse_dfxp_time_expr(para.attrib['dur'])
TypeError: unsupported operand type(s) for +: 'float' and 'NoneType'

There are two lines with invalid timestamps in the downloaded file:

      <p begin="00:00:46.080" dur="00:00:-46.-380" style="left">Dagene bestod av antidepressiva,<br />skolen, hat mot meg selv og andre.</p>
      <p begin="655:35:00.-296" dur="00:00:00.000" style="left">Jeg kranglet med foreldrene mine<br />og låste meg inne på rommet.</p>

youtube-dl should tolerate it.

[yan12125]

Invalid timestamps in TTML are ignored now. In the next version you can embed subtitles with the following command:

$ youtube-dl -v -f hls-2364 --write-sub --sub-format ttml --convert-subtitles srt --embed-subs https://tv.nrk.no/serie/tara-og-vegard/MYNT16000115/sesong-1/episode-1

[PacoH]

Thanks! I had the same problem. --convert-subtitles srt worked without --sub-format ttml. So did --convert-subs, which is from the man page.