How do I set the audio language with --parse-metadata?

[bradenhilton]

I noticed in the section for modifying metadata that language can be set for each stream and was wondering if it was possible/how to set the language of the audio stream. I'm not entirely sure how to go about selecting stream specific metadata.

The language isn't set to what I would expect for YouTube downloads (as in the fields are all empty strings with the preference set to -1 in the info JSON). Perhaps this is an extractor bug that might need an issue opening, but I'd still like to know if this is possible regardless as a workaround. I could be misunderstanding how it works though.

Do I need to use ffmpeg style mappings (s:​a​:0 etc.) or is some other mapping used?

I'd like to do this as my media player can auto select subtitles/captions based on the language of the current audio track, but they all get marked as English so I have to select them manually (I am a native English speaker and so generally do not need English subtitles/captions for English audio).

[pukkandan]

was wondering if it was possible/how to set the language of the audio stream. I'm not entirely sure how to go about selecting stream specific metadata.

No, stream-specific metadata cannot currently be manually set using --parse-metadata. It should be possible to implement though. Open a feature request

Do I need to use ffmpeg style mappings (s:​a​:0 etc.) or is some other mapping used?

You can pass appropriate ffmpeg args using --ppa Merger:<args>. But not sure if that is an ideal solution

The language isn't set to what I would expect for YouTube downloads (as in the fields are all empty strings with the preference set to -1 in the info JSON)

Could you give example URL?

[bradenhilton]

No, stream-specific metadata cannot currently be manually set using --parse-metadata. It should be possible to implement though. Open a feature request

Will do!

You can pass appropriate ffmpeg args using --ppa Merger:<args>. But not sure if that is an ideal solution

I'll give it a try, but I think I would prefer doing it via --parse-metadata if at all possible (if/when it is implemented).

Could you give example URL?

I tried 2 different URLs that I would expect to have wildly different languages and both of them are as I described above when looking at the info JSON:

https://www.youtube.com/watch?v=PEKkdIT8JPM -> I'd expect something like this to have Korean audio (kor/ko/ko-KR).
https://www.youtube.com/watch?v=HeKuX063bcs -> I'd expect this to have English audio, but the JSON has blank languages and the file seems to just default to English when I inspect with mediainfo.

I'm happy to move this to an issue though if you'd prefer, as you have actually answered the question.

[pukkandan]

I see what u mean now. But youtube doesn't give any language information for those videos. I think the uploader has to set it (not sure)

Here's an example with language data available

$ yt-dlp -F WaOKSUlf4TM
[youtube] WaOKSUlf4TM: Downloading webpage
[youtube] WaOKSUlf4TM: Downloading android player API JSON
[youtube] WaOKSUlf4TM: Downloading android agegate player API JSON
[youtube] WaOKSUlf4TM: Downloading web agegate player API JSON
[info] Available formats for WaOKSUlf4TM:
ID    EXT  RESOLUTION FPS |  FILESIZE    TBR  PROTO | VCODEC          VBR  ACODEC     ABR  ASR    MORE INFO
----- ---- ---------- --- - ---------- ------ ----- - ------------- ------ --------- ---- ------- ---------------------------------------------------
139-0 m4a  audio only     |  885.42KiB   48k  https |                      mp4a.40.5  48k 22050Hz [en], English descriptive, low, m4a_dash
139-1 m4a  audio only     |  885.42KiB   48k  https |                      mp4a.40.5  48k 22050Hz [en], English original (default), low, m4a_dash
249-0 webm audio only     |  889.24KiB   49k  https |                      opus       49k 48000Hz [en], English descriptive, low, webm_dash
250-0 webm audio only     |  1.14MiB     64k  https |                      opus       64k 48000Hz [en], English descriptive, low, webm_dash
249-1 webm audio only     |  877.56KiB   48k  https |                      opus       48k 48000Hz [en], English original (default), low, webm_dash
250-1 webm audio only     |  1.13MiB     64k  https |                      opus       64k 48000Hz [en], English original (default), low, webm_dash
140-0 m4a  audio only     |  2.29MiB    129k  https |                      mp4a.40.2 129k 44100Hz [en], English descriptive, medium, m4a_dash
140-1 m4a  audio only     |  2.29MiB    129k  https |                      mp4a.40.2 129k 44100Hz [en], English original (default), medium, m4a_dash
251-0 webm audio only     |  2.27MiB    128k  https |                      opus      128k 48000Hz [en], English descriptive, medium, webm_dash
251-1 webm audio only     |  2.26MiB    127k  https |                      opus      127k 48000Hz [en], English original (default), medium, webm_dash
17    3gp  176x144    8   |  1.39MiB     78k  https | mp4v.20.3       78k  mp4a.40.2   0k 22050Hz 144p
394   mp4  256x144    30  |  1.14MiB     64k  https | av01.0.00M.08   64k                         144p, mp4_dash
160   mp4  256x144    30  |  1.01MiB     56k  https | avc1.4d400c     56k                         144p, mp4_dash
278   webm 256x144    30  |  1.38MiB     78k  https | vp9             78k                         144p, webm_dash
395   mp4  426x240    30  |  2.18MiB    123k  https | av01.0.00M.08  123k                         240p, mp4_dash
133   mp4  426x240    30  |  2.18MiB    123k  https | avc1.4d4015    123k                         240p, mp4_dash
242   webm 426x240    30  |  2.58MiB    145k  https | vp9            145k                         240p, webm_dash
396   mp4  640x360    30  |  4.43MiB    250k  https | av01.0.01M.08  250k                         360p, mp4_dash
134   mp4  640x360    30  |  4.29MiB    242k  https | avc1.4d401e    242k                         360p, mp4_dash
18    mp4  640x360    30  |  10.32MiB   582k  https | avc1.42001E    582k  mp4a.40.2   0k 44100Hz 360p
243   webm 640x360    30  |  5.54MiB    312k  https | vp9            312k                         360p, webm_dash
397   mp4  854x480    30  |  8.13MiB    459k  https | av01.0.04M.08  459k                         480p, mp4_dash
135   mp4  854x480    30  |  7.11MiB    401k  https | avc1.4d401f    401k                         480p, mp4_dash
244   webm 854x480    30  |  9.96MiB    563k  https | vp9            563k                         480p, webm_dash
398   mp4  1280x720   30  |  16.62MiB   939k  https | av01.0.05M.08  939k                         720p, mp4_dash
136   mp4  1280x720   30  |  11.98MiB   677k  https | avc1.4d401f    677k                         720p, mp4_dash
22    mp4  1280x720   30  |            2287k  https | avc1.64001F   2287k  mp4a.40.2   0k 44100Hz 720p
247   webm 1280x720   30  |  19.67MiB  1111k  https | vp9           1111k                         720p, webm_dash
399   mp4  1920x1080  30  |  30.05MiB  1698k  https | av01.0.08M.08 1698k                         1080p, mp4_dash
137   mp4  1920x1080  30  |  39.80MiB  2249k  https | avc1.640028   2249k                         1080p, mp4_dash
248   webm 1920x1080  30  |  35.69MiB  2016k  https | vp9           2016k                         1080p, webm_dash
400   mp4  2560x1440  30  |  86.39MiB  4882k  https | av01.0.12M.08 4882k                         1440p, mp4_dash
271   webm 2560x1440  30  |  97.87MiB  5531k  https | vp9           5531k                         1440p, webm_dash
401   mp4  3840x2160  30  |  176.14MiB 9954k  https | av01.0.12M.08 9954k                         2160p, mp4_dash
313   webm 3840x2160  30  |  240.13MiB 13570k https | vp9           13570k                        2160p, webm_dash

I will see if the language data is available somewhere in the API, but don't get your hopes up

[bradenhilton]

I had suspected something like this might be the case, hence me wanting to just force an arbitrary value into the tag manually.

I'll mark answered for now and open the feature request soon.

Thanks for your timely responses and assistance.

[bradenhilton]

Just a quick follow up.

You can pass appropriate ffmpeg args using --ppa Merger:<args>. But not sure if that is an ideal solution

--ppa 'Merger+ffmpeg:-metadata:s:a:0 language=LANG' as well as --ppa 'Metadata+ffmpeg:... etc. do indeed accomplish this.