Videos might be in more than one language

[conferences-gesticulees]

• PeerTube version or commit: 1.0.0-alpha.8

• What is the expected behaviour?

A video might be in multiple languages. The movie Le Jeune Karl Marx is such an example: actors in the original vesrsion express themselves in German, English, and French.

A video might contain subtitles in a language that is different from the language spoken by the characters.

Maybe it would be worthwhile to have a language field per audio track, and one language field for the video. These language field should allow to select more than one language. The language field for the video could also allow to select sign languages.

• What do you see instead?

When editing metadata for a video, the dropdown only allows to select one language.

[rigelk]

#420 should at least add support for sign languages until we engage in the bigger refactoring of the multiple language support.

[NoMoreCRAPTion]

#420 should at least add support for sign languages until we engage in the bigger refactoring of the multiple language support.

Glottolog is the best documented database for this. It includes ISO-639 along with peer-reviewed papers on the existence of these languages and variations.

It's probably just easier to auto-update / recompile (every time PeerTube rolls out a new release) based on maintained tables of ISO-639 along with cross-references to other attributions like Glottocode so one doesn't need to manually add language support.

Think Ethnologue is the only one which maintains active numbers based on government surveys and other sources though (including sign languages), but access is paywalled. Doesn't stop web archivers from bypassing the paywall though.

It's probably better to use ISO-639 and Glottocode since a lot of language experts use YouTube to resurrect or revitalize small languages. They just use clunky workarounds like renaming English subtitle tracks.

As seen in this example here:

YouTube's method of renaming subtitle track is hidden and most uploaders don't even know you can rename the subtitle tracks.

I hope the oddity of the above use case helps differentiate PeerTube from YouTube. Especially since PeerTube's main selling point is to decentralize, and language extinctions are driven by the state trying to centralize everything for convenience and subject marginalized people to nationalism and capitalism or forcing small language speakers to prescribe to state-approved political ideologies and corporate interests.

Here is the downloadable files for Glottolog. As you can see, they match the ISO numbers with the glottocodes alongside the names of the language. And Glottolog has public repository on Github. The data is released under Creative Commons Attribution 4.0 International License. And the best part is that dataset is collaborative.

[xeruf]

If I am understanding that correctly, this is about multiple tracks in parallel?
So in the above example, if I don't understand French, I could watch the movie in German and English, with the French translated into one of these two?

Related to #939 anyways

[J-J-Chiarella]

Late to the party (woot, woot!), but in regards to what xeruf said, there are two phenomena:

(1) A movie with multiple audio tracks. Animated movies may have multiple dubs as these can copy everything except vocals and cleanly do dubs.

(2) A movie with multiple languages of dialog. If I watch a movie about East Asia in the 1930s, I will hear Mandarin Chinese, Korean, and Japanese. There will be subtitles for Korean speakers that will only appear and translate Chinese speech and Japanese speech. There will be subtitles for Chinese speakers that will only appear to translate Korean speech and Japanese speech. (This ignores subtitles for hearing learners who would appreciate subtitles even if they speak the language somewhat, and it ignores subtitles for the deaf. That's East Asia for you . . .)