Start a new topic

Source language issues

Yesterday, I created the Ishmael Beah : World Humanitarian Day 2012 Amara page from the homonymous http://www.youtube.com/watch?v=T1XxGRPY1kA YouTube page (1). The YouTube page does not indicate a source language.

  • If I try to use the Create subtitles now link below the player, the dialog box says: "This video is in Gujarati" whereas it is in English, and gives me no option to change that.
  • If I try to use the upload link below the player, the Translated from drop list for selecting a source language in the dialog box is blank and unusable.
See the two attached screenshot

So how can one either create subtitles with the Amara software, or upload them, indicating the correct source language, i.e. English?

(1) I also added the Amara video page to the World Humanitarian Day Amara team: though this doesn't seem prima facie relevant to the issues, I'm giving this info because you never know what developers may change in  Amara teams that will affect subtitling

Hello Claude, 
It seems like an odd situation what happened with your video, were you able to resolve it by re-uploading the video, or did it sort itself out? Checking up on the video it seems that this is no longer an issue. Thanks for letting us know this happened to you. 

Best,

Jules
Hi Jules,

What happened is that Berley started the English subs: see the history of revisions for them. 

I was reluctant to ignore the Gujarati indication because when there is a droplist to choose the original language from, there's a warning to check that this indication is correct as it cannot be changed. But once Berley had ignored it, I completed the subs.

However, if you try to start a new translation, the dialog box still says "This video is in Gujarati": see attached Beah_subtitle_Gujarati_source_language2.jpg screenshot.

Actually, I'm puzzled by the developer's insistence on original language when their drop list cannot cover all instances. E.g. there's a project for subtitling in English the videos of the Irish Deaf Archive YT channel: most are in Irish Sign Language, which is not in the drop list. So what should we put?
Then when I subtitled Revoir le concert : En Grèce, la revanche des rêves, avec Angélique Ionatos, which is half in French, half in Greek, actually perhaps more in Greek than in French, I tossed a coin and put French as original language. And for Sir Tim and Gordon Brown: How Can the Web Accelerate Social and Economic Change? (Geneva University recording), I plunked for English in spite of the long intros in French.

No doubt developers have their reasons for that insistence. But then they should offer an "other" item we can fill ourselves in that drop list for these kinds of cases.
Hi Claude,

I have changed the Ishmael Beah video so it now accurately shows English as the original language.  You are right that there are reasons why the original language has been established and is important that may not be apparent to those not as involved with the backstage and coding side: however devs are aware of the limitations of it and in the new data model many of these things will be improved and resolved. 

For now, the language chosen as original should be the language mostly used in the video even if there are multiple languages. What you've been doing, selecting one which you'd think is the main one is the recommended option.

Thanks for bringing to our notice the Irish Sign Language issue, I've added the request to include this language so that it can be selected as the original language going forth, I hope this helps!

Best regards,

Jules



:-) @ "the language chosen as original should be the language mostly used in the video even if there are multiple languages": when I was a secondary school teacher in Ticino (Switzerland), the Dept of education had these census surveys with one question about "primary language" every year. Regularly, one kid in my class would say "Hey Teach, I don't know what my primary language is, I speak Italian in school and X [Albanese, French, German...] at home."  And I'd reply: "The issue has been raised and I'll raise it again. Meanwhile, write one language this year and the other one next year".

Eventually, the folks who prepared the survey relented and renamed the entry "primary languages" - could that work with the Amara database, maybe?

Then re sign languages: there may not be quite as many as spoken languages, but there are many nonetheless.

So can't we have an "other" item that we can fill ourselves?
Claude, I find myself answering that "one year one answer and one year another answer" whenever someone asks me where I'm from! 

The other field sounds like a good idea from the users perspective. However, (and this I post as my non-technical self) I can imagine it is much more complicated to program since different languages are completely different from one another. I wonder how the system would know if "other" means a RTL language, if it has special characters, what alphabet should be displayed, and how other users would see it. 

In the next few months we'll be seeing a new (and improved) Amara platform, perhaps then we'll have solutions for these and other issues. In the meantime, we are adding any new languages users request, so if you can think of other languages we might add to the list to make work easier for translators, please let us know!
Hi Jules,

Thanks for your hypotheses which seem globally justified. Yet there are still some details that it would be nice if the developers came here to directly explain. e.g.:

If it is a matter of encoding the right sets of characters, as you suggest,
  1. Doesn't the basic UTF-8 encoding used by Amara work for any language?
  2. Why should the original language of the video matter?
  3. Taking again the case of the "French subs" I made indicating "French" as the original language of Revoir le concert : En Grèce, la revanche des rêves, avec Angélique Ionatos, what would have changed if I had used "Greek" for the subs and for the original language?
  4. Isn't the important factor what languages the video is getting subtitled into?
  5. As subs cannot be written in sign language, then how does the original language indication make encoding sense when a video is in a signed language?
Could it not rather be that developers insist so much on original language because they've made that a search filter for finding videos?
  1. If so, what sense does such a filter have when it forces users to indicate only one language when a video is actually bi- or multi-lingual?
  2. Would it not make more sense to have a multiple choice list - with that "other" item subtitlers could fill themselves?
Why do some video pages have a blank, unusable source language drop-list when you create the first subtitles?

Why can't users change the source language indication themselves in case of mistakes (see (1))?

 

***

 

But in practice, thanks for the proposal to add more languages to the source language drop list for the time being. Could we please have:

  • Irish Sign Language
  • Israeli Sign Language
  • Italian Sign Language
  • French Sign Language
  • British Sign Language

to start with?

 

And re Irish Sign Language, if it gets added, could the indicated source language be changed from American Sign Language to Irish Sign Language in:

Or would that be too complicated?

 

(1) Then there is http://www.universalsubtitles.org/en/videos/J5tFpZpIWfOw/info/irish-deaf-archives-john-mccandless/ which says Afar for the original language: possibly the result of a slip of the mouse

I now have the same problem that many had one year ago. The video 'Sign language for a good education' chose Vietnamese as its original language. Original language matters because the name of the video will appear in this language and in this case, when Vietnamese is a language that not many speak, the video will not likely to be picked to work on by others. 

Good point, An - I mentioned this change of complete URL in the ticket I created about your other mention, in the Remove Subtitles or Videos topic, of this issue with 'Sign language for a good education' . I wish I remembered how the full URLs worked before, when there was no video language setting and Amara was called Universal Subtitles. It worked without it, though.
Login or Signup to post a comment