Start a new topic

July 3rd Amara Update

Dear Amara users,

We're happy to announce we made an update to Amara yesterday. Below are a list of features and fixes. Some of these may or may not apply to every team or every team member, if you have questions about these, please reply to this thread. We're looking to have another update for Amara up in the coming weeks, and I'll make sure to update everyone on the forums with what's new.

Best,
Darren

Features
We restored the ability to upload unsynced TXT transcripts.
We added some new languages: 'bug', 'skx', 'prs', and 'haz'.
We reverted the 32-char, 4-line display back to what it was before.
 
Bug Fixes for teams that use Workflows:
"Send back" at approval step now correctly moves task back to review step (it had been pushing it all the way back to transcribe/translate step).
We're now properly allowing resumption of work that was started without a preexisting task.
We fixed a bug that disabled completion of a review task when the reviewer had uploaded a draft.
We're now correctly reassigning approve tasks that result when an approver sends a task back and the review resubmits it.

Fixes for teams that don't use Workflows:
We're no longer mistakenly moderating subtitles in workflow-free teams.

Fixes for all teams:
We're no longer adding redundant title data to language name when syncing subtitles to Youtube.

Yes as you say now we can send txt files but there is a bug when we are using it. Letters "à,è,ê,ç" doesn't works, these are important letters in french (see attached files). Please don't tell me we have to wait an other week for this little bug.
Secondly we still 
can't past a text in a transcription, which is so much easier than to save a document on our computer. 
 
Anyway i'm not the kind of person to complain every time but here I have to say Amara was perfect so far 1 month ago, hope to have a simple software as before.
Hi Darren,

I second Verone's comment about accented - and possibly all not alphanumeric - letters: they are absolutely essential in heaps of languages.

So if Amara can only process them correctly when the .txt file is encoded as UTF-8, which seems to be the case (see the exchange about these letters between Verone and me in Next release of the Amara software: date? testing?) Amara should roll back the copy-paste possibility, where the Amara software did the UTF-8 encoding itself. Anyway, what was wrong with copy-pasting? Less esthetically streamlined to have 2 options from a geek's viewpoint? As subtitlers' though, we definitely prefer something less elegant but that works.

Re: "We're no longer mistakenly moderating subtitles in workflow-free teams." Wahoo and thanks!

However, what about formerly imposed moderation that's still blocking collaborative subtitling in these teams?
In the Music Captioning team alone, there are still  8 sets of subtitles  with a visible task - that cannot be collaboratively edited (you'll find them easily in the Tasks page, Darren, but for non members who can't view that page, see the attached screenshots).
Then there are the subs that are moderation-blocked, but without a visible task, like e.g.
  • French subs for "Damien Lee & Gavin Hatchel - Im Not Who I Was (Cover)",
  • Spanish, Korean, Portuguese, Brazilian and Italian subs for "Coldplay - Paradise"
  • Polish subs for "Carly Rae Jepsen - Call Me Maybe"
  • possibly other ones: I found the above via the Diigo list of blocked subs I started making at your request, Darren, but gave up updating after June 15, as the moderation was continuously developing new collaboration-blocking ways

Now if it were only a matter of editing subs, we could message the members of the team, telling them to remove the still moderation-blocked video from the team and then re-add it if they want to edit the blocked subs.

But moderation-blocked subs are also unavailable in the player for people who need subs: and that defeats the purpose of the Music Captioning team, which is to be "[t]he place where music is captioned to bridge the gap between hearing and deaf world. " And that is also the goal stated in the video of Amara's home page:

We started Universal Subtitles because we believe every video on the web should be subtitle-able. Millions of deaf and hard-of-hearing viewers require subtitles to access video.

(From 0.00 to 0:11; true, that's still the same video that was already there when Amara was called Universal Subtitles, but there has been no announcement indicating that this motivation was dropped with the change of name)

So, can Amara people identify these still moderation-blocked subtitles and unblock them, please?

Hi Claude, and Verone,

Claude-- In the past you had asked that I don't perform fixes to tasks and blocked subtitles for Music Captioning team. I want to respect the fact that you are the admin of that team, and won't mess with anything there unless you give me the go-ahead. I'll wait to hear your final word before making these fixes, if you want me to perform them, I will make sure to backup any subs that might be affected.

Verone-- I'm going to look into the possibility of restoring copy/pasting text into Amara's transcription interface, as well as the issue with the letters "à,è,ê,ç" not displaying correctly. Have you noticed that these letters don't show up at all, or only when uploading .txt files? The more information the better, since this is likely an issue that will need to be passed to our developers.

Best,
Darren
Hi Darren,

Re removing the remaining blocks caused by the imposed moderation
I objected to further Amara fixes  during the imposed moderation after one deleted hours of work I had done on a long PBS NewsHour video, reverting to a former version and wrongly labeling it as well. As it had also happened to another NewsHour contributor, and what with the  workarounds we had to keep thinking up in order to still collaborate in spite of the glitchy moderation, I didn't want further work - mine or other people's - deleted by this "fix".

But if you save the .srt files first, thanks: please go ahead.

re the transcript issue mentioned by Verone: it's also a very important one in the Music Captioning team, where we have videos in many languages: so we need to be able to use transcripts that remain true, with all accented etc letters, once processed by Amara. So it would be great if you could convince the developers to let users copy-paste again: thanks.

Best,

Claude
Hi Daren, this accent problem appear when you are uploading a txt file. (The only way for us to work for the moment)
Dear Darren,

Thank you for having removed the 8 blocks with visible tasks in Music Captioning. I hope you also manage to remove the blocks without visible task that are still affecting 7 subtitle sets.

Best,

Claude
Dear Darren,

Thank you for having removed the 8 blocks with visible tasks in Music Captioning. I hope you also manage to remove the blocks without visible task that are still affecting 7 subtitle sets.
And thanks now for having removed those without visible task too.

Best,

Claude
Login or Signup to post a comment