Re: Git Inclusion Summit

[Kaartic Sivaraam <kaartic.sivaraam@gmail.com>]

On 28-07-2020 21:55, Taylor Blau wrote:
> On Tue, Jul 28, 2020 at 10:07:26AM +0000, Eric Wong wrote:
>>
>> Jitsi w/ audio-only certainly seems to be a step in the right
>> direction and would be more inclusive.
>>
>> Is there any speech-to-text transcription done for the hearing
>> (or extremely bandwidth) impaired?
> 
> It looks like such a thing exists:
> https://jitsi.org/blog/a-speech-to-text-prototype/.
> 

Yeah. A link with current information appears to be:

https://github.com/jitsi/jigasi#using-jigasi-to-transcribe-a-jitsi-meet-conference

>> It'd ideally go to #git on IRC (or something that doesn't
>> require a browser to trigger swap storms on old systems).
>>
>> Even for people with good hearing, acceptable audio quality for
>> speech seems tricky to get right, being dependent on mics,
>> bandwidth, codecs, background noise, speaker/earphone quality,
>> etc.
> 
> I haven't look hard enough to see if it supports redirecting its output
> to an IRC channel, but my guess is that it probably doesn't.

I believe you're right. From the link above:

> Currently Jigasi can send speech-to-text results to the chat of a
> Jitsi Meet room as either plain text or JSON. If it's send in JSON,
> Jitsi Meet will provide subtitles in the left corner of the video,
> while plain text will just be posted in the chat.

> In either
> case, hopefully disabling audio and video is possible within Jitsi's web
> UI, and you should be able to read or write in the chat as well as read
> the transcription.
> 

I think there's just one catch. The transcription as of now appears to
use Google Cloud speech-to-text API. Well, its Google. I'll let you make
your own inferences. In any case, it hopefully wouldn't be a concern for
people who _read_ the transcription.

-- 
Sivaraam