Nuance – We Want To Talk To You

April 1, 2012 § 1 Comment

Nuance – what a great name for a company. We’ve been reading about you for years, and in the news, this article now:

We encourage these projects and especially if they also lead to much more QUALITY speech-to-text systems – full VERBATIM and also REAL TIME  – for so many millions of us who are deaf, deafened, have hearing loss, or use captioning for everyday communications for many other good reasons.

Being able to train a system to one’s own voice, and then talk to so many different machines may be fun, useful too!

At the same time, we urgently need accurate speech to text systems – we need improvements as soon as your company, Google, and others can provide.

We doubt that these systems will replace human professional caption providers anytime soon, and yet…

Speaking for one person only here – half of me wants you to try a whole lot harder. And the other half of me knows how important real time professionals are for so many of us now, in so many everyday situations, and will remain vital also. Captioning is our language too – we want it universally, so there should be room for all professionals and all systems too.

Nuance – tell us how we can help. We hope to talk with you Nuance, in Burlington, MA, USA one day soon. We would like to learn a lot more, and tell you about the CCAC also.



Tagged: , , , ,

§ One Response to Nuance – We Want To Talk To You

  • Among the points I wish the NYT article had explained:

    Does Nuance work at all in the direction of non-personalized speech-to-text, i.e. working with any voice? Because that is of essence for an evolution towards automatic captioning.

    The other day, a friend and I were trying Google’s spech-to-text as presently offered on my android tablet. We were dictating a text, taking turns. There was no significant difference whether she or I was dictating. Well, just one case where her saying “pompier” (French for fireman) came out as “font chier” (vulg. French for “are annoying”), whereas the software wrote “pompier” correctly when I said it. But then, she also paused a trifle longer between the syllables than I did, so Google speech-to-text seems “voice agnostic” and works darned well for dictation already (that’s why Apple making such a doo-dah about their Nuance-powered Siri was a bit ludicrous).

    Of course, speech-to-text needs to work with normal speech too, not just with careful dictation. There it’ll be a matter of the variety of spoken texts samples the software can refer to. And possibly, Google, who can use all the speech parts in all YT videos might have a solid advantage.

What’s this?

You are currently reading Nuance – We Want To Talk To You at CCAC Blog.


%d bloggers like this: