Real time spoken language translator

I've seen people being overly optimistic about this. From what I've read it's nothing new. All they've done is tied a bluetooth headset to an app that most likely takes STT > Translate > TTS.

Not sure on what speech to text engine they're using yet (would have to be a fair jump from modern tech) nor do I know the translation engine but it seems all but confirmed it's using googles.

The only way I'll buy this is if they call it Babel :D.
 
I can see this ending badly given how awful some speech recognition is. "Excuse me, what is your name?" Becomes "**** on me, I think you are gay?" Etc... :p
 
I can see this ending badly given how awful some speech recognition is. "Excuse me, what is your name?" Becomes "**** on me, I think you are gay?" Etc... :p

I tried it for the first time in years the other day, was driving and wanted to reply to a message to say I was running late - so tried the voice to text thing - it worked perfectly. Was just a default ios keyboard option, no training. I had internet so it could well have sent it to California to apply several gigaflops to it, but I got the right result so who cares :)
 
I tried it for the first time in years the other day, was driving and wanted to reply to a message to say I was running late - so tried the voice to text thing - it worked perfectly. Was just a default ios keyboard option, no training. I had internet so it could well have sent it to California to apply several gigaflops to it, but I got the right result so who cares :)

+1 My google speech recognition rarely gets anything wrong these days. Except welsh place names, then I might as well be speaking Chinese...
 
I tried it for the first time in years the other day, was driving and wanted to reply to a message to say I was running late - so tried the voice to text thing - it worked perfectly. Was just a default ios keyboard option, no training. I had internet so it could well have sent it to California to apply several gigaflops to it, but I got the right result so who cares :)

I find it works perfectly when you speak clearly, I've sent a long email using voice recognition when driving before, and I looked at it when I arrived at work and it came out fine. I was a bit worried that it might not have done and I might have inadvertently sent something daft!

The second you speak a little too quickly, or if there is some background noise etc, it all gets a bit temperamental. So I'm not sure how practical it will be in social situations, public places, etc.

I'd prefer to make sure it has actually detected what I've said first, but if I did that it would no longer be "real time".

Interesting one though, if implemented properly. And of course these things are forever improving!
 
Back
Top Bottom