IBM has developed a computerised voice that is almost indistinguishable from a human.

http://www.telegraph.co.uk/scienceandtechnology/technology/technologynews/4420798/IBM-develop-most-realistic-computerised-voice.html

by vaksel • 17 years ago

27 25 17 years ago

telegraph.co.uk

nobbynutz2• 17 years ago

Does anyone have a link for any further details?

quoderat• 17 years ago

Here you go:

http://www.patents.com/Generating-paralinguistic-phenomena-v...

But that said, it mostly just sounds like a way to insert a bunch of annoying "umms" and "ahs" into artificial conversation.

lunchbox2• 17 years ago

So it looks like this system has people use simple textual markup. From the patent:

"For example, the developer may specify: <prosody style="bad news">Well, \sigh I cannot answer that question;</prosody>"

sosuke• 17 years ago

Perfect candidate for EmotionML? http://www.w3.org/2005/Incubator/emotion/XGR-emotionml/

ars• 17 years ago

Um, I don't, like, want that, er, in my, you know, like, conversations with robots.

wallflower5• 17 years ago

Acapela Group's voices are already pretty good. I bought Lucy, the British one. She sounds almost like a voice-response automated phone system.

http://www.acapela-group.com/text-to-speech-interactive-demo...

nobbynutz2• 17 years ago

if you don't mind me asking how much was "Lucy" roughly? $100's, $1,000's, $10,000's, more? I couldn't find a price list on the site.

lunchbox2• 17 years ago

Lucy costs $35 USD, plus the cost of the text-to-speech program (e.g. TextAloud, which I think is <$50): http://www.nextup.com/acapela.html

(Acapela doesn't sell directly to consumers; they license their technology to resellers.)

markessien4• 17 years ago

You use this in an API or just personally? Is it possible to use these voices over an API (for non-commercial use)?

wallflower5• 17 years ago

I don't use their API. I just use the TextAloud program to occasionally read websites and documentation to me. (Don't do this at home folks - let's you surf the web without 'reading' it - it obviously is not at the 'generate your own audiobook level' [1] - however it's way beyond E.T. and that speech synthesizer)

API product page (iPhone version available now too!): http://www.acapela-group.com/acapela-multimedia-8-speech-sol...

[1] That would require bridging the gap from 'Dry eyes. dry eyes' monotone to an actor doing a dramatic script reading. Vocal variety, appropriate pauses, inflections, tone - wow - a lot of things. Possible startup opportunity here?

As an aside, it takes much longer to listen than to read (I'm thinking 10x for some material).

furyg3• 17 years ago

Wow. They even have tons of foreign languages. This would have been extremely useful when I was learning how to pronounce words in Dutch (still might be ;) ).

wallflower5• 17 years ago

I'm so close to buying the Spanish voice to practice my Spanish listening skills.

mhb• 17 years ago

No place to, uh, hear it?

abdulhaq• 17 years ago

I'm pretty sure I've spoken to a number of these computerised voices at some customer service call centres.

randomtask• 17 years ago

Here's a page with some sample audio clips. There are some papers that are some credited to the guy mentioned in the article on the publications page

http://www.research.ibm.com/tts/

TrevorJ• 17 years ago

Sounds like they still have some trouble with phoneme cadence and blending.

TooMuchNick• 17 years ago

IBM's new business plan: Run the works of the Clash through their software, crank it through Songsmith and make millions on the YouTube adsense.

IBM has developed a computerised voice that is almost indistinguishable from a human.

Related Stories