Amazon announces "Amazon Polly" to convert text to native level natural language, supports 24 languages ​​and 47 voices

ByStephen Poff

Amazon Web Services (AWS) is using "artificial intelligence service"Amazon Polly"At the event" AWS re: Invent 2016 "conducted in Las Vegas. Amazon Polly is a text-to-speech service that supports 24 languages ​​and 47 voices.

Amazon Polly - sentence to voice, 47 voices and 24 languages ​​| Amazon Web Services blog
https://aws.amazon.com/jp/blogs/news/polly-text-to-speech-in-47-voices-and-24-languages/

Amazon Polly - Lifelike Text-to-Speech
https://aws.amazon.com/jp/polly/

Amazon Lex - Build Conversational Voice & Text Interfaces | AWS Blog
https://aws.amazon.com/jp/blogs/aws/amazon-lex-build-conversational-voice-text-interfaces/

Amazon Polly, announced by AWS, is a cloud service that reads spoken text by voice. Amazon Polly uses its own tools and applications to convert text to speech and supports 47 languages ​​including male & female voice and 24 languages ​​including Japanese.


Polly's feature is an accurate recognition function that has been developed with native speakers. For example, "live" in the sentence "I live in Seattle" and "Live from New York" is an isomorphic word having the same spelling but different pronunciation and meaning, Amazon Polly is an isomorphism word It recognizes and reads with different pronunciation.

The Polly console has already been released and any AWS member can try it.

Amazon Polly
https://console.aws.amazon.com/polly

Polly took 5 million monthly for free, above that it costs $ 0.000004 per letter, which is about $ 0.004 per minute of voice. With Polly, there is a possibility that it is easier to develop voice-enabled applications than before.

AWS is another artificial intelligence service, "Amazon LexAlso announced. Amazon Lex is the company's personal assistant "Amazon AlexaIt is a service that can construct a conversational interface like a chat bot using the deep learning technology adopted by.


Amazon Lex supports two types of interactive interface with text and voice. Alexa is installed in the Wi - Fi speaker "Amazon Echo" of voice control, but it is not yet released in Japan. Using Alexa's natural language processing machine, it is possible to develop text format chat bots and speech recognizable applications.

Amazon Lex has released a preview version for the Northern Virginia (North Virginia) region. Price is free for 10,000 text requests and 5000 voice requests for the first year only. After that, there will be a fee of $ 0.75 (about 85 yen) for every 1000 text requests and $ 4 (about 470 yen) for every 1000 voice requests.