Amazon Web Services on Wednesday announced the addition of Hindi language support for Amazon Polly, a machine learning service that turns text into lifelike speech, allowing users to create applications that talk and build new categories of speech-enabled products. With the addition of Hindi language support, the company now has its first bilingual voice for Amazon Polly. Aditi, first released as an Indian English voice to the Amazon Polly portfolio in November 2017, now also speaks fluent Hindi.
Aditi and Raveena are the two Indian English voices currently available on Amazon Polly, along with a variety of other voices in multiple languages to synthesise speech from text. According to the 2011 Census of India, Hindi is spoken by over 500 million people worldwide. The addition of Hindi language support for Amazon Polly will enhance the experience for Indian customers.
Organisations are using the new bilingual Indian English/Hindi voice to enhance automated customer engagement applications, interactive voice responses (IVR), audio news, and targeted vernacular language services.
“Today, there is a demand for content in Hindi and Indian English across all genres of voice applications — from training videos to e-learning, corporate narrations and voice bots. Amazon Polly makes it easy to switch between languages and voices, based on the customer’s requirements. We introduced the bilingual support in response to requests from customers for high-quality speech synthesis in both Hindi and Indian English, especially in scenarios where the two languages are mixed together and spoken. Customers can now provide mixed input in a variety of dialects including Devanagari Hindi, Romanised Hindi, and English, as well as complete Hindi text input in Devanagari script,” said Navdeep Manaktala, head of business development at Amazon Internet Services Private Limited.
One of India’s largest insurance marketplaces, PolicyBazaar.com, is using Amazon Polly’s bilingual support for their in-house interactive voice response (IVR) calling service and is looking to innovate further with the new release.
Try deep learning using MATLAB