Google taught AI to say “more clearly and more natural”

Google will update the engine to convert the text into Speech Services on Android devices, making the sound of the voices more pure and natural.

According to developers, users will not have to do anything – the modernization of the technology will “occur behind the scenes”. The update will “significantly improve” the quality of generated speech, especially with regard to clarity and naturalness, they added.

421 Voice in 67 languages ​​will receive a new speech model and synthesizer. The current sound of American English will automatically change to a speech created using “more fresh data”.

The developers also showed samples of updated votes in other languages.

The company will distribute an update for Speech Services to all 64-bit Android devices through the Google Play app store over the next few weeks.

Recall that in September Openai introduced a system of speech recognition with an open source code Whisper, which provides transcription in several languages.

In May 2021, Google introduced the Lamda neural network model, which communicates as a living person and supports a relaxed dialogue on different topics.

Subscribe to FORKLOG news at Telegram: Forklog AI – all news from the world of AI!