The popularity of smart assistants has gradually increased the demand for text-to-speech. In recent years, Google has introduced cloud-based artificial intelligence machine learning services. It recently announced the launch of a new text-to-speech API that can turn text into natural pronunciation and support various voices. Application requirements.
The Text-to-speech API supports 12 languages and converts them into 32 natural languages. The computer's written content is converted into spoken language. This API also matches the new WaveNet pronunciation model. , make the pronunciation more natural and accurate, the operation speed is 1000 times faster than the original, it takes only 50 milliseconds to calculate the 1 second voice message, and the resolution of the pronunciation sample is also raised from 8 bits to 16 bits. Google said that in the English WaveNet test, more than 20% of people think that computer pronunciation is better than real-time pronunciation.
Different Internet of Things or smart assistants now require text-to-speech services. Google's own services such as maps, search, etc. have built-in text-to-speech services. Now that APIs are introduced, third-party Internet of Things applications such as TV , cars, etc. can benefit from this technology, so that the interaction between the computer and the user can be more natural and smooth.
