Google is launching a new AI speech synthesizer as part of its machine learning cloud tool suite. The cloud-to-speech service will work for any developer or business that needs speech synthesis, whether it's an application, a website, or a virtual assistant. Interestingly, Cloud Text-to-Speech is powered by WaveNet, which was created by Google's AI subsidiary DeepMind in the UK. Since Google acquired DeepMind in 2014, it has been exploring how to turn a company's AI talent into a tangible product. Now, using DeepMind's algorithm reduces the cost of electricity in Google's data centers by 40%, and DeepMind will be in the health care arena. However, integrating WaveNet directly into its cloud services may be even more important, especially as Google tries to win cloud business from Amazon and Microsoft. Second, DeepMind's AI speech synthesis technology is one of the most advanced and realistic technologies in the industry. Most speech synthesizers (including Apple's Siri) use so-called join compositing, where the program stores a single syllable - sounds such as "ba", "sht" and "oo" - and flies them together to form words and sentences. This method has been quite good for many years, but it still sounds awkward. In contrast, WaveNet uses machine learning to generate audio from scratch. It actually analyzes the waveforms in a huge human speech database and recreates them at a rate of 24,000 samples per second. The end result includes subtle sounds such as lip sounds and accents. WaveNet was first integrated into Google Assistant in October last year (although only in Japanese and English) and is now available for selected voices in Cloud Text-To-Speech. According to Google, the new service offers 32 different voices, which can be said to be 12 languages, and users can customize the pitch and speed. CCTV Tool, Network Installation Tool, CCTV Installation Tool, Stripping Tool, Crimping Tool Chinasky Electronics Co., Ltd. , https://www.chinacctvproducts.com
March 17, 2023