This archive report was first published on 14 August 2019.
On August 14, 2019, Chinese search engine Sogou made headlines at the China Online Literature+ conference by unveiling 'lifelike' avatars of Chinese authors Yue Guan and Bu Xin Tian Shang Diao Xian Bing. These avatars were created from video recordings and are designed to read popular novels in the authors' voices.
The development is part of a growing trend in China's audiobook market, which is expected to more than double between 2016 and 2020, reaching 7.8bn Chinese yuan (£900m) a year, according to think tank iiMedia.
Appetite for audiobooks in China is on the rise, mirroring trends in the West. While text-to-speech technology can quickly generate an audio version of a book, most people prefer audiobooks that are 'professionally narrated' by authors, actors, or famous public figures.
Advances in machine learning and speech-to-text technologies are making digitized voices more lifelike. For example, Lyrebird allows clients to create custom 'vocal avatars' from just a one-minute recording of their voices. The company showcases its services with two audio recordings of former US President Barack Obama and current US President Donald Trump, created using machine learning.
Meanwhile, London-based tech start-up DeepZen says its synthetic speech technology can create multiple versions of an audiobook within hours, cutting down time and costs by an estimated 90%.
