The Xinhua News Agency replaced the human speaker with a synthesized voice CGI model

The future is gradually approaching, the technologies penetrate deeper and deeper into all spheres of our life. One of the manifestations is 3D technology. Recently it became known that the Chinese news agency Xinhua decided to use a virtual announcer for the news program.

So far it is unclear which technological innovations were used to create the model, but the virtual announcer looks quite realistic. Yes, it reads the news a little monotonously and sometimes mimics, but everything is within the normal range.


The agency used the movements and facial expressions of a real person for its model - they do this quite often. The “prototype” was a real employee of the company, Zhang Zhao. The voice of the speaker is fully computerized, but not as monotonous as that of many “screen speakers.” By the way, Xinhua has a virtual model for Chinese news. The image, facial expressions are superimposed on the voice, so that the articulation looks almost natural. Artificial intelligence helps in the processing of voice and “pictures” —the project uses a set of specialized algorithms.

The models were developed using the Sogou technology startup. According to representatives of the project, virtual speakers have “unlimited possibilities”, and their use costs the company less than working with people.

Virtual newsman can read the news as long as desired, at least a day away. With regard to limited facial expressions and synthesized voice, their company promises to improve - good, technology develops very quickly, gradually improving. The moment is not far off when real presenters and announcers cannot be distinguished from virtual ones.