Yandex introduced a system for automatic voice-over translation of video clips.

The system uses machine learning algorithms that automatically recognize and translate the speaker’s speech, and also generate a Russian-language version of the text. It is broadcast almost in real time, taking into account the speech tempo. Biometric algorithms also determine the gender of the speaker in order to choose the appropriate voice.

The technology is at the prototype stage and translates speech in the video only from English into Russian. The company continues to develop the service and plans to make the announcement of its public version in the near future.

Contacts:

Andrew Zakonov, CPO Smart Devices and Alice, Yandex

Original (in Russian)