YouTube integrates automatic audio dubbing through AI
[15:55 Mon,31.July 2023 by blip]
For quite a while now, YouTube has been testing an AI-powered dubbing feature that can easily and quickly sync videos into multiple languages. The feature was developed under the name Aloud as part of Google's Area 120 incubator program and was officially adopted by YouTube in June. It's not entirely clear how many Creator Channels currently have access to the feature in Creator Studio; there are said to be at least several hundred by now. However, dubbing does not seem to have been released to all users yet, nor are all languages supported.
Audio track selection at the MrBeast YouTube channel
In order to dub the spoken words in a video, a transcript is output first, which can be corrected manually if necessary. In dubbing, the new audio tracks are then "spoken" by synthetic voices, as is also known from similar external services. According to Google, it is planned to be able to imitate the original voices in the future, and automatic adaptation of lip movements is also planned, probably next year - for example, Flawless AI had already presented something similar for the cinema with Truesync.
The alternative audio tracks can be activated - if available - in the video settings, where the resolution etc. can also be changed. The sensation-heavy YouTube channel MrBeast, for example, one of the largest with 172 million subscribers, is allowed to use the function as a beta tester from the beginning and offers videos in 14 languages (see below), including French, Turkish or Japanese, but no German.
Automatic dubbing abruptly expands the potential reach for many YouTube videos. Subtitles already had a great effect in this respect. At the same time, dubbing studios, which previously provided dubbing for large YouTube channels for a fee, are now out of luck. Because automatic YouTube dubbing is not only simple, but also free.
;)
