Home technology software deep voice software helps cloning just 3.7 seconds of audio

Deep voice software helps cloning just 3.7 seconds of audio

Software

CIO Bulletin
2018-03-08

Can you believe?! It is possible for a Chinese tech giant Baidu to clone fake voice. Researchers at the tech giant divulged advancement in Deep Voice which is designed for cloning voices. Research from Tech giant makes us hard to believe any piece of media on the internet.

Cloning is done in just 3.7 seconds of audio. A year back, the technology required 30 minutes of audio to create a fake audio clip. But now, your work is done with better results in seconds. “Voice cloning is expected to have significant applications in the direction of personalization in human-machine interfaces,” the researchers write in a Baidu blog article on the study.

This technology can change a female voice to male and from British accent to American. AI is helping in learning to mimic different styles of speaking and takes personalizing text-to-speech to a whole new level.

There was lot of developments made in this field. In 2016 Adobe demonstrated its VoCo software and Montreal-based AI startup Lyrebird can work on text-to-speech using one minute of audio. If this is all that is needed then this technology can help a lot in fabricating press conference, news segment and interview.