3 ms·
I found the Obama video at the end very interesting. It would be a neat next step to map non-Obama audio to the generated video. For example, pull audio from
by 33W 9y ago
I found the Obama video at the end very interesting. It would be a neat next step to map non-Obama audio to the generated video. For example, pull audio from an Obama impersonator.
- visarga 9y ago> pull audio from an Obama impersonator We can impersonate voices with neural nets. We can clone timbre and style, and this tech is being used commercially by Baidu at the very least (keyword: Deep Voice 3).