At the AZ conference we had a booth to collect My Stories. However, this room was a very busy and noisy room. It had the merch perch, book sales, and forums booths as well.
We did the best we could with a directional mic, but still a lot of background is picked up.
The request is to find and utilize tools and techniques that would result in a clean audio track.
One approach is to find tools that greatly reduce background noise and enhance the primary voice. I’m personally skeptical of how far we can take those tools.
The alternative idea is similar to what was used for the Javier Milei to translate is epic WEF speech to English while maintaining his voice, accent, mannerisms, etc.
We can use AI tools to extract a good transcript. Train another AI on their voice with what good clips we can find in the audio track. Then dub over based on the transcript with their trained voice. A completely clean audio track.
Replied in this thread, we will provide samples for experimentation.
Hello and have to say I’m missing you guys, not a lot, but enough.
I’ve just listened to both recordings and the original audio is useable, but of course I was hoping for all my work to already have been done, so I could sit back and take all the credit.
May I ask the status of the deep fakes please. What can I do?
