4 ms·

I signed up for Descript and did the Overdub training a few weeks ago. It's impressive, albeit not a perfect match for my actual voice. You train the voice by r

I signed up for Descript and did the Overdub training a few weeks ago. It's impressive, albeit not a perfect match for my actual voice. You train the voice by reading 30 minutes of Wizard of Oz using a high quality microphone. I used a Yeti in a closet at my house. I do plan to record the additional hour of supplemental reading which might help make it more authentic. (Side note - I didn't know Dorothy's shoes were actual silver in the book) My use case for Overdub is to quickly correct errors in recorded demo/how-to screencast videos for my startup without having to do redo the audio. Here's a sample of the Overdub output vs. my actual recorded voice if you're curious: https://web.descript.com/084a416c-57d4-4df8-980b-24e0df82532f/b1d3d https://web.descript.com/084a416c-57d4-4df8-980b-24e0df82532...

kundan2510 6y ago
How does it compare with the training data you submitted to create your overdub voice? Sometimes, the voice sounds different because of the recording conditions, room tone, etc. and overdub voice copies these environmental factors as well.
- andygcook 6y ago
  Here's a sample clip of the training audio: https://web.descript.com/afdff9c8-17d0-41bb-84ec-77709120709b/e0fb9 https://web.descript.com/afdff9c8-17d0-41bb-84ec-77709120709... I recorded on a high quality microphone in a small closest with a foam mattress pad and blankets surrounding me. It was pretty soundproof. Upon re-listening to the audio, it does sound a little tinny, which likely has to do with the microphone settings. I think the issue with the monotone is that my reading voice is different from my talking voice. When I'm freely talking, I'm more dramatic. When I'm reading, I'm less so because the only time I read out loud nowadays is to children before bedtime. I do plan to resubmit the training script with a different style, but recording the full audio was admittedly tiring so I haven't sent that in yet. I'm sure with some experiment I could get it closer to reality. I did get some helpful advice from someone on the Descript team on how to use punctuation and misspellings to fix some of the intonation. Overall, it's really impressive technology and I can think of a few different uses for my business for my own voice.
jeremyw 6y ago
Thanks for the sample. There's a flatness to the audio in recreation, but the major difference is lack of inflection. I imagine the latter might be tunable in a future incarnation.
- andygcook 6y ago
  You can record your voice in different styles, which I think would help with inflection. I haven't had enough time to really experiment with it yet. The starting point is definitely an impressive effort on the Overdub/Descript team's side.

kundan2510 6y ago

How does it compare with the training data you submitted to create your overdub voice? Sometimes, the voice sounds different because of the recording conditions, room tone, etc. and overdub voice copies these environmental factors as well.

andygcook 6y ago
Here's a sample clip of the training audio: https://web.descript.com/afdff9c8-17d0-41bb-84ec-77709120709b/e0fb9 https://web.descript.com/afdff9c8-17d0-41bb-84ec-77709120709... I recorded on a high quality microphone in a small closest with a foam mattress pad and blankets surrounding me. It was pretty soundproof. Upon re-listening to the audio, it does sound a little tinny, which likely has to do with the microphone settings. I think the issue with the monotone is that my reading voice is different from my talking voice. When I'm freely talking, I'm more dramatic. When I'm reading, I'm less so because the only time I read out loud nowadays is to children before bedtime. I do plan to resubmit the training script with a different style, but recording the full audio was admittedly tiring so I haven't sent that in yet. I'm sure with some experiment I could get it closer to reality. I did get some helpful advice from someone on the Descript team on how to use punctuation and misspellings to fix some of the intonation. Overall, it's really impressive technology and I can think of a few different uses for my business for my own voice.

jeremyw 6y ago

Thanks for the sample. There's a flatness to the audio in recreation, but the major difference is lack of inflection. I imagine the latter might be tunable in a future incarnation.

andygcook 6y ago
You can record your voice in different styles, which I think would help with inflection. I haven't had enough time to really experiment with it yet. The starting point is definitely an impressive effort on the Overdub/Descript team's side.