Training someone to change the perceived gender of their own voice is a similar process to training them to speak with a new accent. I recently prototyped a gender-affirming voice therapy app for transgender women: https://own-voice-dev.netlify.app/
So now I’m curious if our two applications use the same approach under the hood! I have a detailed tech stack rundown on this page: https://own-voice-dev.netlify.app/#/info but in brief I use OpenVoice v2 to convert someone’s voice toward a differently-gendered reference voice, but leave it at only like 30%-60% transformed so it still sounds like them and not the reference. This transformed voice then becomes a practice reference for drilling.
I know it’s bad form to discuss that, but this is a completely normative comment, with no opinion given so far from being a flame. And it adds to the technical discussion!
Probably should make sure you're thinking about your actual UX when your third party voice cloning service cuts you off. Not a great sign when I can't even go through the demo without hitting a back end error from a third party.
I got ```
Voice cloning failed (403): {"detail":{"type":"authorization_error","code":"forbidden","message":"You have reached your monthly limit of voice add/edit operations (95). Please consider upgrading your subscription to increase your
```
Definitely something that you should have made sure doesn't leak into the front end UI.
The only accent I treat differently is the unsolicited spam call.
Face to face, I work with Indians, Ukrainians, Vietnamese, Mexicans, Brazilians, Russians, Guatemalans, Cubans, Italians and I love to hear the twang each person has.
Curious what you mean by treated differently or why you needed to put so much effort into hiding your accent? Are you sure it wasn't just general English? Because of course you may get treated differently if you can't speak English...but I think the accent is not really a problem.
Cool. Thank you for testing it. It can have multiple purposes. For people that want to do video presentations for instance and have a more native accent. Or yes, for a podcast. What do you think I should improve?
Voice cloning failed (403): {"detail":{"type":"authorization_error","code":"forbidden","message":"You have reached your monthly limit of voice add/edit operations (95). Please consider upgrading your subscription to increase your limit.","status":"voice_add_edit_limit_reached","request_id":"zzz"}}
This is good, although it left a fairly antipodean tone with a nerdy style when I selected British English. The source was a fairly heavy french accent, with the outcome being quite amusing.
The token limits seem to have kicked in now (my first go was a few hours back and this is just my second attempt)
So now I’m curious if our two applications use the same approach under the hood! I have a detailed tech stack rundown on this page: https://own-voice-dev.netlify.app/#/info but in brief I use OpenVoice v2 to convert someone’s voice toward a differently-gendered reference voice, but leave it at only like 30%-60% transformed so it still sounds like them and not the reference. This transformed voice then becomes a practice reference for drilling.