Is there a reason why no female voice is available in this great project?
Do we have more data on male speech than female speech?
Show HN: Neural text to speech with dozens of celebrity voices
151–160 of 174 posts
Re: Show HN: Neural text to speech with dozens of celebrity voices
#152Earlier quoted context omitted.
Photos were being doctored long before then, even. For example, https://en.wikipedia.org/wiki/Censorship_of_images_in_the_So...
I'm reminded of a comment by someone that studied the old Eugenics movement. Mentioned finding that one of the major books on the subject the author had carefully edited photo's of 'degenerate people' to look even more crude an unintelligent. Also photographer friend of mine said; a great photographer doesn't need to Photoshop anything to lie to you.
> Also photographer friend of mine said; a great photographer doesn't need to Photoshop anything to lie to you.
I couldn't agree more if they used the word lie in the more general sense as a synonym of deception. The availability of fakes may not become a problem because the most effective deception doesn't involve telling untruths.
To give an example, suppose Russia Today and Fox News report on the same event. There's a set of facts. RT picks a subset and reports it from their point of view. Fox picks another subset and presents their view. The resulting articles may give readers vastly different interpretation of the event and no untruths had to be involved.
Re: Show HN: Neural text to speech with dozens of celebrity voices
#153I've built a lot of celebrity text to speech models and host them online: https://vo.codes It has celebrities like Sir David Attenborough and Arnold Schwarzenegger, a bunch of the presidents, and also some engineers: PG, Sam Altman, Peter Thiel, Mark Zuckerberg I'm not far away from a working "real time" [1] voice conversion (VC) system. This turns a source voice into a target voice. The most difficult part is gettin…
What is the reason for them being almost all men?
My guess would be that there are more interesting and recognizable voices by men than by women. Not everything is a CIS male's conspiracy you know.
Re: Show HN: Neural text to speech with dozens of celebrity voices
#154Is there a reason why no female voice is available in this great project? Do we have more data on male speech than female speech?
What iconic female voices would you add?
Re: Show HN: Neural text to speech with dozens of celebrity voices
#155I've built a lot of celebrity text to speech models and host them online: https://vo.codes It has celebrities like Sir David Attenborough and Arnold Schwarzenegger, a bunch of the presidents, and also some engineers: PG, Sam Altman, Peter Thiel, Mark Zuckerberg I'm not far away from a working "real time" [1] voice conversion (VC) system. This turns a source voice into a target voice. The most difficult part is gettin…
What is the reason for them being almost all men?
It's extremely hard to get women's voices to sound right in TTS, at least back when I was working on it. It was a striking difference, and I'm not sure why.
e.g. I tried to do an Alyx version of https://www.youtube.com/watch?v=koU3L7WBz_s but it came out sounding nothing like her.
Re: Show HN: Neural text to speech with dozens of celebrity voices
#156Re: Show HN: Neural text to speech with dozens of celebrity voices
#157Earlier quoted context omitted.
Yep, that's what we're doing at https://replicastudios.com - though we have strict ethical guidelines to not clone a person's voice without their permission.
Can this service be used with preexisting recordings? I have family that are losing the ability to speak and I’d love to be able to give them that back somehow.
Johnkpaul, you might want to check out https://news.ycombinator.com/item?id=23490115 , it has a plethora of helpful ideas on topic.
Re: Show HN: Neural text to speech with dozens of celebrity voices
#158Wondering whether there is an androgynous TTS voice that sounds neither male nor female.
Re: Show HN: Neural text to speech with dozens of celebrity voices
#159Is there a reason why no female voice is available in this great project? Do we have more data on male speech than female speech?
What iconic female voices would you add?
> What iconic female voices would you add?
Cate Blanchett, Sarah Silverman, Katey Sagal, Jennifer Tilly, Laura Prepon, Viola Davis, Judi Dench, Whoopi Goldberg, Julie Andrews, Lake Bell, Jane Lynch, Joan Rivers, Martha Stewart, Katharine Hepburn, Sarah Vowell, Shoreh Aghdashloo...
Re: Show HN: Neural text to speech with dozens of celebrity voices
#160Earlier quoted context omitted.
Have you had any success with using speaker embeddings to generate voices with fewer samples of speech? I did some cursory experiments but I couldn't get too far beyond getting pitch similar to the target speaker. My reasoning for this approach: IMO, if the model learns a "universal human voice", it shouldn't need too much additional information to get a target voice.
I did! I tried creating a multi-speaker embedding model for practical concerns: saving on memory costs. I'm going to have to add additional layers, because it didn't fit individual speakers very well. I wish I'd saved audio results to share. I might be able to publish my findings if I look around for the model files. I think you're right in that if we can get such a model to work, training new embeddings won't requir…
Hmm. Would a multi-speaker model be able to interpolate between voices (eg. halfway between Morgan Freeman and James Earl Jones)?