Latent Diffusion Models Are Zero-Shot Speech and Singing Synthesizers

Re:Latent Diffusion Models Are Zero-Shot Speech and Singing ... by gigel82 on Wednesday April 19, 2023 @12:48PM (#35630405)

I like https://github.com/neonbjb/tortoise-tts ; it doesn't do singing, but the voice reproduction is very good and -most importantly- it's open source and you can run it locally.

[ Parent | on HN ]

Re:Latent Diffusion Models Are Zero-Shot Speech and Singing ... by samuelzxu on Wednesday April 19, 2023 @01:22PM (#35630792)

Woah! How is this not more popular? I don't see it referenced in the naturalspeech2 paper anywhere.

[ Parent | on HN ]

Re:Latent Diffusion Models Are Zero-Shot Speech and Singing ... by woodson on Wednesday April 19, 2023 @05:59PM (#35633843)

That's likely because it wasn't published as a paper anywhere (not even arxiv) and the author then joined OpenAI and development ceased.

[ Parent | on HN ]

Re:Latent Diffusion Models Are Zero-Shot Speech and Singing ... by omeze on Wednesday April 19, 2023 @09:54PM (#35635461)

unfortunately it can't be used for real-time use cases, but YourTTS is just as good and faster (except... non-commercial)

[ Parent | on HN ]