Tags
objectsexperiments
Contributor
James Parker
Date
March 1, 2024
Folgezettel
13a
There are a growing number of online services and open source software projects for creating a custom voice for speech synthesis. We can learn a lot about their similarities and differences by using them, comparing the relationship between input and output, listening to their output across iterations, and searching for information about the way they have been implemented (in terms of neural network architecture, computing infrastructure, training data, etc.)
Commercial systems:
- Resemble AI
- ElevenLabs
- PlayHT
- Lovo.AI
Open source systems:
- Bark (Suno AI)
- SV2TTS (CorentinJ)
- GPT-SoVITS
- Parakeet