5. Google’s Speech Commands Dataset
- Details: Although primarily focused on keyword recognition, this dataset contains a wide range of short speech samples that can be used for TTS model training.
- How to Use: Available here, this dataset is under the Apache 2.0 license.
Licensing Considerations:
When using these sources, always check the specific licensing terms, especially for commercial use, to ensure compliance. For instance:
- CC0 (Public Domain) means you can use it freely for any purpose.
- Apache 2.0 allows modification and commercial use, provided the original license is maintained.
Are you looking to use any specific TTS cloning models to implement these voices?