This repository is a curated list of academic papers on speech synthesis, targeting researchers and engineers in the field. It provides a structured overview of key advancements and methodologies in text-to-speech (TTS) technology, enabling users to quickly identify foundational and state-of-the-art research.

How It Works

The list categorizes papers across various sub-fields of speech synthesis, including TTS frontends, acoustic models (autoregressive and non-autoregressive), vocoders, and specialized areas like expressive TTS and voice conversion. Papers are annotated with a star (★) indicating over 50 citations, serving as a guide for beginners to prioritize influential works.

Highlighted Details

• Comprehensive categorization of TTS research, from foundational models like Tacotron and WaveNet to recent diffusion-based approaches.
• Includes papers on specialized TTS tasks such as expressive speech, multi-speaker synthesis, and voice conversion.
• Papers are marked with citations counts (★) to highlight highly influential works.
• Covers a wide range of techniques including autoregressive, non-autoregressive, flow-based, GAN-based, and diffusion-based models.

Maintenance & Community

This is a community-driven list, welcoming recommendations for new papers. It references other curated lists like "awesome-speech-recognition-speech-synthesis-papers" and "awesome-tts-samples."

Licensing & Compatibility

The repository itself does not contain code, only a list of papers. Licensing information would pertain to the individual papers or their associated code repositories, which are not directly hosted here.

Limitations & Caveats

The list is a compilation of research papers and does not provide implementations, code, or datasets. The "★" citation count is a manual annotation and may not be exhaustive or perfectly up-to-date.