# SqueezeWave: Extremely Lightweight Vocoders for On-device Speech Synthesis

arXiv: https://arxiv.org/abs/2001.05685 · HTML: https://mlsanity.com/paper/2001.05685/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 16 Jan 2020.

## Built on by (most cited first, 5 of 5)

- [A Survey on Neural Speech Synthesis](https://mlsanity.com/paper/2106.15561/) (2021)
- [Lightweight and High-Fidelity End-to-End Text-to-Speech with Multi-Band Generation and Inverse Short-Time Fourier Transform](https://mlsanity.com/paper/2210.15975/) (2022)
- [Improving Generalization for AI-Synthesized Voice Detection](https://mlsanity.com/paper/2412.19279/) (2024)
- [On the Interplay Between Sparsity, Naturalness, Intelligibility, and Prosody in Speech Synthesis](https://mlsanity.com/paper/2110.01147/) (2021)
- [Multimodal Perception System for Real Open Environment](https://mlsanity.com/paper/2410.07926/) (2024)

## Abstract (arXiv)

Automatic speech synthesis is a challenging task that is becoming increasingly important as edge devices begin to interact with users through speech. Typical text-to-speech pipelines include a vocoder, which translates intermediate audio representations into an audio waveform. Most existing vocoders are difficult to parallelize since each generated sample is conditioned on previous samples. WaveGlow is a flow-based feed-forward alternative to these auto-regressive models (Prenger et al., 2019). However, while WaveGlow can be easily parallelized, the model is too expensive for real-time speech synthesis on the edge. This paper presents SqueezeWave, a family of lightweight vocoders based on WaveGlow that can generate audio of similar quality to WaveGlow with 61x - 214x fewer MACs. Code, trained models, and generated audio are publicly available at https://github.com/tianrengao/SqueezeWave.