# Diffusion Synthesizer for Efficient Multilingual Speech to Speech Translation

arXiv: https://arxiv.org/abs/2406.10223 · HTML: https://mlsanity.com/paper/2406.10223/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 29 of its 32 references that are in the map.
- First posted on arXiv 14 Jun 2024.

## Builds on (most cited first, 29 of 29)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Flow Matching for Generative Modeling](https://mlsanity.com/paper/2210.02747/) (2022)
- [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://mlsanity.com/paper/2104.09864/) (2021)
- [Robust Speech Recognition via Large-Scale Weak Supervision](https://mlsanity.com/paper/2212.04356/) (2022)
- [FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness](https://mlsanity.com/paper/2205.14135/) (2022)
- [Decoupled Weight Decay Regularization](https://mlsanity.com/paper/1711.05101/) (2017)
- [SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition](https://mlsanity.com/paper/1904.08779/) (2019)
- [HiFi-GAN: Generative Adversarial Networks for Efficient and High Fidelity Speech Synthesis](https://mlsanity.com/paper/2010.05646/) (2020)
- [SoundStream: An End-to-End Neural Audio Codec](https://mlsanity.com/paper/2107.03312/) (2021)
- [High Fidelity Neural Audio Compression](https://mlsanity.com/paper/2210.13438/) (2022)
- [MLS: A Large-Scale Multilingual Dataset for Speech Research](https://mlsanity.com/paper/2012.03411/) (2020)
- [Voicebox: Text-Guided Multilingual Universal Speech Generation at Scale](https://mlsanity.com/paper/2306.15687/) (2023)
- [AudioLDM 2: Learning Holistic Audio Generation with Self-supervised Pretraining](https://mlsanity.com/paper/2308.05734/) (2023)
- [AudioPaLM: A Large Language Model That Can Speak and Listen](https://mlsanity.com/paper/2306.12925/) (2023)
- [NaturalSpeech 2: Latent Diffusion Models are Natural and Zero-Shot Speech and Singing Synthesizers](https://mlsanity.com/paper/2304.09116/) (2023)
- [Seamless: Multilingual Expressive and Streaming Speech Translation](https://mlsanity.com/paper/2312.05187/) (2023)
- [StyleTTS 2: Towards Human-Level Text-to-Speech through Style Diffusion and Adversarial Training with Large Speech Language Models](https://mlsanity.com/paper/2306.07691/) (2023)
- [SONAR: Sentence-Level Multimodal and Language-Agnostic Representations](https://mlsanity.com/paper/2308.11466/) (2023)
- [The People's Speech: A Large-Scale Diverse English Speech Recognition Dataset for Commercial Usage](https://mlsanity.com/paper/2111.09344/) (2021)
- [VioLA: Unified Codec Language Models for Speech Recognition, Synthesis, and Translation](https://mlsanity.com/paper/2305.16107/) (2023)
- [CVSS Corpus and Massively Multilingual Speech-to-Speech Translation](https://mlsanity.com/paper/2201.03713/) (2022)
- [Non-Attentive Tacotron: Robust and Controllable Neural TTS Synthesis Including Unsupervised Duration Modeling](https://mlsanity.com/paper/2010.04301/) (2020)
- [UnitY: Two-pass Direct Speech-to-speech Translation with Discrete Units](https://mlsanity.com/paper/2212.08055/) (2022)
- [TorchAudio-Squim: Reference-less Speech Quality and Intelligibility measures in TorchAudio](https://mlsanity.com/paper/2304.01448/) (2023)
- [Generative Pre-training for Speech with Flow Matching](https://mlsanity.com/paper/2310.16338/) (2023)
- [SpeechMatrix: A Large-Scale Mined Corpus of Multilingual Speech-to-Speech Translations](https://mlsanity.com/paper/2211.04508/) (2022)
- [Leveraging unsupervised and weakly-supervised data to improve direct speech-to-speech translation](https://mlsanity.com/paper/2203.13339/) (2022)
- [Translatotron 3: Speech to Speech Translation with Monolingual Data](https://mlsanity.com/paper/2305.17547/) (2023)
- [StyleS2ST: Zero-shot Style Transfer for Direct Speech-to-speech Translation](https://mlsanity.com/paper/2305.17732/) (2023)

## Abstract (arXiv)

We introduce DiffuseST, a low-latency, direct speech-to-speech translation system capable of preserving the input speaker's voice zero-shot while translating from multiple source languages into English. We experiment with the synthesizer component of the architecture, comparing a Tacotron-based synthesizer to a novel diffusion-based synthesizer. We find the diffusion-based synthesizer to improve MOS and PESQ audio quality metrics by 23% each and speaker similarity by 5% while maintaining comparable BLEU scores. Despite having more than double the parameter count, the diffusion synthesizer has lower latency, allowing the entire model to run more than 5× faster than real-time.