# RetrieverTTS: Modeling Decomposed Factors for Text-Based Speech Insertion

arXiv: https://arxiv.org/abs/2206.13865 · HTML: https://mlsanity.com/paper/2206.13865/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 28 Jun 2022.

## Built on by (most cited first, 10 of 10)

- [NaturalSpeech 2: Latent Diffusion Models are Natural and Zero-Shot Speech and Singing Synthesizers](https://mlsanity.com/paper/2304.09116/) (2023)
- [VoiceCraft: Zero-Shot Speech Editing and Text-to-Speech in the Wild](https://mlsanity.com/paper/2403.16973/) (2024)
- [Autoregressive Diffusion Transformer for Text-to-Speech Synthesis](https://mlsanity.com/paper/2406.05551/) (2024)
- [DiffVoice: Text-to-Speech with Latent Diffusion](https://mlsanity.com/paper/2304.11750/) (2023)
- [SongCreator: Lyrics-based Universal Song Generation](https://mlsanity.com/paper/2409.06029/) (2024)
- [Improving Code-Switching and Named Entity Recognition in ASR with Speech Editing based Data Augmentation](https://mlsanity.com/paper/2306.08588/) (2023)
- [ELF: Encoding Speaker-Specific Latent Speech Feature for Speech Synthesis](https://mlsanity.com/paper/2311.11745/) (2023)
- [Speech Editing -- a Summary](https://mlsanity.com/paper/2407.17172/) (2024)
- [RephraseTTS: Dynamic Length Text based Speech Insertion with Speaker Style Transfer](https://mlsanity.com/paper/2508.17031/) (2025)
- [Speak, Edit, Repeat: High-Fidelity Voice Editing and Zero-Shot TTS with Cross-Attentive Mamba](https://mlsanity.com/paper/2510.04738/) (2025)

## Abstract (arXiv)

This paper proposes a new "decompose-and-edit" paradigm for the text-based speech insertion task that facilitates arbitrary-length speech insertion and even full sentence generation. In the proposed paradigm, global and local factors in speech are explicitly decomposed and separately manipulated to achieve high speaker similarity and continuous prosody. Specifically, we proposed to represent the global factors by multiple tokens, which are extracted by cross-attention operation and then injected back by link-attention operation. Due to the rich representation of global factors, we manage to achieve high speaker similarity in a zero-shot manner. In addition, we introduce a prosody smoothing task to make the local prosody factor context-aware and therefore achieve satisfactory prosody continuity. We further achieve high voice quality with an adversarial training stage. In the subjective test, our method achieves state-of-the-art performance in both naturalness and similarity. Audio samples can be found at https://ydcustc.github.io/retrieverTTS-demo/.