# InstructTTS: Modelling Expressive TTS in Discrete Latent Space with Natural Language Style Prompt

arXiv: https://arxiv.org/abs/2301.13662 · HTML: https://mlsanity.com/paper/2301.13662/

## Key facts

- Cited by 92 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 31 Jan 2023.

## Built on by (most cited first, 30 of 92)

- [NaturalSpeech 3: Zero-Shot Speech Synthesis with Factorized Codec and Diffusion Models](https://mlsanity.com/paper/2403.03100/) (2024)
- [Mini-Omni: Language Models Can Hear, Talk While Thinking in Streaming](https://mlsanity.com/paper/2408.16725/) (2024)
- [Audiobox: Unified Audio Generation with Natural Language Prompts](https://mlsanity.com/paper/2312.15821/) (2023)
- [VoiceCraft: Zero-Shot Speech Editing and Text-to-Speech in the Wild](https://mlsanity.com/paper/2403.16973/) (2024)
- [Natural language guidance of high-fidelity text-to-speech with synthetic annotations](https://mlsanity.com/paper/2402.01912/) (2024)
- [VDocRAG: Retrieval-Augmented Generation over Visually-Rich Documents](https://mlsanity.com/paper/2504.09795/) (2025)
- [TextrolSpeech: A Text Style Control Speech Corpus With Codec Language Text-to-Speech Models](https://mlsanity.com/paper/2308.14430/) (2023)
- [PromptTTS 2: Describing and Generating Voices with Text Prompt](https://mlsanity.com/paper/2309.02285/) (2023)
- [MegaTTS 3: Sparse Alignment Enhanced Latent Diffusion Transformer for Zero-Shot Speech Synthesis](https://mlsanity.com/paper/2502.18924/) (2025)
- [HierSpeech++: Bridging the Gap between Semantic and Acoustic Representation of Speech by Hierarchical Variational Inference for Zero-shot Speech Synthesis](https://mlsanity.com/paper/2311.12454/) (2023)
- [EmoVoice: LLM-based Emotional Text-To-Speech Model with Freestyle Text Prompting](https://mlsanity.com/paper/2504.12867/) (2025)
- [OpenVoice: Versatile Instant Voice Cloning](https://mlsanity.com/paper/2312.01479/) (2023)
- [Make-A-Voice: Unified Voice Synthesis With Discrete Representation](https://mlsanity.com/paper/2305.19269/) (2023)
- [PromptTTS++: Controlling Speaker Identity in Prompt-Based Text-to-Speech Using Natural Language Descriptions](https://mlsanity.com/paper/2309.08140/) (2023)
- [VoiceLDM: Text-to-Speech with Environmental Context](https://mlsanity.com/paper/2309.13664/) (2023)
- [SimpleSpeech: Towards Simple and Efficient Text-to-Speech with Scalar Latent Transformer Diffusion Models](https://mlsanity.com/paper/2406.02328/) (2024)
- [Towards Controllable Speech Synthesis in the Era of Large Language Models: A Systematic Survey](https://mlsanity.com/paper/2412.06602/) (2024)
- [Enhancing Zero-shot Text-to-Speech Synthesis with Human Feedback](https://mlsanity.com/paper/2406.00654/) (2024)
- [SpeechCraft: A Fine-grained Expressive Speech Dataset with Natural Language Description](https://mlsanity.com/paper/2408.13608/) (2024)
- [Scaling Rich Style-Prompted Text-to-Speech Datasets](https://mlsanity.com/paper/2503.04713/) (2025)
- [PromptVC: Flexible Stylistic Voice Conversion in Latent Space Driven by Natural Language Prompts](https://mlsanity.com/paper/2309.09262/) (2023)
- [ControlSpeech: Towards Simultaneous and Independent Zero-shot Speaker Cloning and Zero-shot Language Style Control](https://mlsanity.com/paper/2406.01205/) (2024)
- [SimpleSpeech 2: Towards Simple and Efficient Text-to-Speech with Flow-based Scalar Latent Transformer Diffusion Models](https://mlsanity.com/paper/2408.13893/) (2024)
- [Audio-Aware Large Language Models as Judges for Speaking Styles](https://mlsanity.com/paper/2506.05984/) (2025)
- [InstructTTSEval: Benchmarking Complex Natural-Language Instruction Following in Text-to-Speech Systems](https://mlsanity.com/paper/2506.16381/) (2025)
- [EmotiveTalk: Expressive Talking Head Generation through Audio Information Decoupling and Emotional Video Diffusion](https://mlsanity.com/paper/2411.16726/) (2024)
- [EmoSpeech: Guiding FastSpeech2 Towards Emotional Text to Speech](https://mlsanity.com/paper/2307.00024/) (2023)
- [LibriTTS-P: A Corpus with Speaking Style and Speaker Identity Prompts for Text-to-Speech and Style Captioning](https://mlsanity.com/paper/2406.07969/) (2024)
- [FLEURS-R: A Restored Multilingual Speech Corpus for Generation Tasks](https://mlsanity.com/paper/2408.06227/) (2024)
- [EmoKnob: Enhance Voice Cloning with Fine-Grained Emotion Control](https://mlsanity.com/paper/2410.00316/) (2024)

## Abstract (arXiv)

Expressive text-to-speech (TTS) aims to synthesize different speaking style speech according to human's demands. Nowadays, there are two common ways to control speaking styles: (1) Pre-defining a group of speaking style and using categorical index to denote different speaking style. However, there are limitations in the diversity of expressiveness, as these models can only generate the pre-defined styles. (2) Using reference speech as style input, which results in a problem that the extracted style information is not intuitive or interpretable. In this study, we attempt to use natural language as style prompt to control the styles in the synthetic speech, e.g., "Sigh tone in full of sad mood with some helpless feeling". Considering that there is no existing TTS corpus which is proper to benchmark this novel task, we first construct a speech corpus, whose speech samples are annotated with not only content transcriptions but also style descriptions in natural language. Then we propose an expressive TTS model, named as InstructTTS, which is novel in the sense of following aspects: (1) We fully take the advantage of self-supervised learning and cross-modal metric learning, and propose a novel three-stage training procedure to obtain a robust sentence embedding model, which can effectively capture semantic information from the style prompts and control the speaking style in the generated speech. (2) We propose to model acoustic features in discrete latent space and train a novel discrete diffusion probabilistic model to generate vector-quantized (VQ) acoustic tokens rather than the commonly-used mel spectrogram. (3) We jointly apply mutual information (MI) estimation and minimization during acoustic model training to minimize style-speaker and style-content MI, avoiding possible content and speaker information leakage from the style prompt.