mlsanityData to 7 Oct 2026

TTS-PRISM: A Perceptual Reasoning and Interpretable Speech Model for Fine-Grained Diagnosis

arXiv 2604.22225 ↗First posted 24 Apr 2026cs.CL eess.AScited by 2

Not on the map. Why?

Abstract (arXiv)

While generative text-to-speech (TTS) models approach human-level quality, monolithic metrics fail to diagnose fine-grained acoustic artifacts or explain perceptual collapse. To address this, we propose TTS-PRISM, a multi-dimensional diagnostic framework for Mandarin. First, we establish a 12-dimensional schema spanning stability to advanced expressiveness. Second, we design a targeted synthesis pipeline with adversarial perturbations and expert anchors to build a high-quality diagnostic dataset. Third, schema-driven instruction tuning embeds explicit scoring criteria and reasoning into an efficient end-to-end model. Experiments on a 1,600-sample Gold Test Set show TTS-PRISM outperforms generalist models in human alignment. Profiling six TTS paradigms establishes intuitive diagnostic flags that reveal fine-grained capability differences. TTS-PRISM is open-source, with code and checkpoints at https://github.com/xiaomi-research/tts-prism.

Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Builds on 30 of its 33 references that are in the map. First posted on arXiv 24 Apr 2026.

Builds on

30 of 33 references.

References · 30

  1. 2025Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilitiescited by 3,655
  2. 2025Qwen3-Omni Technical Reportcited by 451
  3. 2024F5-TTS: A Fairytaler that Fakes Fluent and Faithful Speech with Flow Matchingcited by 353
  4. 2024CosyVoice 2: Scalable Streaming Speech Synthesis with Large Language Modelscited by 282
  5. 2024MT-Bench-101: A Fine-Grained Benchmark for Evaluating Large Language Models in Multi-Turn Dialoguescited by 247
  6. 2021NISQA: A Deep CNN-Self-Attention Model for Multidimensional Speech Quality Prediction with Crowdsourced Datasetscited by 189
  7. 2021WenetSpeech: A 10000+ Hours Multi-domain Mandarin Corpus for Speech Recognitioncited by 189
  8. 2024MaskGCT: Zero-Shot Text-to-Speech with Masked Generative Codec Transformercited by 159
  9. 2025CosyVoice 3: Towards In-the-wild Speech Generation via Scaling-up and Post-trainingcited by 142
  10. 2020AISHELL-3: A Multi-speaker Mandarin TTS Corpus and the Baselinescited by 111
  11. 2026Qwen3-TTS Technical Reportcited by 106
  12. 2025IndexTTS2: A Breakthrough in Emotionally Expressive and Duration-Controlled Auto-Regressive Zero-Shot Text-to-Speechcited by 85
18 more
  1. 2025MiMo-Audio: Audio Language Models are Few-Shot Learnerscited by 58
  2. 2024SpeechAlign: Aligning Speech Generation to Human Preferencescited by 45
  3. 2025VoxCPM: Tokenizer-Free TTS for Context-Aware Speech Generation and True-to-Life Voice Cloningcited by 39
  4. 2024SpeechBERTScore: Reference-Aware Automatic Evaluation of Speech Generation Leveraging NLP Evaluation Metricscited by 38
  5. 2025Step-Audio-R1 Technical Reportcited by 35
  6. 2025FireRedTTS-2: Towards Long Conversational Speech Generation for Podcast and Chatbotcited by 34
  7. 2024Emo-DPO: Controllable Emotional Speech Synthesis through Direct Preference Optimizationcited by 31
  8. 2021How do Voices from Past Speech Synthesis Challenges Compare Today?cited by 21
  9. 2025AudioJudge: Understanding What Works in Large Audio Model Based Speech Evaluationcited by 20
  10. 2022SOMOS: The Samsung Open MOS Dataset for the Evaluation of Neural Text-to-Speech Synthesiscited by 18
  11. 2025Advancing Zero-shot Text-to-Speech Intelligibility across Diverse Domains via Preference Alignmentcited by 18
  12. 2025WavReward: Spoken Dialogue Models With Generalist Reward Evaluatorscited by 15
  13. 2025SpeechJudge: Towards Human-Level Judgment for Speech Naturalnesscited by 15
  14. 2025NVSpeech: An Integrated and Scalable Pipeline for Human-Like Speech Modeling with Paralinguistic Vocalizationscited by 14
  15. 2025SpeechLLM-as-Judges: Towards General and Interpretable Speech Quality Evaluationcited by 14
  16. 2025VStyle: A Benchmark for Voice Style Adaptation with Spoken Instructionscited by 9
  17. 2025Audio Turing Test: Benchmarking the Human-likeness of Large Language Model-based Text-to-Speech Systems in Chinesecited by 4
  18. 2025Read to Hear: A Zero-Shot Pronunciation Assessment Using Textual Descriptions and LLMscited by 4

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026STEB: A Speech-to-Speech Translation Expressiveness Benchmark for Evaluating Beyond Translation Fidelitycited by 1
  2. 2026COT-TTS: Audio Context-Aware Text-to-Speech with Chain-of-Thought Reasoningcited by 0