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

Xi Wang, Jie Wang, Xingchen Song, Baijun Song, Jingran Xie +6 more
4/24/2026
cs.CLeess.AS

Official code: xiaomi-research/tts-prism (52 stars, Apache-2.0, last commit 2026-04-27).

Abstract

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.

View on arXivView PDF

Code Implementations(1)

xiaomi-research/tts-prismOfficial100%
Apache-2.0
524PythonApr 10, 20265 months agoApache-2.0

Similar Papers

PortaSpeech: Portable and High-Quality Generative Text-to-Speech

84% match1 repo1.0k2021

How Generative Spoken Language Modeling Encodes Noisy Speech: Investigation from Phonetics to Syntactics

82% match2023

AVGen-Bench: A Task-Driven Benchmark for Multi-Granular Evaluation of Text-to-Audio-Video Generation

82% match2026

SPEARBench: A Benchmark for Naturalness Evaluation in Streaming Speech-to-Speech Language Models

82% match2026

XSQ-AST: An Explainable Audio Spectrogram Transformer Framework for Localising Synthetic Speech Artifacts

82% match2026

Improving Robustness of LLM-based Speech Synthesis by Learning Monotonic Alignment

81% match2024

Cite this paper

@article{wang2026ttsprism,
  title  = {TTS-PRISM: A Perceptual Reasoning and Interpretable Speech Model for Fine-Grained Diagnosis},
  author = {Xi Wang and Jie Wang and Xingchen Song and Baijun Song and Jingran Xie and Jiahe Shao and Zijian Lin and Di Wu and Meng Meng and Jian Luan and Zhiyong Wu},
  year   = {2026},
  eprint = {2604.22225},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2604.22225v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2604.22225.svg)](https://codeofpaper.com/paper/2604.22225)

Discussion