mlsanityData to 14 Oct 2026

An Evaluation Framework for Text-to-Speech Voice Reconstruction

arXiv 2606.21343 ↗First posted 19 Jun 2026eess.AS cs.CL cs.SDcited by 1

Not on the map. Why?

Abstract (arXiv)

Voice reconstruction using Text-to-Speech (TTS) offers a communication method for people with speech disorders, which aims to retain their speaker identity while improving intelligibility. Previous work generally relies on Mean Opinion Score (MOS) to evaluate naturalness and speaker similarity, but this has limited sensitivity and reliability. We propose an evaluation framework with subjective and objective components. Subjectively, we evaluate perceived intelligibility and speaker identity using Best Worst Scaling (BWS) with situational framing. Objectively, we demonstrate that standard measures fail to predict reconstruction success for highly unintelligible speakers, so we introduce a novel dual-reference distributional measure to assess the trade-off between intelligibility and speaker identity. By evaluating the output of 17 zero-shot TTS systems for 193 speakers, we show that our framework provides a reliable and task-aligned approach for assessing voice reconstruction.

Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Builds on 29 of its 52 references that are in the map. First posted on arXiv 19 Jun 2026.

Builds on

29 of 52 references.

References · 29

  1. 2022Robust Speech Recognition via Large-Scale Weak Supervisioncited by 4,057
  2. 2019LibriTTS: A Corpus Derived from LibriSpeech for Text-to-Speechcited by 649
  3. 2022UTMOS: UTokyo-SaruLab System for VoiceMOS Challenge 2022cited by 390
  4. 2024F5-TTS: A Fairytaler that Fakes Fluent and Faithful Speech with Flow Matchingcited by 361
  5. 2024XTTS: a Massively Multilingual Zero-Shot Text-to-Speech Modelcited by 179
  6. 2024MaskGCT: Zero-Shot Text-to-Speech with Masked Generative Codec Transformercited by 161
  7. 2023StyleTTS 2: Towards Human-Level Text-to-Speech through Style Diffusion and Adversarial Training with Large Speech Language Modelscited by 152
  8. 2024E2 TTS: Embarrassingly Easy Fully Non-Autoregressive Zero-Shot TTScited by 152
  9. 2024VoiceCraft: Zero-Shot Speech Editing and Text-to-Speech in the Wildcited by 120
  10. 2026Qwen3-TTS Technical Reportcited by 118
  11. 2022Wespeaker: A Research and Production oriented Speaker Embedding Learning Toolkitcited by 97
  12. 2025IndexTTS2: A Breakthrough in Emotionally Expressive and Duration-Controlled Auto-Regressive Zero-Shot Text-to-Speechcited by 90
17 more
  1. 2020Universal Phone Recognition with a Multilingual Allophone Systemcited by 81
  2. 2023Better speech synthesis through scalingcited by 63
  3. 2024Fish-Speech: Leveraging Large Language Models for Advanced Multilingual Text-to-Speech Synthesiscited by 48
  4. 2025Vevo: Controllable Zero-Shot Voice Imitation with Self-Supervised Disentanglementcited by 35
  5. 2023OpenVoice: Versatile Instant Voice Cloningcited by 32
  6. 2024TTSDS -- Text-to-Speech Distribution Scorecited by 14
  7. 2024AccentBox: Towards High-Fidelity Zero-Shot Accent Generationcited by 9
  8. 2025Personalized Fine-Tuning with Controllable Synthetic Speech from LLM-Generated Transcripts for Dysarthric Speech Recognitioncited by 6
  9. 2025DiffDSR: Dysarthric Speech Reconstruction Using Latent Diffusion Modelcited by 5
  10. 2024CoLM-DSR: Leveraging Neural Codec Language Modeling for Multi-Modal Dysarthric Speech Reconstructioncited by 4
  11. 2025Good practices for evaluation of synthesized speechcited by 4
  12. 2025The State Of TTS: A Case Study with Human Fooling Ratescited by 4
  13. 2025Can we reconstruct a dysarthric voice with the large speech model Parler TTS?cited by 3
  14. 2025Finding My Voice: Generative Reconstruction of Disordered Speech for Automated Clinical Evaluationcited by 3
  15. 2024Creating Personalized Synthetic Voices from Articulation Impaired Speech Using Augmented Reconstruction Losscited by 2
  16. 2025Facilitating Personalized TTS for Dysarthric Speakers Using Knowledge Anchoring and Curriculum Learningcited by 2
  17. 2025Layer-wise Analysis for Quality of Multilingual Synthesized Speechcited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026A Cross Community Agenda for Speech AIcited by 0