SonoEdit: Null-Space Constrained Knowledge Editing for Pronunciation Correction in LLM-Based TTS

• URL: http://arxiv.org/abs/2601.17086v1
• Date: Fri, 23 Jan 2026 08:40:49 GMT
• Title: SonoEdit: Null-Space Constrained Knowledge Editing for Pronunciation Correction in LLM-Based TTS
• Authors: Ayush Pratap Singh, Harshit Singh, Nityanand Mathur, Akshat Mandloi, Sudarshan Kamath,
• Abstract summary: We introduce SonoEdit, a model editing technique that surgically corrects pronunciation errors in pre-trained TTS models without retraining.<n>Instead of costly finetuning or explicit phoneme injection, we propose a parsimonious alternative based on Null-Space Pronunciation Editing.
• Score: 1.392548092257887
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Neural text-to-speech (TTS) systems systematically mispronounce low-resource proper nouns, particularly non-English names, brands, and geographic locations, due to their underrepresentation in predominantly English training corpora. Existing solutions typically rely on expensive multilingual data collection, supervised finetuning, or manual phonetic annotation, which limits the deployment of TTS systems in linguistically diverse settings. We introduce SonoEdit, a model editing technique that surgically corrects pronunciation errors in pre-trained TTS models without retraining. Instead of costly finetuning or explicit phoneme injection, we propose a parsimonious alternative based on Null-Space Pronunciation Editing, which performs a single-shot parameter update to modify the pronunciation of specific words while provably preserving all other model behavior. We first adapt Acoustic Causal Tracing to identify the Transformer layers responsible for text-to-pronunciation mapping. We then apply Null-Space Constrained Editing to compute a closed-form weight update that corrects the target pronunciation while remaining mathematically orthogonal to the subspace governing general speech generation. This constrained update steers the model's acoustic output toward a desired pronunciation exemplar while guaranteeing zero first-order change on a preserved speech corpus.

Related papers

• Counterfactual Activation Editing for Post-hoc Prosody and Mispronunciation Correction in TTS Models [19.852233854729235]
  Existing approaches for prosody manipulation often depend on specialized modules or additional training, limiting their capacity for post-hoc adjustments.<n>We introduce Counterfactual Activation Editing, a model-agnostic method that manipulates internal representations in a pre-trained TTS model to achieve post-hoc control of prosody and pronunciation.<n> Experimental results show that our method effectively adjusts prosodic features and corrects mispronunciations while preserving synthesis quality.
  arXiv  Detail & Related papers  (2025-06-01T04:33:37Z)
• FluentEditor2: Text-based Speech Editing by Modeling Multi-Scale Acoustic and Prosody Consistency [40.95700389032375]
  Text-based speech editing (TSE) allows users to edit speech by modifying the corresponding text directly without altering the original recording.<n>Current TSE techniques often focus on minimizing discrepancies between generated speech and reference within edited regions during training to achieve fluent TSE performance.<n>We propose a new fluency speech editing scheme based on our previous textitFluentEditor model, termed textittextbfFluentEditor2.
  arXiv  Detail & Related papers  (2024-09-28T10:18:35Z)
• HyPoradise: An Open Baseline for Generative Speech Recognition with Large Language Models [81.56455625624041]
  We introduce the first open-source benchmark to utilize external large language models (LLMs) for ASR error correction. The proposed benchmark contains a novel dataset, HyPoradise (HP), encompassing more than 334,000 pairs of N-best hypotheses. LLMs with reasonable prompt and its generative capability can even correct those tokens that are missing in N-best list.
  arXiv  Detail & Related papers  (2023-09-27T14:44:10Z)
• DTW-SiameseNet: Dynamic Time Warped Siamese Network for Mispronunciation Detection and Correction [1.8322859214908722]
  We present a highly-precise, PDA-compatible pronunciation learning framework for the task of TTS mispronunciation detection and correction. We also propose a novel mispronunciation detection model called DTW-SiameseNet, which employs metric learning with a Siamese architecture for Dynamic Time Warping (DTW) with triplet loss. Human evaluation shows our proposed approach improves pronunciation accuracy on average by 6% compared to strong phoneme-based and audio-based baselines.
  arXiv  Detail & Related papers  (2023-03-01T01:53:11Z)
• Thutmose Tagger: Single-pass neural model for Inverse Text Normalization [76.87664008338317]
  Inverse text normalization (ITN) is an essential post-processing step in automatic speech recognition. We present a dataset preparation method based on the granular alignment of ITN examples. One-to-one correspondence between tags and input words improves the interpretability of the model's predictions.
  arXiv  Detail & Related papers  (2022-07-29T20:39:02Z)
• Few-Shot Cross-Lingual TTS Using Transferable Phoneme Embedding [55.989376102986654]
  This paper studies a transferable phoneme embedding framework that aims to deal with the cross-lingual text-to-speech problem under the few-shot setting. We propose a framework that consists of a phoneme-based TTS model and a codebook module to project phonemes from different languages into a learned latent space.
  arXiv  Detail & Related papers  (2022-06-27T11:24:40Z)
• Revisiting End-to-End Speech-to-Text Translation From Scratch [48.203394370942505]
  End-to-end (E2E) speech-to-text translation (ST) often depends on pretraining its encoder and/or decoder using source transcripts via speech recognition or text translation tasks. In this paper, we explore the extent to which the quality of E2E ST trained on speech-translation pairs alone can be improved.
  arXiv  Detail & Related papers  (2022-06-09T15:39:19Z)
• Dict-TTS: Learning to Pronounce with Prior Dictionary Knowledge for Text-to-Speech [88.22544315633687]
  Polyphone disambiguation aims to capture accurate pronunciation knowledge from natural text sequences for reliable Text-to-speech systems. We propose Dict-TTS, a semantic-aware generative text-to-speech model with an online website dictionary. Experimental results in three languages show that our model outperforms several strong baseline models in terms of pronunciation accuracy.
  arXiv  Detail & Related papers  (2022-06-05T10:50:34Z)
• Towards Language Modelling in the Speech Domain Using Sub-word Linguistic Units [56.52704348773307]
  We propose a novel LSTM-based generative speech LM based on linguistic units including syllables and phonemes. With a limited dataset, orders of magnitude smaller than that required by contemporary generative models, our model closely approximates babbling speech. We show the effect of training with auxiliary text LMs, multitask learning objectives, and auxiliary articulatory features.
  arXiv  Detail & Related papers  (2021-10-31T22:48:30Z)
• Class LM and word mapping for contextual biasing in End-to-End ASR [4.989480853499918]
  In recent years, all-neural, end-to-end (E2E) ASR systems gained rapid interest in the speech recognition community. In this paper, we propose an algorithm to train a context aware E2E model and allow the beam search to traverse into the context FST during inference. Although an E2E model does not need pronunciation dictionary, it's interesting to make use of existing pronunciation knowledge to improve accuracy.
  arXiv  Detail & Related papers  (2020-07-10T20:58:44Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.