TokAN is presented, a token-based accent normalization framework that operates on self-supervised discrete speech tokens extracted from a L1-L2 jointly trained vector-quantization (VQ) tokenizer, without the need of synthetic supervisory speech.
Abstract
Accent normalization (AN) seeks to convert non-native (L2) accented speech into standard (L1) speech while preserving speaker identity. The current techniques either require naturally recorded parallel L1-L2 speech for training, or suffer from quality degradation when supervised by synthesized targets. In this paper, we present TokAN, a token-based accent normalization framework that operates on self-supervised discrete speech tokens extracted from a L1-L2 jointly trained vector-quantization (VQ) tokenizer, without the need of synthetic supervisory speech. An autoregressive encoder-decoder model performs token-to-token conversion, translating L2-accented token sequences into the tokens of standard voice. We also introduce reinforcement learning (RL) post-training based on Group Relative Policy Optimization (GRPO), using word error rate and accent classifier confidence as complementary rewards. A non-autoregressive flow-matching synthesizer recovers the Mel-spectrogram from the converted tokens, conditioned on the source speaker embedding. We also develop a flow-matching duration predictor that supports total-duration-aware synthesis, making TokAN applicable to duration-critical tasks such as voice dubbing and live casting. Experiments on seven English accents demonstrate that TokAN reduced the word error rate from 12.40% to 9.89% after supervised fine-tuning, and further to 9.23% after RL post-training, consistently outperforming frame-to-frame, direct flow-matching, and prompt-based token-conversion baselines in terms of accent reduction and intelligibility.
Semantic speech tokens should preserve linguistic content while suppressing speaker- and duration-dependent variation inherited from acoustic inputs. We propose Iterative Semantic Token Purification (ISTP), an alternating speech-to-unit (S2U) and text-to-unit (T2U) training procedure guided by text predictability. Starting from an initial S2U tokenizer, each iteration trains a T2U model on its deduplicated token sequences. The decoded T2U predictions then serve as connectionist temporal classification targets for a newly initialized S2U model, whose outputs supervise the next T2U model. This cycle progressively aligns the two token generators and biases the token space toward information recoverable from text. Experiments on Mandarin and English show substantially improved S2U--T2U agreement. Independently trained de-tokenizers further show that the refined S2U and T2U tokens retain sufficient content for high-intelligibility voice conversion and text-to-speech synthesis. In voice conversion, the generated speaking rate follows the reference more closely. The refined tokens also exhibit substantially improved cross-speaker consistency and reduced probe-recoverable speaker information.
Hanlin Zhang, Daxin Tan, Dehua Tao et al.· 0 citations
Discrete speech tokenizers aim to disentangle semantic from acoustic information, yet targets from self-supervised learning (SSL) models like HuBERT retain non-linguistic variation: speaker identity, prosody, and channel conditions leak into the tokens, inflating entropy. Our key insight is that when enough speakers utter the same words under varying conditions, linguistic content is the only shared factor. We propose PINT (Parallel INvariant Tokenization), which fine-tunes an SSL encoder with alignment losses across parallel utterances and augmentations to distill this shared residual. PINT collapses identical words onto consistent token sequences, drastically reducing conditional entropy. Unlike ASR text, PINT tokens preserve frame-level temporal grounding and serve as drop-in semantic targets for audio codecs. Experiments show a 98.7% relative reduction in speaker probe accuracy (93.1% to 1.2%), a 42% lower ABX error rate, and 27-30% lower LM perplexity versus baselines, confirming that the right invariance is key to efficient learning.
Laurin Wagner, Bernhard Thallinger, Miroslav Stankovič et al.· arXiv.org· 2 citations
In current zero-shot text-to-speech systems, conventional semantic tokenizers are typically optimized using supervised automatic speech recognition or self-supervised learning objectives. However, due to the inherent nature of speech, semantic and acoustic information cannot be completely decoupled, and ASR-based tokenizers discard acoustic details to focus on linguistic content; models relying on them usually struggle to achieve optimal speaker similarity. Furthermore, these tokenizers are optimized independently and lack direct supervision from downstream acoustic generation tasks. This isolated training creates a feature gap between the extracted discrete tokens and the continuous space required by acoustic models, fundamentally bottlenecking the upper bound of synthesis quality. To bridge this gap, we propose Phoenix TTS, a unified framework that tightly couples representation learning with generative acoustic modeling. Specifically, our speech tokenizer is optimized to reconstruct self-supervised features to maintain semantic richness, while simultaneously receiving direct supervision from a Flow Matching training loss. Through this joint training paradigm, the extracted discrete tokens successfully preserve essential semantic information and natively align with the feature space of the downstream Flow Matching model. Comprehensive evaluations highlight the efficiency and effectiveness of Phoenix TTS. Trained on 110K hours of data, the system achieves excellent speech intelligibility, yielding WER that consistently falls below that of ground-truth recordings. Simultaneously, it maintains robust zero-shot speaker similarity, rivaling or outperforming several prominent large-scale baselines. Furthermore, as an advantageous byproduct of this unified training, the learned tokenizer can be seamlessly adapted to zero-shot voice conversion tasks without requiring task-specific fine-tuning.
Peijie Chen, Zhuanling Zha, Zhipeng Nie et al.· 0 citations
ParaASR is introduced, an ASR system that leverages Multi-Token Prediction (MTP) to let a 4B LLM decoder emit multiple tokens per forward step and shows that decoder scaling, low-latency inference, and long-context transcription need not be competing goals when future-token proposals are anchored by the acoustic signal and guarded by autoregressive verification.
Voice conversion (VC) is an emerging technology in speech processing that aims to modify an utterance so that it sounds as though it were spoken by another speaker while preserving its linguistic content. High-quality voice conversion has broad applications, including speech synthesis, assistive communication, and entertainment applications such as multilingual dubbing. However, current embedding-guided voice conversion (EGVC) frameworks often struggle with generalization and naturalness under regional data-scarcity conditions. This study explores these limitations by evaluating an EGNVC framework adapted for low-resource regional-language pairs. The proposed framework incorporates the Harvest pitch-extraction algorithm alongside pretrained speaker representations to guide cross-gender pitch transitions while attempting to preserve speaker-identity profiles. Experimental results show that, although the framework successfully shifts macro-level pitch contours across genders, spectral alterations lead to substantial acoustic distortion and reduced intelligibility. Specifically, Kannada speech conversion achieved a localized objective intelligibility score of STOI = 0.12, whereas Malayalam transformations exhibited substantial spectral variation, with an MCD of 169.75, highlighting significant language-specific barriers to regional voice conversion. Kannada achieved higher intelligibility (STOI = 0.12) than Malayalam, whereas Malayalam required greater spectral modification (MCD = 169.75), indicating language-specific challenges in voice conversion. These baseline metrics delineate the empirical limitations of current embedding-guided architectures for Dravidian languages and indicate that substantial advances in spectral mapping are required before such systems can be integrated into real-time assistive or localized voice-synthesis applications.
B. A, Singh S. P., Dhiraj Sunehra· International Research Journ...· 0 citations
ReLMCodec is a low-bitrate single-codebook speech codec built upon a preserve--control--refine principle that moves the empirical single-stream predictability--reconstruction frontier in the evaluations, with gains that carry over to downstream text-to-speech (TTS) synthesis in both intelligibility and speaker similarity.
Zixiang Wan, Xusheng Yang, Zhengmeng Wang et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
