Searcharxiv⌕ Search
Home/Archive/arXiv

arXiv · 2508.05385

A Scalable Pipeline for Enabling Non-Verbal Speech Generation and Understanding

Runchuan Ye·Yixuan Zhou·Renjie Yu·Zijian Lin·Kehan Li·Xiang Li·Xin Liu·Guoyang Zeng·Zhiyong Wu
View original source ↗Read source document ↗Export metadata ↓

Abstract

Non-verbal Vocalizations (NVs), such as laughter and sighs, are vital for conveying emotion and intention in human speech, yet most existing speech systems neglect them, which severely compromises communicative richness and emotional intelligence. Existing methods for NVs acquisition are either costly and unscalable (relying on manual annotation/recording) or unnatural (relying on rule-based synthesis). To address these limitations, we propose a highly scalable automatic annotation framework to label non-verbal phenomena from natural speech, which is low-cost, easily extendable, and inherently diverse and natural. This framework leverages a unified detection model to accurately identify NVs in natural speech and integrates them with transcripts via temporal-semantic alignment method. Using this framework, we created and released \textbf{NonVerbalSpeech-38K}, a diverse, real-world dataset featuring 38,718 samples across 10 NV categories collected from in-the-wild media. Experimental results demonstrate that our dataset provides superior controllability for NVs generation and achieves comparable performance for NVs understanding.

Explore related subjects

cs.SDeess.AS

Keep this discovery

BibTeXRIS

Runchuan Ye, Yixuan Zhou, Renjie Yu, Zijian Lin, Kehan Li, Xiang Li, Xin Liu, Guoyang Zeng, Zhiyong Wu. 2025-08-07. A Scalable Pipeline for Enabling Non-Verbal Speech Generation and Understanding. https://arxiv.org/abs/2508.05385

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Xiaomi-CocktailASR-1 Technical Report

Recently, large language model (LLM) based ASR models have achieved significant progress, yet they generally lack support for multi-speaker scenarios, where the cocktail party problem remains a critical bottleneck for further advancing ASR. Existing TS-ASR methods, including end-to-end architectures with speaker embeddings and latest LLM-based explorations suffer from degraded single-speaker performance and the inability to reject when the target speaker is absent. In this paper, we propose Xiaomi-CocktailASR-1, an LLM-based end-to-end TS-ASR architecture. By utilizing reference speech as voiceprint prompts, it directly transcribes the target speaker's speech without requiring speech separation. Xiaomi-CocktailASR-1 maintains competitive performance in single-speaker scenarios, comparable to mainstream ASR models. It also features a negative sample rejection capability, outputting empty text when the target speaker is absent from the mixed speech. Additionally, Xiaomi-CocktailASR-1 supports a Chain-of-Thought (CoT) reasoning mode to provide explicit reasoning steps. Extensive experiments on various synthetic and real-world multispeaker benchmarks demonstrate that Xiaomi-CocktailASR-1 achieves state-of-the-art performance, effectively addressing the cocktail party problem through a unified architecture that balances multispeaker and single-speaker recognition accuracy, along with rejection capability.

cs.SD↗
arXiv2026

EConv-TasNet: Efficient Conv-TasNet for Effective Speech Separation

Conv-TasNet has served as a strong baseline for time-domain speech separation, and many studies have extended it with advanced architectures such as dual-path networks, U-Nets, and attention mechanisms. However, these methods often introduce high computational cost and complexity, limiting their deployment in resource-constrained scenarios. To address this issue, we propose eConv-TasNet, an efficient variant of Conv-TasNet that improves both effectiveness and efficiency without relying on resource-intensive modules. The proposed model consists of a group-wise early-splitting (GES) module and a multi-group feature aggregation (MGFA) module. GES generates discriminative speaker embeddings at intermediate stages, while MGFA progressively aggregates these group-level representations for refined mask estimation. Experimental results show that eConv-TasNet reduces model size by 22.4%, accelerates inference by 18.9%, and improves SI-SNRi by 14.0%-28.0% across three public benchmarks. Moreover, it achieves competitive performance compared with state-of-the-art methods while requiring significantly fewer parameters and lower inference cost. These results demonstrate a favorable efficiency-effectiveness trade-off for edge deployment.

cs.SD↗