Attention Is Not Always the Answer: Optimizing Voice Activity Detection with Simple Feature Fusion

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Tripathi, Kumud, Kumar, Chowdam Venkata, Wasnik, Pankaj
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912409667502080
author Tripathi, Kumud
Kumar, Chowdam Venkata
Wasnik, Pankaj
author_facet Tripathi, Kumud
Kumar, Chowdam Venkata
Wasnik, Pankaj
contents Voice Activity Detection (VAD) plays a key role in speech processing, often utilizing hand-crafted or neural features. This study examines the effectiveness of Mel-Frequency Cepstral Coefficients (MFCCs) and pre-trained model (PTM) features, including wav2vec 2.0, HuBERT, WavLM, UniSpeech, MMS, and Whisper. We propose FusionVAD, a unified framework that combines both feature types using three fusion strategies: concatenation, addition, and cross-attention (CA). Experimental results reveal that simple fusion techniques, particularly addition, outperform CA in both accuracy and efficiency. Fusion-based models consistently surpass single-feature models, highlighting the complementary nature of MFCCs and PTM features. Notably, our best-performing fusion model exceeds the state-of-the-art Pyannote across multiple datasets, achieving an absolute average improvement of 2.04%. These results confirm that simple feature fusion enhances VAD robustness while maintaining computational efficiency.
format Preprint
id arxiv_https___arxiv_org_abs_2506_01365
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Attention Is Not Always the Answer: Optimizing Voice Activity Detection with Simple Feature Fusion
Tripathi, Kumud
Kumar, Chowdam Venkata
Wasnik, Pankaj
Sound
Computation and Language
Audio and Speech Processing
Voice Activity Detection (VAD) plays a key role in speech processing, often utilizing hand-crafted or neural features. This study examines the effectiveness of Mel-Frequency Cepstral Coefficients (MFCCs) and pre-trained model (PTM) features, including wav2vec 2.0, HuBERT, WavLM, UniSpeech, MMS, and Whisper. We propose FusionVAD, a unified framework that combines both feature types using three fusion strategies: concatenation, addition, and cross-attention (CA). Experimental results reveal that simple fusion techniques, particularly addition, outperform CA in both accuracy and efficiency. Fusion-based models consistently surpass single-feature models, highlighting the complementary nature of MFCCs and PTM features. Notably, our best-performing fusion model exceeds the state-of-the-art Pyannote across multiple datasets, achieving an absolute average improvement of 2.04%. These results confirm that simple feature fusion enhances VAD robustness while maintaining computational efficiency.
title Attention Is Not Always the Answer: Optimizing Voice Activity Detection with Simple Feature Fusion
topic Sound
Computation and Language
Audio and Speech Processing
url https://arxiv.org/abs/2506.01365