# MarbleNet: Deep 1D Time-Channel Separable Convolutional Neural Network for Voice Activity Detection

Source: [MarbleNet: Deep 1D Time-Channel Separable Convolutional Neural Network for Voice Activity Detection](<https://www.sota2.com/research/paper/68fa694f1e209131cea216fb>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-25T22:57:51.173Z

Authors: Fei Jia, Somshubra Majumdar, Boris Ginsburg

Published: 2020-10-26T20:26:05.000Z

## Abstract

We present MarbleNet, an end-to-end neural network for Voice Activity Detection (VAD). MarbleNet is a deep residual network composed from blocks of 1D time-channel separable convolution, batch-normalization, ReLU and dropout layers. When compared to a state-of-the-art VAD model, MarbleNet is able to achieve similar performance with roughly 1/10-th the parameter cost. We further conduct extensive ablation studies on different training methods and choices of parameters in order to study the robustness of MarbleNet in real-world VAD tasks.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Voice Activity Detection on HAVIC (test)](<https://www.sota2.com/research/sota/voice-activity-detection-on-havic-test.md>): Voice Activity Detection; HAVIC (test); AUC-ROC: 0.804.

- [Voice Activity Detection on AVA-Speech (test)](<https://www.sota2.com/research/sota/voice-activity-detection-on-ava-speech-test.md>): Voice Activity Detection; AVA-Speech (test); AUC-ROC: 85.8.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2010.13886>)

[PDF](<https://arxiv.org/pdf/2010.13886.pdf>)