LibriVAD: A Large-Scale Noise-Augmented Dataset for Voice Activity Detection

LibriVAD is a large-scale, noise-augmented dataset for Voice Activity Detection (VAD) generated from the LibriSpeech corpus. The dataset was created by LibriVAD and was last updated on March 17, 2026. It is designed for training and evaluating VAD models in noisy environments.

Use Cases

• Training Voice Activity Detection models based on the noise-augmented audio described.
• Benchmarking VAD algorithm performance in noisy conditions based on the dataset's stated purpose.
• Improving speech processing pipeline robustness using the large-scale, augmented audio data.

Strengths

• Dataset is based on the established LibriSpeech corpus, suggesting a foundation of clean speech data.
• Explicitly designed for Voice Activity Detection, indicating a clear task focus.
• Described as 'large-scale' and 'noise-augmented', which are relevant features for model training.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

LibriVAD

Collection Method

Generated from the LibriSpeech corpus using scripts for noise augmentation.

Freshness

Last updated 2026-03-17 10:29:07; freshness should be verified.