Bengali Noisy and Clean Paired Audio for Speech Enhancement

A Bengali language dataset containing paired noisy and clean audio files. The data is described as ready for machine learning tasks. The dataset was sourced from Kaggle, but details on its creator, size, and update date are unavailable.

Use Cases

• Train speech denoising models based on the provided noisy-clean audio pairs.
• Benchmark audio enhancement algorithms for Bengali speech.
• Develop noise classification systems based on the types of noise present in the audio.
• Create synthetic training data for robust Bengali speech recognition systems.

Strengths

• Dataset provides aligned noisy and clean audio pairs, which are a standard format for supervised speech enhancement.
• Data is described as ready for machine learning, suggesting it may be pre-processed.

Limitations

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

Provenance

Source

Kaggle

Collection Method

Method of data gathering is unknown.

Time Range

Temporal coverage is unknown.

Freshness

Last update date is unknown; freshness unverified.

Geography

Data is in the Bengali language, suggesting a likely geographic focus on Bangladesh or parts of India.