IndicTTS Bengali: High-Quality Speech Recordings for Text-to-Speech Research

Bengali Indic TTS Dataset contains high-quality speech recordings with corresponding text transcriptions. It is derived from the Indic TTS Database project, specifically using Bengali monolingual recordings from four native speakers. The dataset was authored by Abdullah500 and last updated on 2026-03-28.

Use Cases

• Train text-to-speech models based on high-quality Bengali audio recordings.
• Develop speech synthesis systems based on recordings from multiple native speakers.
• Benchmark TTS model performance based on the described audio format and sampling rate.
• Create voice cloning or voice conversion systems based on the male and female speaker data.

Strengths

• Contains high-quality speech recordings as described.
• Includes recordings from 4 native Bengali speakers (2 male, 2 female).
• Audio is in WAV format with a 48kHz sampling rate.

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

Indic TTS Database project

Collection Method

Derived from Bengali monolingual recordings.

Freshness

Last updated 2026-03-28 03:26:57; freshness should be verified.