Hindi Speech Recordings From 401 Indian Speakers

Comprising 240 hours of Hindi speech recorded by 401 Indian speakers in both quiet and noisy environments. The recording content covers economic, entertainment, news, and spoken language topics, with all texts manually transcribed for high accuracy.

Use Cases

• Train automatic speech recognition models on 240 hours of manually transcribed Hindi audio.
• Develop noise-robust speech models using recordings from both quiet and noisy environments.
• Analyze speech patterns across diverse content domains like economic, entertainment, and news topics.

Strengths

• 240 hours of audio data provides a substantial corpus for model training.
• Recordings from 401 distinct Indian speakers offer speaker diversity.
• Manual transcription of all texts ensures high label accuracy for supervised learning.

Limitations

• The dataset is described as a sample of a larger paid dataset, so the full scope is unknown.
• No information is provided on audio quality specifications, sampling rates, or file formats.
• The geographic coverage is limited to Indian speakers, which may not generalize to other Hindi dialects.

Provenance

Source

Nexdata

Collection Method

Recorded by 401 Indian speakers reading text in both quiet and noisy environments.

Freshness

Last updated on 2025-04-23.

Geography

India