UrduSpeech: Multi-Style Speech Corpus for Urdu and Kashmiri

UrduSpeech is a high-quality multi-style speech corpus containing approximately 51,600 audio-text pairs for Urdu and Kashmiri languages. It was created by humairawan and last updated on December 11, 2025. The dataset includes professionally recorded audio with diverse speaking styles, emotional expressions, and gender representation.

Use Cases

• Training text-to-speech models based on professionally recorded audio.
• Building automatic speech recognition systems based on audio-text pairs.
• Developing expressive speech synthesis models based on diverse speaking styles and emotional expressions.
• Researching speech patterns and gender representation in Urdu and Kashmiri.

Strengths

• Approximately 51,600 audio-text pairs provide a substantial corpus.
• Contains professionally recorded audio, suggesting high-quality source material.
• Includes diverse speaking styles, emotional expressions, and gender representation.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

huggingface

Collection Method

Professionally recorded audio.

Freshness

Last updated 2025-12-11 20:02:18; freshness should be verified.