High-Fidelity Conversational Speech in Four Asian Languages

Featuring high-quality conversational audio samples for Automatic Speech Recognition tasks in Vietnamese, Korean, Arabic, and Filipino. It includes paired audio and transcripts of natural, non-scripted speech, featuring both single-speaker and dual-speaker interactions. Audio specifications include a sampling rate of 16 kHz to 24 kHz and a 16-bit bit depth.

Use Cases

• Train an Automatic Speech Recognition model on paired audio and transcripts for the four supported languages.
• Fine-tune a speech model on non-scripted conversational speech to improve performance on natural dialogue.
• Analyze acoustic features and transcription accuracy differences between single-speaker and dual-speaker interactions.
• Benchmark ASR system performance across the Vietnamese, Korean, Arabic, and Filipino languages using high-fidelity audio.

Strengths

• Includes audio samples in four distinct languages: Vietnamese, Korean, Arabic, and Filipino.
• Audio is high-fidelity with a sampling rate of 16 kHz to 24 kHz and a 16-bit bit depth.
• Contains both single-speaker and dual-speaker conversational interactions.
• Features natural, non-scripted conversational speech paired with transcripts.

Limitations

• The total number of audio samples, dataset size, and specific row count are unknown.
• Specific audio file formats, column structure, and detailed metadata are not provided.
• Geographic origin and temporal coverage of the speech samples are unspecified.

Provenance

Source

humyn-labs on Hugging Face.

Collection Method

Curated collection of conversational audio samples.

Freshness

Last updated on 2026-03-13.