All datasets

Korean Speech: Contributor Recordings

Korean Speech: Contributor Recordings

About

Korean speech recordings from the global contributor audio collection, with validated speech paired with written text. Samples are captured through a dedicated recording app with consistent audio formatting and transcript metadata.

Specifications

File types
wav
Audio format
wav
Audio codec
PCM
Sampling rate Hz
48,000
Bit depth
16
Channels
1
Bitrate kbps
768
Transcript source
Ground Truth
Collection source
Contributor recording app
Speaker format
Single-speaker recordings
Language code
ko

Sample details

Bank account issue

A single-speaker Korean recording describing a banking or account-support scenario.

Smart home setup

A single-speaker Korean recording about configuring a smart-home device or workflow.

Poseidon AI, Inc. © 2026
All rights reserved