Japanese speech database
About this dataset
1. Product Overview
The Japanese speech database contains a total of 295,083 entries, spanning 245 hours. Recordings were made in Japan, with speakers including both adults and children. The corpus consists of everyday conversational expressions, and the audio files are in 16 kHz, 16-bit WAV format.
2. Product Directory Structure
Directory Structure
Data directory structure
└─DOC (text folder)
│├─content.txt (text file)
└─SPEECHDATA (data folder)
│├─JA (block number)
││ ││ JA000001.wav (audio file)
Naming Conventions
CORPUS/USAGE/BLOCK_ID/SPEECH_IDe.g./SPEECHDATA/JA/JA000001.wav
Directory Name · Content · Notes
CORPUS · · Speech database ID
USAGE · SPEECHDATA · Audio file folder
BLOCK_ID · JA · Block number
SPEECH_ID · 0000001.wav · WAV file
Gender · Number of entries · Percentage
Adult Male · 89,723 · 30.4%
Adult Female · 113,066 · 38.3%
Children · 92,294 · 31.3%
Total · 295,083 · 100%
3. Speaker Information
The speakers include both adults and children, with the following breakdown:
Download Sample Data
Samples
- Audio sample 1
- Audio sample 2
- Audio sample 3