Japanese Short Sentence Speech Database

Feature

This product has been developed to meet the needs of our overseas clients who wish to test the capabilities of their Japanese-language speech codec algorithms or telecommunication speech terminals. Samples in this database have been used in subjective standardization tests for the ITU-T 16 kbit/s wideband codec and 4kbit/s speech codec algorithms. Merging the database files enables users to make speech samples comprising sentence pairs and time structure as given in ITU-T Recommendation P.800 and the subjective test plan of 3GPP.

Price

385,000 JP Yen for four CD disks, each containing utterances of 100 male or 100 female speakers.

220,000 JP Yen for two CD disks, each containing utterances of 50 male or 50 female speakers.

Please note: When clients order CD sets of both genders, the total price is double the above price.
Note: A 5% consumption tax will be added for shipments within Japan.

Page Top

Specification

Recording

All speech samples are recorded using digital equipment, with all recording complying with ITU-T Recommendation P.800. All recording is done using a high-quality, low-noise condenser microphone for measuring sound. Puff noise is prevented through a screening process.

Speaker

A total of 100 male or 100 female native Japanese speakers, ranging from 15 to 65 years in age, are employed. They are given no special instructions regarding utterances, and their native accent is preserved unless it results in readings or lexical meanings that are incorrect.

Sentences

Each speaker utters 200 short Japanese sentences. The speaker set for each sentence differs slightly due to the removal, in some cases, of speakers with incorrect utterances or extraneous noise.

Media

All speech samples are recorded on ISO 9660-formatted CD disks. Speech data are sampled at 16-bit and 16kHz rates and formed into raw (binary) data files.* Speech samples are digitized at 16-bit and 16kHz rates. All files are stored in ISO9660-formatted CD-ROMs.
The active power level of each sample is normalized to-26 dBov according to the ITU-T Rec.P.56 algorithm.

*: Files can be changed to Windows WAV format using included software under the DOS window. A byte change is necessary to apply them to the UNIX system.