Back Home Next  

| Japanese | English |

MEASUREMENT OF JAPANESE SYLLABLES /A, KA, SA, TA, NA, HA, MA, YA, RA, WA/
(Analysis of the Japanese voice 6)

 
Date: 12:00, 30 Sep. 2002
Place: Nagoya, Japan
Microphone: SONY ECM-MS957
Amplifier: SONY DAT WALKMAN TCD-D100
Personal computer: DELL INSPIRON 7500
OS: Windows 2000 Professional
Software: DSSF3
WAVE sound file: voice6.wav (44.1kHz / Stereo / 16.25sec / 2.73MB)

Ten Japanese syllables "a, ka, sa, ta, na, ha, ma, ya, ra, wa" were measured. The integration time and the running step were set as 0.01 and 0.005 sec.

In the combination of a consonant and a vowel, a frequency characteristic at certain moment is not meaningful. The power spectrum is not displayed. Only the autocorrelation is measured later.

This is the calculation condition window of SA.

In the following, only the ACF in which the te value becomes minimum is considered. It is because the most information is included at this time.

This is the time change of the sound level. Ten utterances can be seen.

Result of "a" is not shown here. See the previous report.

 
"Ka" 2.56 second after measurement start, 10msec after utterance start
t
1=1.86msec / f1=0.46 / te=6.26msec / F(0)(SPL)=-20.59dB

At 10 ms after the utterance starts, the low frequency components are not analyzed yet. Most information is included at this time. That is the reason why this point is chosen for the analysis.

 
"Sa" 4 second after measurement start, 60 msec after utterance start
t
1=0.09msec / f1=0.64 / te=2.48msec / ƒ³(0) (SPL)=-35.09dB

 
"Ta" 5.595 second after measurement start, 10 msec after utterance start
t
1=1.93msec / f1=0.43 / te=5.07msec / ƒ³(0) (SPL)=-23.18dB

 
"Na" 7.13 second after measurement start, 25 msec after utterance start
t
1=4.35msec / f1=0.98 / te=21.94msec / ƒ³(0) (SPL)=-28.25dB

 
"Ha" 8.665 second after measurement start, 40 msec after utterance start
t
1=0.7msec / f1=0.16 / te=3.22msec / ƒ³(0) (SPL)=-39.14dB

 
"Ma" 10.135 second after measurement start, 15msec after utterance start
t
1=4.08msec / f1=1 / te=8.86msec / ƒ³(0) (SPL)=-36dB

 
"Ya" 11.645 second after measurement start, 15msec after utterance start
t
1=0.11msec / f1=1 / te=10.09msec / ƒ³(0) (SPL)=-36dB

 
"Ra" 13.13 second after measurement start, 20msec after utterance start
t
1=2.59msec / f1=0.18 / te=7.29msec / ƒ³(0) (SPL)=-28.68dB

 
"Wa" 14.68 second after measurement start, 5msec after utterance start
t
1=0.25msec / f1=0.69 / te=7.42msec / ƒ³(0) (SPL)=-55.12dB

The figure was made by the screen copy function of SA. The ACF is measured for the utterance of "wa".

In the figure below, time change of four parameters (SPL, t1, f1, WIACC) are shown.

In the following table, the ACF parameters for ten syllables are summarized with the previously measured data.
  time(ms) utterance(ms) t1(ms) f1 te(ms)
a tune do 0.625 35 1.13 0.27 6.80
a tune re 2.040 10 1.22 0.32 4.30
a tune mi 3.575 20 1.20 0.32 6.27
a tune fa 5.075 20 1.22 0.38 4.94
a tune so 6.605 25 1.20 0.66 7.76
a tune ra 8.140 5 1.18 0.49 6.45
a tune si 9.650 10 1.30 0.30 4.91
a tune do 11.255 20 1.25 0.50 4.45
a vowel compare   5 1.18 0.44 3.98
i   10 0.27 0.97 6.78
u   10 2.18 1.00 5.05
e   15 0.29 0.50 15.48
o   15 1.18 0.35 21.80
a   25 1.20 0.45 8.46
Ka   10 1.86 0.46 6.26
Sa   60 0.09 0.64 2.48
Ta   10 1.93 0.43 5.07
Na   25 4.35 0.98 21.94
Ha   40 0.70 0.16 3.22
Ma   15 4.08 1.00 8.86
Ya   15 0.11 1.00 10.09
Ra   20 2.59 0.18 7.29
Wa   5 0.25 0.69 7.42

September 2002 by Masatsugu Sakurai


Back Home Next