THE JOURNAL OF THE ACOUSTICAL SOCIETY OF JAPAN

Automatic Recognition of Nasal Consonants

Minoru Shigenaga, Hitoshi Ariizumi

Article type: Article
1965 Volume 21 Issue 5 Pages 263-271
Published: September 30, 1965
Released on J-STAGE: June 02, 2017

DOIhttps://doi.org/10.20697/jasj.21.5_263

JOURNAL FREE ACCESS

Show abstractHide abstract

A method of recognizing /m/ and /n/ of monosyllables and words in real time is reported. The method consists of three parts; segmentation of nasal consonants, recognition of the following vowel, and discrimination between /m/ and /n/ according to the result of vowel recognition. In order to extract the nasal part, /e, a, o, u, w/ are excluded from nasals by comparing the output of 300 c/s LPF with that of 500〜1600 c/s BPF, and /i, j/ are excluded by comparing 500〜1600 c/s with 2800〜5000 c/s. Voiceless stops are easily omitted by comparing the output of 300 c/s LPF with that of 700 c/s HPF, and this circuit is also used for excluding vowels. In order to exclude the beginning and end part of /u/ more certainly, the output wave form of 400〜1000 c/s BPF is used. The parts in which the envelope of that wave varies rapidly are excluded from the nasal part. For excluding voiced stops, fricatives, and flappeds, fundamental frquency components are extracted from filtering output of 400〜1000 c/s BPF and the parts in which the fundamental frequency components exist continuously are considered to be likely nasal. The low intensity parts of original speech waves are considered to be non-nasal parts. The segment which satisfies above six conditions is decided to be nasal consonant after leaving out 12 msec of the continuum. According to this method the initial part of the nasal consonant is often missed but the boundary between the nasal consonant and the following vowel is pointed out exactly. Concerning about the discrimination between /m/ and /n/, the components of two specific frequencies are compared just before the boundary between the nasal consonant and the following vowel for distinguishing /mi/ and /me/ from /ni/ and /ne/ respectively. For the nasals followed by /a, o, u/, F_2 loci are used directly. Though /me/ and /ne/ in words are not discriminated satisfactorily by the method due to the individual variations, the others are recognized correctly over 80% for 180 samples of three male speakers.

View full abstract

Download PDF (823K)
On the Relationship Between the Absorption Coefficient and Flow Resistance of the Cloths

Otoichi Kitamura, Minoru Sasaki

Article type: Article
1965 Volume 21 Issue 5 Pages 272-280
Published: September 30, 1965
Released on J-STAGE: June 02, 2017

DOIhttps://doi.org/10.20697/jasj.21.5_272

JOURNAL FREE ACCESS

Show abstractHide abstract

The normal incident absorption coefficient α_0 of many cloths placed at a certain distance from a rigid wall was measured. The peak value of α_0 due to energy loss by viscous resistance of air at the distances of 1/4λ and 3/4λ from the rigid wall was obtained and the flow resistance R_f of the cloths was measured. The relationship between the peak value of α_0 and the flow resistance R_f was investigated experimentally and theoretically. The following results were obtained. 1) The samples are classed into three groups in relation to absorption mechanisms; viscous resistance type, resonance type, and intermediate type between the viscous resistance type and the resonance type. 2) In case that R_f of a cloth is in the range between 10 and 40 rayls, α_0 increases with R_f, and when R_f is larger than 50 rayls, α_0 decreases with increasing R_f. When R_f of a cloth is 40〜50 rayls, α_0 therefore shows the maximum value (about 100%). 3) Theoretically, in case that R_f of a cloth is equall to 42 rayls, the peak value of α_0 becomes 100% under the assumption of modelized cylindrical narrow pores distributed homogeneously.

View full abstract

Download PDF (1071K)
Statistical Studies on the Level Fluctuation of Random Noise (I) : Explicit Representations of the Probability Density Distribution of Random Noise

Mitsuo Ota

Article type: Article
1965 Volume 21 Issue 5 Pages 281-294
Published: September 30, 1965
Released on J-STAGE: June 02, 2017

DOIhttps://doi.org/10.20697/jasj.21.5_281

JOURNAL FREE ACCESS

Show abstractHide abstract

We are well aware of the fact that in almost all of the random noise measurement we face the measurement of frequency distribution of level fluctuation of random noise. More explicitly, when the instantaneous readings of a sound level meter are recorded at every five seconds (df. JIS Z 8731-1957), infinitely many different types of frequency distribution of random noise fluctuation such as city noise are experimentally found. However, the statistical methods of analysing unificatively these experimental results of frequency distribution have not yet been introduced. In previous papers, a joint characteristic function in the form of Hankel transform applicable as a postern to probability problem of many correlative physical quantities fluctuating only in positive region was proposed. Then it was demonstrated that this function is more effective for the analysis of the random noise current than the characteristics function of Fourier or Laplace type which has been used so far. In this paper, the problem of what the fundamental and universal property will become important when we try to obtain the general explicit expression of probability density distribution of random noise. First, the relation between a general statistical treatment of the frequency distribution of random noise such as city noise and the probability method using the characteristics function of Hankel type is cleared up. Then, a special case of interest with a physical quantity as mentioned above is treated, and explicit expressions of the probability density distribution of random noise in the forms of statistical Laguere expansion series or other expansion series are presented under the statistical method using the characteristic function of Hankel type. We must call our attention to the fact that the individual character in the statistical property of each random noise is reflected in two parameters and each expansion coefficient of the probability distribution. Further, from the above point of view, it has been pointed out that a lognormal probability distribution can be approximately derived as a universal expression of the probability distribution of random noise. Finally, detailed experimental considerations of city noise enough to corroborate the above theories are give in the following three cases: (a) probability expression in the form of statistical Laguerre expansion series, (b) an approximation to the gamma distribution, (c) a universal approximation to the lognormal distribution. The statistical method described in this paper is also applicable to other wide fields of measurement on random phenomena, since the probability variables defined only in positive region are fundamental.

View full abstract

Download PDF (1408K)
[title in Japanese]

Article type: Appendix
1965 Volume 21 Issue 5 Pages 298-
Published: September 30, 1965
Released on J-STAGE: June 02, 2017

DOIhttps://doi.org/10.20697/jasj.21.5_298

JOURNAL FREE ACCESS

Download PDF (62K)
[title in Japanese]

[in Japanese]

Article type: Article
1965 Volume 21 Issue 5 Pages 299-302
Published: September 30, 1965
Released on J-STAGE: June 02, 2017

DOIhttps://doi.org/10.20697/jasj.21.5_299

JOURNAL FREE ACCESS

Download PDF (493K)
[title in Japanese]

[in Japanese]

Article type: Article
1965 Volume 21 Issue 5 Pages 302-304
Published: September 30, 1965
Released on J-STAGE: June 02, 2017

DOIhttps://doi.org/10.20697/jasj.21.5_302

JOURNAL FREE ACCESS

Download PDF (432K)
[title in Japanese]

[in Japanese]

Article type: Article
1965 Volume 21 Issue 5 Pages 305-308
Published: September 30, 1965
Released on J-STAGE: June 02, 2017

DOIhttps://doi.org/10.20697/jasj.21.5_305

JOURNAL FREE ACCESS

Download PDF (473K)

Register with J-STAGE for free!