This paper deals with the effect of linear phase variation on total acoustic power generated from an axial-flow turbomachinery. In a previous paper (J. Acoust. Soc. Jpn.(E) 2, 203-211), the authors derived a simple expression on the power when all the fluctuating forces on rotor (or stator) blades are in phase. The expression is extended to the case where the phase of the fluctuations varies linearly with the circumferential distance. This is a model for turbomachines with differentn umber of stator and rotor blades. It is shown that the Fourier components of the fluctuation contributing to the power are related to the phase variation and that the dependence of some parameters on the power has the same form as that where the fluctuationsa re in phase. Furthermore, it is suggested that the noise with discrete frequencies is emitted mainly from rotor blades not from stator ones. To obtain a further understanding an estimation of the fluctuationa re made using a lineart heory and important effectos f blade geometry and flow configuration are pointed out.

View full abstract

This paper describes the recognition system of Korean digits spoken in isolation. Especially investigated are the bandwidth and the center frequency allocation of filters for extracting the spectral envelope from the power spectrum obtained by fast Fourier transform (FFT). This experiment is carried out as a preliminary investigationo f the recognition of Korean speech. The filterasr e realizedb y multiplying the rectangular frequency windows by the spectra obtained by FFT. Regarding the bandwidth of band pass filterst, h ree types of filteras re compared. The firsto ne is of the constant frequency bandwidth in the linear frequency scale. The second one is of the constant bandwidth in the linear-logarithmicf requency scale (lineari n the range lower than about 1, 000Hz and logarithmic in the higher frequency range). The third one is of the constant bandwidth in the logarithmic frequency scale. Experiments are carried out on both speaker dependent and speaker independent recognition of 10 Korean digits uttered by five males and five females (1, 000 test samples). On the speaker dependent mode a 99.3% correct recognition rate is obtained using 9 channel filtersw hose bandwidths are constant in the logarithmic frequency scale. The same rate is obtained using 12 channel filterwsh ose bandwidths are constant in the linear-logarithmifcr equency scale.

View full abstract

It is well-known that the levelf luctuationsa ppearing in the road trafficn oise environment are brought on by various type causes and can be roughly classifieidn to two great divisionso n the road trafficf low pattern and the surrounding sound field. Accordingly, if the equivalent models for the above two divisions-traffifcl ow and its surrounding fielda re well establishedi n a flexiblef orm, of reflectintgh e complex situationo f actual road trafficn oise field, i t will be possible to predict preciselya nd/or to control effectivelyt he road trafficn oise. In this paper, from the above point of view, a stochastic model of road trafficn oise with a generalized P oisson type flow is firsit ntroduced with many informations of actual situation. Next, an equivalent type sound propagation model is newly proposed in a form of mixing of two idealizeds ound propagation environments-a free sound field and a diffused sound field. Furthermore, a methodological trialt o estimate the above equivalent sound propagation characteristicosf the compound environment by measuring the actual road trafficn oise data is derived from the system-theoreticaplo int of view. In thisc ase of estimation, t he well-known stochastic approximation method is used in order to overcome the difficuitoyf non-Gaussian and non-linear characteristicesx istingi n the actual environmental noise phenomena. Finally, the legitimacy of this estimation method is confirmed by applying it to the trafficn oise data actuallyo bserved in a big city.

View full abstract

This paper covers the experiment on the Japanese syllable recognition system with the variable duration dictionary by using a 16 bits microcomputer (MN1613). We reported a syllable recognition experiment in the previous paper (Transactions of the Commitee on Speech Research, Acoust. Soc. Jpn., 137-144, S82-18, 1982) by using super minicomputer VAX-11/780. Correct recognition rate 94% was obtained and the validity of the variable duration dictionary algorithm was confirmed. Some improvement in the algorithm has been done to adopt the algorithm to the microcomputer. Ninetyfive percents correct recognition rate is obtained on the microcomputer system. Segmentation technique is newly employed and the duration, from -51.2ms to +51.2ms for vowels and voicelesss top, -102.4ms to +51.2ms for voiced consonants, -51.2ms to 102.4ms for semivowels are chosen in the variable duration dictionary. Moreover we proposae techniqfuoer s egmentaticon sidertihnegb andpowedri fferebnecteween malea nd femalsep eakerasn da lsop roposae new algoritthom d iscrimin/astiet/ o /ci/and /su/t o /tsu/N.i nety-fipever centcso rrecrte cognitiraont ef or1 0 speakerhsas beeno btainewdi tht he6 8 Japanesmeo no-syllablEeasc.h syllabilses poken1 0 or2 0 timefs orb othm alea ndf emalsep eakeirnst hiesx periment.

View full abstract