The localization of a single acoustic source on the horizontal plane using phase difference spectrum images is discussed. The azimuth for source is identified from the general linear relationship, which is extracted from the measured phase difference spectrum after filtering. The phase difference spectrum is introduced as the quasi-stationary cross-spectral phase between the sound signals detected simultaneously by two sensors. Acoustic source localization in an anechoic chamber having a metal base plate using two types of sound signals, white noise and voice, indicated that the phase difference spectrum was not affected with respect to the sound pressure level but was affected with respect to the azimuth for source. Although the phase difference spectrum measured in a reverberant room had less continuity as a function of frequency, a linear distribution of the images obtained from the data (dots) was observed on the frequency - phase difference plane. Using the phase difference spectrum images, the azimuths for various sources, which radiated any kind of broadband sound on separated time schedule, were precisely identified even in the reverberant room.
We can communicate with others in a noisy environment. This phenomenon is known as a “Cocktail Party Effect” and is one of the most important binaural functions. This paper addresses a frequency domain binaural model that plays the role of a binaural function based on an interaural phase and level difference. The proposed model is evaluated not only as a front-end of the speech recognition system, but also as a speech enhancer. According to the evaluation, when the direction of arrival of the target signal and noise differs by 10°, recognition rates improve in comparison with the previous time domain binaural model (TDBM) in any cases. Furthermore, recognition rates show more than 90% when the signal to noise ratio (SNR) is higher than approximately 5 dB. On the other hand, SNR and coherence of the frequency domain binaural model, which is obtained for an evaluation of the speech enhancer, show superior results over the TDBM.
Two kinds of psychological experiments are conducted to reconfirm the effect of late sound from directions other than lateral on LEV, which was reported in our previous paper, and to clarify the degrees of contribution of directional late energy to listener envelopment (LEV). In the first experiment, the levels of late sound arriving from four directions, namely, lateral, frontal, overhead, and back, are independently varied. The results reconfirm that not only the lateral level, but also the levels of late sound from above and behind the listener affect LEV significantly. In the second experiment, directional late energy ratios are varied keeping the total level of late energy constant. The results indicate the degrees of contribution of lateral, overhead, and back late energy to LEV.
In this paper, we present the effectiveness of a software tool for eliminating nonlinear distortions of loudspeaker systems through some experiments. The nonlinear distortions affect the sound quality of loudspeaker systems. We have presented some identification methods of loudspeaker systems and some compensation methods of nonlinear distortions. However, the software tool for compensating the distortions has not been realized yet. We therefore develop the software tool. Experimental results show that the 2nd- and 3nd-order nonlinear distortions of a loudspeaker system can be reduced in the range of 10 [dB] to 20 [dB] by the developed tool when noise level is below 70 [dBA].