Development Of Hardware and Software Complex "Dummy Head" with the Use of Non-professional Equipment

Abstract

This paper presents the results of research on the fundamental possibility of creating a hardware and software complex for acoustic examination of premises using non-professional equipment. The "Dummy Head" complex is designed to measure the characteristics of the binaural impulse and assess the intelligibility of two-channel speech. The peculiarity of this complex is that it contains electro-acoustic equipment of various qualities, including household appliances with mediocre quality characteristics.  

The paper shows that for such a creation it is necessary to adjust the characteristics of the measuring path taking into account the non-uniformity of the amplitude-frequency characteristics of the subsystem "speaker-microphone". Since  the amplitude-frequency characteristic of the speaker-microphone subsystem takes the smallest values at the edges of  the frequency range, and the variance of the mutual spectrum estimation of the speaker-microphone system is the largest at the right edge of the frequency range, it is advisable to use the regularization method to achieve the required calculation accuracy. Such a correction can be performed by controlled division of the frequency response of the speaker-room-microphone system to the previously obtained estimate of the amplitude-frequency characteristic of the speaker-microphone sub system. The problem of such calculations is the distribution operation, because the amplitude-frequency response of  the speaker-microphone subsystem may contain small numerical values that will overflow the bit grid of the computer system and will fall on the computer. However, it is obvious that if to ensure proper control of the amplitude-frequency characteristics of the speaker-microphone subsystem, such a distribution can be practically implemented. 

The paper also shows that the results of language intelligibility assessment in two small and medium-sized classrooms showed that speech intelligibility in the middle of the room may be less than near the wall of the room. These results are consistent with the results of previous studies, where language intelligibility was assessed by objective and subjective methods for other audiences. It should be noted that although estimates of the widely used C50 coefficient are consistent with estimates of speech intelligibility, the phenomenon of increased speech intelligibility near the walls of the room is more pronounced when using estimates of speech intelligibility. Using a two-channel measurement system configured in this way, experimental studies were conducted to assess the intelligibility of speech distorted by reverberation.