Shigeaki Aoki

Sound Image Rendering System For Headphones

We have developed a sound image rendering system for headphones that enables anyone to easily create realistic 3-D sound. This paper presents the structure of this system, and the results of subjective tests confirm the systems feasibility in affecting sound. This system has been used in the making of music CDs.

Channel-number-compressed multi-channel acoustic echo canceller for high-presence teleconferencing system with large display

Sound localization is important to make conversation easy between local and remote sites in a teleconference. This requires a multi-channel sound system having a multi-channel acoustic echo canceller (MAEC). The appropriate number of channels is determined from a trade-off between high presence and MAEC performance, so it is not possible to increase the channel number by much. We propose a channel-number-compressed MAEC to provide teleconferencing systems that exhibit high presence. The channel number of the MAEC inputs is compressed and that of its outputs is expanded.

Expansion of listening area with good localization in audio conferencing

Audio conferencing quality can be improved by introducing the localization effect for participants at both sites. This paper proposes a new method which can expand the listening area with good localization for each participant in audio conferencing. The method can suppress interference between sounds radiated to different listening areas by using the localization-in-head and precedence effects and the directivity of a doublet source. The proposed stereo reproduction system employs three pairs of loudspeakers placed on a rectangular table. The effectiveness of the system is confirmed by subjective testing and its performance with various recording methods is also investigated. It can provide good localization not only at the center of the listening area, but also at its sides. In terms of recording method, using a center-placed microphone with two spaced microphones was found to contribute to good localization.

An escape guiding system utilizing the precedence effect for evacuation signal

This research aims at building an escape guiding system using audio signals. The system utilizes the precedence effect for a listener to perceive easily the direction of an emergency exit. In one of the ordinary escape guiding systems, two or more loudspeakers are set on the ceiling of a passageway. Since the precedence effect is generated by delaying suitably the audio guidance signal radiated from the loudspeakers, the proper direction of the emergency exit is recognized. However the effective listening area is limited. That is, the ordinary escape guiding systems is not effective in areas under loudspeakers. In this paper, a new method that the loudspeakers are set beside in a passageway is proposed. The generation and disappearance of the precedence effect of audio signals in the new configuration are investigated. In the listening tests, the time delay and the difference of intensity between loudspeakers are parameters. The installation angle of the loudspeaker is another parameter. Male voice and female voices are used as emergency guidance. The guidance effect of the configuration in setting beside the loudspeaker is confirmed and the test results are discussed.

Audio teleconferencing systems with sound localization effects

In teleconferencing, conversation should flow naturally between separate sites. The acoustic design should, of course, be able to at least convey the correct volume. However, sound localization is also very important for realizing the kind of virtual reality world that is teleconferencing. Coordinating the visual image (a speaker’s face) and the sound image (his or her voice) creates a kind of virtual environment in which the participants at each site feel as if they are having a conversation in the same room. Two prototype teleconferencing systems with sound localization effects have been developed. The more primitive of the two is a small teleconferencing system that can connect two sites with stereo reproduction in a way that utilizes psychological acoustic phenomena. The other is a high‐presence system that aims at achieving a high level of ambience among three separate sites. Its features and key technologies are natural images on a large super‐high‐resolution display, good sound localization from remote locations through the use of a multichannel sound image localization system, and high‐definition images as a result of HDTV coding and transmission. The proposed multichannel sound system comprises four subsystems, one each for sound input, reproduction, acoustic feedback control, and transmission.

Analysis of vertical sound image control with parametric loudspeakers

A parametric loudspeaker is known as a super directional loudspeaker by using nonlinear interaction between high power ultrasounds in air. The parametric loudspeaker is one of the prominent applications of nonlinear acoustics. So far, we had conducted the listening test that the parametric loudspeaker reproduced left and right signals with binaural information. The characteristics of the sound localization with the upper and lower parametric loudspeakers in the vertical direction were investigated by the results of listening tests. The vertical direction of sound localization was able to be controlled when the acoustical axis was set to the right ear. The results were similar as in using ordinary loudspeakers. It was found that the left-right sound localization could be realized only with the upper and lower parametric loudspeakers. By setting the parametric loudspeaker the right ear, that is by setting it only 3 degrees rightward, the direction of sound localization moved about 10 degrees rightward. In o...

Sound source segregation using the inter‐channel differences in both intensity and phase

A method is proposed for segregating sounds from multiple sources recorded on a two‐channel system. It uses the differences in both intensity and phase, between the channels. In addition, the relation between frequency resolution and sound quality of the proposed method is examined. In the tests, three pairs of mixed sounds were used, namely, male speech with female speech, two kinds of female speech (the same person with different phrases), and a cock’s call with female speech. Frequency resolution was varied between 5 and 80 Hz. In the subjective test, subjects listened to the original sound plus sounds segregated at five different resolutions: 5, 10, 20, 40, and 80 Hz. Then they estimated the quality of these sounds on a five‐point scale. Moreover, the correlation coefficient between the original sound and each segregated sound is discussed. In the case of a human voice, it is determined that the optimal frequency resolution is 10 or 20 Hz. The interesting fact is that the optimal frequency resolution ...

Effect of the level difference between leading and trailing stimuli on pitch, loudness, and binaural perception

The total perception for pitch and loudness as a function of level difference between the leading and trailing stimuli is investigated using the same stimulus configuration and measuring the paradigm which clearly showed that the dominance of the leading stimulus depends on the level difference [S. Aoki and T. Houtgast, Hear. Res. (1992)]. A brief stimulus of 20 ms is subdivided into two parts with durations T1 and T2. Each part consists of a sinusoidal signal of either 1 or 2 kHz for pitch perception, or each part consists of a 2‐kHz sinusoidal signal with either + or −3 dB for loudness perception. The measuring paradigm aims at assessing the critical ratio T1/T2 for which both parts contribute equally to the overall sensation of pitch or loudness for these brief stimuli. In contrast to the earlier study on dichotic cues, the effect of the level difference between the leading and trailing stimuli on pitch perception is found to be smaller and discontinuous. The effect on loudness perception is found to b...

The continuity of auditory apparent motion for amplitude‐modulated broadband noise

The thresholds of the switching period Tth’s that produced the continuous or broken auditory apparent motion were measured, when the stimuli were successively presented by using three loudspeakers arranged in the horizontal plane. In the current experiment, the Tth’s were investigated for amplitude‐modulated broadband noise as a function of the modulation frequency fm (ranging between 4–64 Hz) and the modulation index m (ranging between 0.0–1.0). In agreement with the previous study that employed signals including bandpass noise, speech, etc. as stimuli [I. Kinoshita and S. Aoki, Tech. Rep. Hear. Acoust. Soc. Jpn. H93‐2, 1–8 (1993)], the Tth’s were diversified and generally larger than the one for an unmodulated (m=0.0) broadband noise; a monotonical increase in the Tth was indicated with the modulation index (e.g., the Tth ranged between 72–290 ms at a modulation frequency of 16 Hz), while the Tth was reduced as the modulation frequency increased (e.g., the Tth decreased from 360 to 135 ms at a modulatio...

Offset dominance in pitch perception

Possible effects of onset or offset dominance for pitch information are investigated with the same experimental approach, which clearly showed the familiar onset dominance in the perception of binaural cues [Houtgast and Aoki, J. Acoust. Soc. Am. Suppl. 1 87, S64 (1990)]. A brief stimulus, with a duration of 5, 10, 20, or 40 ms, is subdivided in two parts with durations T1 and T2, each part consisting of a sinusoidal signal of either 1 or 2 kHz. The measuring paradigm aims at assessing the critical ratio T1/T2 for which both parts contribute equally to the overall sensation of pitch for these brief stimuli. This critical ratio is found to be greater than one, indicating a dominance of the pitch information presented in the later part of the stimulus. The data can be summarized in a weighting function: the perceptual weight of the pitch information as a function of time before signal offset. This weighting function shows a peak of enhanced weight for the last few milliseconds just before stimulus offset, i...