AES 東京コンベンション 2009
Poster Session P3

A method on multi-channel sound synthesis is proposed for creating virtual acoustic environment in computer graphics productions. Impulse responses with spatial information are generated from virtual sound source distribution. For desired reproduction channel setups, total impulse response is decomposed into channel components, and each response is convolved with source signal. The method is easily adaptable when location and direction of observation point in virtual graphical scene switches frequently.

In this paper, we proposed a new method using HRTFs that contain room reverberations(R-HRTF). The reverberation is not added to the dry sound source separated with HRTF but contained at their measured process in the HRTFs. We measured the HRTFs in a real reverberant environment for directions of azimuth 45, 90, 135 (left side) and elevation from 0 to 90 (step of 10 degrees) degrees then constructed a 3D sound system with the measured R-HRTF with headphones, examine if the sound reality is improved. As a result, we succeed to create 3D spatial sound system with more reality compared with traditional HRTFs sound system.

Artificial reverberation is often used to increase reality and prevent the in-the-head localization in a headphone based 3-D sound system. In traditional method, monaural reverberations were used. In this research, we measured impulse responses of an ordinary room by Four Point Microphone method, and calculated the sound intensity vectors by the Sound Intensity method. From the sound intensity vectors, we obtained the image sound sources. A binaural reverberation was reconstructed by the estimated image sound sources. Comparison experiments were conducted for 3 kinds of reverberations, i.e., monaural reverberation, binaural reverberation and binaural reverberation added with Head-Related Transfer Function. From the results, we could clarify the 3-D sounds reconstructed by binaural reverberation with head-related transfer function has the best spatiality.

We investigated the discrimination of sounds that have interaural correlation coefficient (ICC) of low frequency sound below around 100 Hz by using 1/4 octave band noise in dichotic listening. We also measured ICCs at low frequencies in several rooms with different volumes. The ICCs were obtained from an impulse response recorded by a head and torso simulator. At reference ICC 1.0 the lower threshold is 0.98; at 0.9 it is 0.85, and the upper threshold is 0.96; and at 0.8 the lower one is 0.68, and the upper is 0.88. The equivalent subjective diffuseness for the 1/4 octave band noise with the center frequency of 106 Hz and ICC of 0.9, are obtained by ICC of 0.71 at 250 Hz and ICC of 0.34 at 500 Hz. The measured ICC in a small room with a volume of less than 100 m3 is almost 1, and the ICC in large room with a volume of more than 1700 m3 is about 0.8.

Simplification of head-related transfer functions (HRTFs) is important for effective implementation of their synthesis from computational point of view. It can be found from the frequency resolution of the auditory system that the detailed spectral form of the HRTFs is not evaluated at high frequency region. This may enable the simplification of the HRTFs to some extent. In this paper, the HRTF on the contralateral side was flattened in the higher frequency region than a certain frequency so as to retain the interaural level difference and interaural time difference. To evaluate the influence of simplified HRTFs, a localization test was carried out. The experimental results showed that HRTFs on the contralateral side could be simplified above 4 kHz.

Previously it was reported that the lowest-frequency spectral notch (first notch) on head-related transfer function (HRTF) play a role in sound image localization; high-frequency spectral notches (called N1 and N2) are important for sound image localization in elevation. However, another spectral notch (labeled N0) at relatively lower frequency appear in the same frequency range for sound sources behind a listener. In this study, we examined the spatial distribution of such low-frequency notches based on several subjects' measured HRTFs, and subsequently investigated whether such spectral notches can contribute to sound localization. The results showed that, in some cases, N0 exists at lower frequency than N1, indicating that a N0 provides localization cues for sound sources behind a listener.

This is a report on our challenge and achievement of creating unique performance specifications and sound quality through adoption of smooth curved surface for overall dynamic speaker system design. Enclosure is shaped like an egg and speaker diaphragm is curved to match the overall enclosure form. In order to achieve smooth and even sound radiation without any disturbance, speaker edge surround and frame are not visible from outside and there is no speaker grill. As a result, we were able to develop a speaker system with very high S/N ratio across wide frequency range, allowing reproduction of subtle nuances, and with ability to create superb sound stage. Furthermore, a large number of favorable attributes, in addition to wide dispersion and smooth sound radiation, could be observed.

The 22.2 multichannel sound system is a high-presence sound format that is superior to the 5.1 surround sound system, and NHK intends it to be the format for the next generation of TV broadcasting. NHK started to make Super Hi-Vision (SHV) content for experiments and exhibitions in 2002 and exhibited the 22.2 multichannel sound system as part of the SHV system for the first time at the Aichi World Expo in 2005. This system has also been demonstrated at the National Association of Broadcasters (NAB) convention and the International Broadcasting Convention (IBC). This paper reports on the 22.2 system's method of recording sound fields and the technique of achieving ``3D stereophony'' in post-production for the program "Gift" that was exhibited at NAB2009.

The public viewing of ``Kouhaku Uta Gassenn'', one of the most famous music programs in Japan, was held at NHK Fureai Hall in Tokyo on December 31, 2008. The program was mixed live using the 22.2 multichannel system, which is highly expected to be a potential audio format for future TV broadcasting audio format. This paper introduces an overview of this public viewing, including the 22.2 multichannel sound system flow and live three-dimensional sound mixing technique.