It is assumed SHV will be viewed on screens that are larger than today's Hi-Vision screens, so it will be important to ensure that sound and video are consistent in the vertical and horizontal directions. As such, we are conducting R&D on a 22.2 multichannel audio format (Figure 1) for use with SHV, which expands the earlier 5.1 channel sound system in the vertical direction and enhances the sense of being enveloped in sound.

Figure 1. Speaker arrangement for 22.2 channel audio system

Standardization

During FY2009, we continued to promote standardization efforts for codings and formats related to 22.2 multichannel sound. For coded transport, we revised the MPEG-2 AAC standard to accommodate 22.2 multichannel sound. 22.2 multichannel sound has also been adopted as one of the recommended audio modes for the ARIB STD-B32 standard. We also proposed a "3D Multichannel Studio Sound Format Standard" for use in studios to the ITU-R. At the IEC, proposals for channel naming and transmission procedures for multichannel audio systems have moved to the international voting stage. SMPTE has also issued a standard (SMPTE299-2009) for transmission of 96 kHz sampled audio. Finally, AES has issued a standard (AES11-2009) for synchronization of UHDTV video and digital audio.

22.2 Multichannel sound production systems

We are conducting R&D on a system to simplify and increase the sophistication of 22.2 multichannel sound production. In FY2009, we developed a mixing system able to control the directionality of up to 1000 individual tracks of audio, a one-point microphone making it easy to capture 22.2 multichannel sound in the field, a 22.2-multichannel-sound speaker system for use in production, and an automatic equalizer method for correcting the frequency characteristics for the speaker system (Figure 2). We also developed a new 3D panning method and verified its capabilities in subjective evaluation tests. As a result, we found that the sound localization produced by the system matches human hearing closely in the horizontal direction, but that if the opening in the speakers exceeds 30 in the vertical direction, it begins to diverge significantly from human hearing. Also, in live multi-channel SHV satellite relay tests using the Kizuna Wideband satellite, we conducted the first ever live relay of all channels of 22.2 multichannel sound captured in the field.

Figure 2. 22.2-Channel production systems under development

22.2-channel sound for homes

With the goal of introducing 22.2 multichannel sound into homes, we are advancing research on signal processing that will allow sound reproduction with fewer loudspeakers while maintaining the sound's spatial impression. In FY2009, we developed a method to automatically convert 22 channels into 8 channels, while maintaining sound pressure and directionality at the listening point. We also developed a method for reproducing 22-channel sound using only three forward speakers, by using the Head-Related Transfer Function, which represents the propagation characteristics of sound arriving at both ears from various directions. We also performed experiments to investigate the perception of the apparent sound source's elevation when reproducing sounds on loudspeakers and headphones. This research was in order to improve the spatial reproduction capabilities of the 22.2 multichannel headphone processor. We found that for sound coming from directly in front, the perceptual resolution of the sound's elevation was degraded through loudspeakers when the elevation angles exceeded 70 degrees or through headphones when the angle exceeded 40 degrees .

Psychology of hearing

We are also advancing research related to the processes contributing to the senses of presence and feeling ("Kandoh": deep emotional feelings) in order to evaluate the sound reproduction of high-presence audio systems in terms of a variety of factors. Conventionally, bipolar adjective pairs such as "clarity" and "muffledness" have been used to express impressions about sound, but words like muffled, for example, have negative nuance, so the impression given by a sound can be swayed by emotional preferences. Thus, we have proposed a two-dimensional metric for evaluating sound, with a sonic axis expressing impressions in terms of physical characteristics, and a preference axis expressing level of like and dislike. We have also investigated changes in impression due to differences in intensity when listening to music. As a result, we found that an optimum intensity can be estimated from the maximum value on the preference axis. We also conducted experiments to evaluate changes in the position of loudspeakers on impressions of expansiveness while listening to music. These impressions have a strong correlation with kandoh. Scores evaluating kandoh, together with those evaluating "sense of expansiveness", also have a strong correlation with the interaural cross correlation, which can be measured physically. This shows that kandoh is promoted by the expansiveness of the sound, and scores for kandoh can be estimated from a physical quantity.