Sound-Localization-Test-Based Analysis of a Sound Image Generating System using a Flat Display and Horizontally Attached Acoustic Barriers 2014 0123 Shunki KUMAGAI 2016 1 21
Abstract Based on the recent dramatic advancement of information technologies, computer-supported cooperative work has been attracting much interest from researchers. The ultimate goal of systems that support cooperative work is to provide users with a sense of being in the same room. t-room studies have been conducted to realize that sense. Among them, we focus on a unique sound image generating system that adopts a triple-speaker unit covered by an acoustic barrier. This system has two versions: 1) one that places two speaker units on the left and right sides of the display (left-right model), and 2) another that places the units on the top and bottom sides of the display (top-bottom model). The effect of the left-right model in sound image localization accuracy was shown in both acoustic measurements and (perceptual) localization tests. However, the speaker units on the left and right sides of the display are visual obstacles, and complicate connecting the displays in the wide wall mode. On the other hand, the top-bottom model was only studied in acoustic measurements, although its obtained results are promising for sound localization. The top-bottom model outperforms the left-right model for the visual obstacle problem. Focusing on these features for the top-bottom model, we evaluate its effects on a sound image generating system in sound localization test experiments. Our experimental results are three-fold. The tested top-bottom model achieves relatively high accuracy in the the horizontal direction localization, but not in the vertical direction localization. Most importantly, the results of the top-bottom model are almost the same as the left-right model. These results show the utility of a sound image generating system that places the speaker units on the top and bottom sides of a large display. Keywards 1. Acoustic image generating method 2. Sound image localization i
1 1 1.1............................................. 1 1.2............................................. 1 1.3........................................ 2 2 3 2.1 t-room 1)..................... 3 2.2...................................... 4 2.2.1.................................. 4 2.2.2............... 6 3 8 3.1....................... 8 3.1.1.............................. 8 3.1.2....................................... 9 3.2....................... 10 3.2.1.............................. 10 3.2.2....................................... 10 4 12 5 14 5.1........................................... 14 5.2....................................... 15 6 21 6.1............................................ 21 6.2......................................... 21 23 24 ii
1 1.1 t-room 1) 2) 6 t-room Rodenas 3) 2ch 1ch 4ch t-room Gabriel 4) t-room 6) 1.2 1
1.3 2 t-room 3 4 5 6 2
2 2.1 t-room 1) t-room t-room 4) Fig 2.1 t-room ( 1) ) 3
2.2 HRTF : head-related transfer function HRIR : head-related impulse response t-room 2.2.1 2 ITD : interaural time difference ILD : interaural level difference Blauert Fig. 2.2 Fig 2.2 ( 8) ) 0ms 1ms -1ms 1ms 1600Hz Blauert 0dB ±10dB Fig. 2.3 4
Fig 2.3 ( 8) ) 5kHz 5kHz-10kHz Fig. 2.4 5
Fig 2.4 ( 8) ) 2.2.2 2 0 0 2 6
1ms 50ms Fig 2.5 Fig. 2.5 Sl Sr C 2 R Sl Sr Sl Sr R Sr Sr Sr t-room 7
3 Gabriel t-room 4) 6) 3.1 3.1.1 Fig. 3.1 t-room 60 3 L Fig 3.1 ( : 4) ) 8
3.1.2 2.2.2 Fig. 3.2 C 0 C R Sl Sr R 6 Fig 3.2 9
3.2 3.2.1 Fig. 3.3 60 3 L Fig 3.3 ( : 6) ) 3.2.2 Fig. 3.4 C R 10
Fig 3.4 11
4 Fig. 4.1 Fig 4.1 0 x 1080 (4.1) 0 y 1920 (4.2) 0,0 (1080,1920) 9) Fig. 4.2 ϕ ϕ 0 (4.3) 12
Y L Y R tanϕ tanϕ 0 = W L W R W L +W R (4.3) Fig 4.2 9) Y L Y R (4.3) x (4.4) X L evel = a exp(b x ) + c exp(d x ) (4.4) x = x x (4.5) 540 13
5 5.1 5.1 6 36 18 27 24 1m 12 2m 5.1 5 15 1m 2m 3 5cm 80dB 3 Table 5.1 ONKYO EP150402A04 Behringer REFERENCE AMP. A500 MOTU 828mk3 Hybrid SHARP ON-E602 14
Fig 5.1 ( : : ) 5.2 1m 2m 1m 5) 7) 6 1 11 1 1-5 0 1 2 Fig 5.2 : 1m : 1m 15
Fig 5.3 : 2m : 2m L C R Fig. 5.2 L 4 5 R 2 5) Fig. 5.2 Fig. 5.3 1m 1m 6) 2m 1000Hz 2m 2m 1m 2m Fig 5.4 : 1m : 1m 16
Fig 5.5 : 2m : 2m Fig. 5.4 Fig. 5.5 2m 1m 1m 2m 1m 1m 1m 5) 1m 7) 70 % 70-50 % 50-30 % Fig 5.6 L : : : 17
Fig 5.7 C : : : Fig 5.8 R : : : Fig. 5.6 Fig. 5.11 L C R 1,3,5 1,3,5 1,3,5 2,4 2,4 L C R 1-5 2,3,4 0,1,2 50 % 50-30 % 30-20 % 18
Fig 5.9 L : : : Fig 5.10 C : : : Fig 5.11 R : : : L C R 0,1,2 (2,3,4) 2 19
3 50% 20
6 6.1 1m 2m 1m 2m 6.2 t-room 21
22
1) Keiji Hirata, Yasunori Harada, Toshihiro Takada, Shigemi Aoyagi, Yoshinari Shirai, Naomi Yamashita, and Junji Yamato, The t-room: toward the future phone, NTT Technical Review, vol.4, no.12, pp.26-33 (2006). 2) Shigeaki Aoki,Nobuo Koizumi, Expansion of listening area with good localization in audio conferencing, Proc. ICASSP 87, pp.149-152, 1987. 3) Josep A. Ródenas, Ronald M. Aarts, A. J. E. M. Janssen, Derivation of an optimal directivity pattern for sweet spot widening in stereo sound reproduction, J. Acoust. Soc. Am., 113(1), 267-278, 2003. 4) Gabriel Pablo Nava, Keiji Hirata, Masato Miyoshi, A loudspeaker design for sound image localization on large flatscreens Acous. Sci. & Tech. 31(4) 278-287 2010 5) Gabriel Pablo Nava 2011 6),,, 2015. 7) : 2015 8) 2011 9) Theory of Three-Dimensional Sound Field Reproduction IEICE Fundamentals Review Vol3 No.4 pp33-46 2009 23
i) 2016 2016 24