| Project/Area Number |
13650404
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
情報通信工学
|
|---|
| Research Institution | University of Yamanashi |
|---|
Principal Investigator |
OZAWA Kenji University of Yamanashi, Faculty of Engineering, Associate Professor, 工学部, 助教授 (30204192)
|
|---|
| Project Period (FY) |
2001 – 2002
|
| Project Status |
Completed (Fiscal Year 2002)
|
| Budget Amount *help |
¥3,000,000 (Direct Cost: ¥3,000,000)
Fiscal Year 2002: ¥1,400,000 (Direct Cost: ¥1,400,000)
Fiscal Year 2001: ¥1,600,000 (Direct Cost: ¥1,600,000)
|
|---|
| Keywords | binaural reproduction / 3D spatial information of sound / impulse response / word intelligibility / naturalness / video conference / H.263 coding / competitive talkers / H.263符合化 / スペクトル / 音響特性補正 |
|---|
| Research Abstract |
In order to yield natural communication in distant collaboration via a network, it is important to transmit and reproduce spatial information of sound as well as clear speech signals. The objective of this study was to establish a method to transmit and reproduce of three dimensional spatial information of sound in the receiver side of the sound. The study was conducted in the following two aspects.
(1) High fidelity reproduction of transmitted spatial information by the receiver
To examine the property of the proposed method, the term to correct frequency characteristics of the transmitted acoustical information in the proposed method was investigated by psychoacoustical experiments. The effect of determination of the term by reducing the impulse response was evaluated by subjects. As a result, the impulse response could be reduced to 0.1 in length to conserve the naturalness of the recorded sound. This certificates the validity of the proposed method to be used in an actual condition.
(2) Reconstruction of three dimensional information of sound
To examine the effectiveness of the proposed method, the performance of the method was compared with ordinal methods, i.e. monaural and stereophonic systems, in a video conference with multiple talkers. Three talkers pronounced different words simultaneously, and a listener was asked to listen to the word of the designated talker. The obtained word intelligibility of the proposed method was approximately 20 % higher than those of the ordinal methods. This effectiveness of the proposed method was independent of the quality of accompanying video.
|
