| Project/Area Number |
13650419
|
|---|
| 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 | Akita Prefectural University |
|---|
Principal Investigator |
TAKANE Shouichi Akita Prefectural University, Fac.Sys.Sci.Tech.,Dept.Electr.Info.sys., Assoc.Prof., システム科学技術学部, 助教授 (90236240)
|
|---|
| Project Period (FY) |
2001 – 2002
|
| Project Status |
Completed (Fiscal Year 2002)
|
| Budget Amount *help |
¥3,600,000 (Direct Cost: ¥3,600,000)
Fiscal Year 2002: ¥900,000 (Direct Cost: ¥900,000)
Fiscal Year 2001: ¥2,700,000 (Direct Cost: ¥2,700,000)
|
|---|
| Keywords | HRTF / Individuality / 3-D scanner / BEM / CARP model / 共通極・留数モデル / 留数展開 / 3次元計測装置 / レーザー光 / ダミーヘッド / 頭部伝達関数データベース / スーパーコンピュータ / データベース / 共通極・零モデル |
|---|
| Research Abstract |
In this research project, Boundary Element Method (BEM) was applied to the sound field around the human subject and the Head-Related Transfer Functions (HRTFs) were estimated from the results of the BEM computation. Geometrical shapes of the human subjects required for the analysis using BEM were measured by using optical 3-D scanner operated by a personal computer. It takes about one hour to obtain the entire head shape of a subject. This is much shorter than the direct measurement of his/her HRTFs in whole direction, which takes about six hours. It is also an advantage of the numerical estimation of HRTFs using BEM that the special facilities such as the anechoic room and the traverser of sound source are not required. Our developed 3-D measurement system yields the small measurement error of 6 mm (maximum) and 2 mm (average).
Individual HRTFs were estimated by using BEM and the obtained data. Comparison of the estimated HRTFs with the measured HRTFs showed that the important features such as the peaks in the frequency characteristics were. well estimated. However, there still exist some difference in the entire frequency characteristics. This difference may come from the incorrespondence of the condition of the estimation with that of the measurement. Investigation of such a difference remains as the future works.
As for modelling of HRTFs, interpolation of HRTFs using the Common-Acoustical-Pole and Residue (CAPR) model was proposed. This model represents the arbitrary HRTFs with the poles independent of the source direction and the direction-dependent residues. This model can efficiently model the HRTFs in terms of small number of parameters without the loss of accuracy.
The investigator considers that the results of this research project provide the insight into the new method of HRTF estimation, although some problems still remain unsolved. Further investigation is required on these problems.
|
