• 제목/요약/키워드: 머리전달함수장

검색결과 2건 처리시간 0.014초

머리 전달 함수장 재현을 통한 광대역 입체 음향 구현 (HRTF-field reproduction for robust virtual source imaging)

  • 최정우
    • 한국소음진동공학회:학술대회논문집
    • /
    • 한국소음진동공학회 2007년도 추계학술대회논문집
    • /
    • pp.997-1004
    • /
    • 2007
  • A hybrid technique that combines the advantages of binaural reproduction and sound field reproduction technique is proposed. The concept of HRTF-field, which is defined as the set of HRTFs corresponding to the various head dislocations, enables us to realize virtual source imaging over a wide area. Conventional $2{\times}2$ definition is redefined as a MIMO system composed of multiple control sources and multiple head locations, and HRTF variations corresponding to various head movement are quantified. Through the direct control of HRTF-field, reproduction error induced by head dislocation can be minimized in least-square-error sense, and consequential disturbances on the virtual source image can be reduced within a selected area. Simple lateralization examples are investigated, and the reproduction error of the proposed technique is compared to that of Higher-order Ambisonics.

  • PDF

머리 전달 함수장 재현을 통한 광대역 입체 음향 구현 (HRTF-field Reproduction for Robust Virtual Source Imaging)

  • 최정우
    • 한국소음진동공학회논문집
    • /
    • 제18권2호
    • /
    • pp.199-207
    • /
    • 2008
  • A hybrid technique that combines the advantages of binaural reproduction and sound field reproduction technique is proposed. The concept of HRTF-field, which is defined as the set of HRTFs corresponding to the various head dislocations, enables us to realize virtual source imaging over a wide area. Conventional binaural($2{\times}2$) reproduction system is redefined as a MIMO system composed of multiple control sources and multiple head locations, and HRTF variations corresponding to various head movement are quantified. Through the direct control of HRTF-field, reproduction error induced by head dislocation can be minimized in least-square-error sense, and consequential disturbances on the virtual source image can be reduced within a selected area. Simple lateralization examples are investigated, and the reproduction error of the proposed technique is compared to that of higher-order Ambisonics.