This study proposes computationally efficient binaural rendering for three-dimensional audio spaces. Instead of directly and individually convoluting each head-related impulse response (HRIR) within each sound object’s direction, only HRIRs of pre-defined representative directions are convoluted to the signals which are panned and summed up at representative positions. The panning is conducted adopting appropriate time shifts and gain adjustments to the original source signals, which approximately maintains waveforms of the source signals in synthesized sound images. We compared the obtained subjective and objective reproduction qualities with various representative HRIR direction arrangements using audio objects on a horizontal plane. Investigations indicated that the azimuth differences between representative HRIRs could be around 60?, and we then expect a reduction in computational complexity by approximately 50%–70% when 20–200 audio objects are rendered.
Click to purchase paper as a non-member or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member and would like to subscribe to the E-Library then Join the AES!
This paper costs $33 for non-members and is free for AES members and E-Library subscribers.