以聽覺為基礎之個人化HRTF調整機制

Adaptive Mechanism for Personalized HRTF Based on Auditory Perception

莊盛宇(Sheng-Yu Chuang)

指導教授 ： 張嘉銘

大同大學/電機資訊學院/資訊工程學系所/碩士(2010年)

摘要

近年來3D立體音效技術廣泛的被運用在各種應用之中，要產生逼真的3D立體音效，就必須探討音源定位的問題。由於每個人先天上外觀差異，例如軀幹、肩膀、頭型、耳廓大小的不同，造就每個人獨特的HRTF。HRTF包含了聲波從音源到達耳膜間所有的頻率響應，和音源定位有相當大的關係。要進行音源定位，就必需對個人之HRTF進行準確的調整。
本論文提出了一個演算法，應用其他測試者已量測妥的HRTF資料為基礎進行調整，以合成出個人化的HRTF。首先，我們利用樣本資料的HRTF去製作半環繞音效。接著讓測試者去聆聽這些合成之半環繞音效，找出其中感覺有聲音發散現象的區段。利用多頻率聲音測試其他樣本同一位置的HRTF資料，用沒有聲音發散現象之片段來替換原來之片段。
將整個HRTF測試且代換完畢，讓測試者進行最後的聆聽測試，若有定位不準處再重覆替換測試，直到完成調整。利用這樣的系統，就可以不用經由繁瑣複雜，且具危險性的侵入式測量，而得到每個人獨特的HRTF資料。

關鍵字

聽覺心理學 ； 3D立體音效

並列摘要

In recent years the 3D three-dimensional sound effect technology is widely used in applications. In order to produce 3D three-dimensional sound effect similar to real, the sound source location must be discussed. Because appearance of each person is different to others, for example the torso, the shoulder, the head, and auricle. They accomplish each person's unique HRTF. All the effects of sound waves from sound source to eardrum are summarized and called the HRTF. A precise personalized HRTF is required to produce a simulated sound in a specified location.
In this paper proposed, an algorithm to synthesize the personalized HRTF from other subjects whose HRTF are measured by putting a microphone in their auditory meatus. First, one subject's HRTF is chosen randomly to produce a simulated sound with circular moving course. Then the listener listen to the simulated sound, and try to find the location of sound source which not disperse. Record the positions and synthesize several sound waves consist of multiple frequencies by convolution at same positions with HRTF from other subjects. Then the listener listen to these sound waves to find the best sound wave in which all the different frequencies are felt at the same location. The best piece of HRTF will replace to be the listener's HRTF.
After test the entire HRTF and substitute the wrong piece of HRTF, a simulated circular-moving sound is produced to perform final test. If the location course are not smooth, another multiple frequency test is needed until the adjustment is complete. By applying the algorithm, a personalized HRTF may be obtained without tedious complex, and the risky measurement.

並列關鍵字

HRTF ； HRIR

參考文獻

[2]Alan Kraemer, “Two speakers are better than 5.1,” IEEE Spectrum, May, 2001.

[3]J. W. Strutt (Lord Rayleigh), “On our perception of sound direction,” Philosophical Magazine, vol.13, pp. 214-232, 1907.

[4]張淑君,“空間音效模式化與數位實現之研究,” 國立成功大學，May 2002.

[5]D. Ramani, D. N. Zotkin, and N. A. Gumerov, “Interpolation and range extrapolation of HRTFs,” IEEE ICASSP, 2004.

[8]D. N. Zotkin, Ramani Duraiswami, Larry S. Davis, Ankur Mohan, and Vikas Raykar, “Virtual audio system customization using visual matching of ear parameters,” Proc. ICPR 2002, Quebec City, Canada, August 2002.