Questions & Answers

The following are answers to some of the questions that were raised during the course of our seminar on Spatial audio. We would appreciate any comments/suggestions. You can contact us via email at dturaga@ece or alkanhal@ece .



Q.1 Technical details about Quadraphonic sound, Matrixing etc.

Ans. These can be found at the following site. They are not organized too well, but these are the best notes we could find on the web.

http://www.rdrop.com/~dano/tunes/quad.html  

Q.2 Information on the auditory cues used by the systems today to help us localize sounds.

Ans. Surprisingly, we have not been able to locate any information regarding this. However, it is quite reasonable to assume (as was pointed out by Prof. Chen in class) that the current systems exploit only the second class of auditory cues (i.e. intensity) for sound localization and do not take into account the first class of cues (Transfer function of the outer ear, head and shoulder).



Q.3 More details on how HRTF's are measured.

Ans. The answer is taken from Wightman, Kistler, "Headphone simulation of free-field listening.I ", J. Acoust. Soc. Am. 85(2), Feb. 1989.

 The approach is based on well-understood linear filtering principle.

 Let:

x1(t) represent an electrical signal that drives a loudspeaker in free field.
y1(t) represent the resultant electrical signal from a probe microphone positioned at a listner's eardrum.
x2(t) represent an electrical signal that derive the headphone, where
y2(t) the resultant microphone response.

Given x1(t), our goal is to produce x2(t) such that y2(t) equals y1(t).
We do this by designing a linear filter that transforms x1(t) into x2(t).

The design is described in the frequency domain (X1(jw), X2(jw), Y1(jw), Y2(jw) are the frequency transforms of x1(t), x2(t), x3(t), x4(t)).
We know

Y1=X1LFM
where L is the loud speaker transform function., F is the free-field-to-ear-drum transfer function and M the microphone transfer function.

We also know

Y2=X2HM
where H represents the headphone to eardrum transfer function and M is as above.
We want Y1 = Y2 and setting the two equations equal and computing the value of X2 we get
X2 = X1LF/H
This equation shows that the desired transfer function T is given by T=LF/H. Thus if the signal x1(t) is passed through this filter and the resultant x2(t) is transduced by the headphone, the signal recorded by the microphone will be y1(t)



Q.4 Description of Holophonics Technology

Ans. We couldn't find detailed information on the holophoinic technology, but from the information that is available on the web site http://www.holophonics.com it seems to be a new approach for 3D audio. On the other hand, some people believe nothing is special with the holophonics. They think of it as the same as the binaural technology. For such an opinion, you can refer to http://www.cs.colostate.edu/~meier/floyd/echoes4.html#5.1. Here is a summary of what holophonics has said about its technology and you can go to its web site for more information.
 
1. It uses a special holophonics head "Ringo" and an associated processing unit for recordings " No details. They are secretive". This technique is used to capture the full spectrum of essential information travelling from the ear to the brain in the recording environment.
 
2. It works very well with speakers and requires no decoding. Initial tests show that the 3D space in the movie theater is 10 times more realistic with Holophonics than with current surround sound.

3. This technology is based on a new understanding of the physiology of human hearing, and this new knowledge doesn't base on that we receive a sound in a passive manner "It doesn't base on the Head Related Transfer Functions". It bases on that human ear generates its own reference tone which interferes with incoming sound to create the necessary spatial information.

Back to the 3D Audio seminar page..