Re: distance perception in virtual environment (Kevin Austin )


Subject: Re: distance perception in virtual environment
From:    Kevin Austin  <kevin.austin@xxxxxxxx>
Date:    Sun, 17 Apr 2011 00:57:45 -0400
List-Archive:<http://lists.mcgill.ca/scripts/wa.exe?LIST=AUDITORY>

An interesting question, but I think I need a little clarification before being able to respond more fully. I take it that by "virtual environment", you mean only sounds that are being produced by loudspeakers. I have produced multi-channel / multi-speaker concerts for more than 30 years, out of doors and indoors, and the idea of "distance" was always complex and perplexing. I take it that the question is not about the sense of how far away the loudspeaker is, but rather, how far away the [illusion of the] sound is. Out of doors, my experience has been that a quiet sound close by, does not sound the same as a loud sound far away. This is made more complex if the sound is a "concrete" sound rather than a completely synthesized sound. Indoors, this is more complex by an order of many magnitudes, largely because it is source dependent. Short sounds can achieve high directionality, especially from point source materials. What makes a sound (indoors) appear to be close or far away? With voice, I have found that people easily identify as human voice as close or far away, almost entirely independent of the amplitude of the signal. The signal of the human voice changes its quality depending whether it is loud or soft. One of the first multi-speaker concerts I produced in a large hall (some 22 years ago) was a revelation. We had placed speakers at close range, medium and great distance, as well as speakers in stairwells, the entrance, facing walls etc. At the intermission, we noted that the 'effects' were so weak, that all of the medium and distant speakers were brought in to 1/2 their distance. By the end of the concert, we knew there was a perceptual problem thinking about "distance". The concerts were presented with the audience sitting on stage, facing out into the concert hall. 'Distance' perception, based on the distance of the loudspeaker, as an abstract concept simply didn't work. In listening to many many hundreds of electroacoustic works, in stereo, in stereo through a sound projection system, and multi-channel works, from 4 to 8 channels, it became clear that the most effective 'distance perception' arose from particular source materials that contained 'distance' perceptual cues. The two strongest being reverberation and lowpass filtering. Two rather exceptional [early] works, IMHO, as demonstrations of distance are John Chowning's Turenas, an almost adequate stereo mix (excerpt) being found here: http://modisti.com/musicbox/?p=4298, and Stria, http://www.youtube.com/watch?v=988jPjs1gao. These are carried to another dimension when heard in the original four-channel versions. In Turenas, I hear the use, on occasion, of two (or more) reverberation environments simultaneously, creating the effect (to my ears) of being in two different acoustical spaces at the same time. Another strong example of distance is that of the boy soprano in Stockhausen's Gesang der jünglinge http://www.youtube.com/watch?v=3XfeWp2y1Lk. Even the opening few moments allude to distance (in the stereo mix). This rather weak illusion is clarified in the four-channel original where different choirs appear in the four channels rather clearly (to me) at different distances. And in dealing with virtual distance, there is the possibility of virtual direction and virtual velocity. The sounds sometime change their distance, and carry a single unique acoustical perspective with them, into one loudspeaker. Again the four-channel version is several orders of magnitude clearer in this regard, IME. Stockhausen's Kontakte is a study in direction, distance, movement, velocity, immersion and sonic texture. I have prepared a very fundamental auditory guide to this, running close to 90 minutes, which can be found at: http://cec.concordia.ca/econtact/12_4/kontakte/index.php In Kontakte, there are many places (alluded to in the Analysis Notes), where it is not easy to separate distance from velocity, immersion and texture. Again, the way to hear this, in my experience, is in the four-channel version, with a high quality sound system, in an excellent acoustical environment. This cannot be considered to be a scientific experiment as such, as it is impossible to create a "control group". My experience is that the perception of these 'higher order' parametric values of sound events, is directly linked to the listener's auditory sensibilities and abilities. Please keep us informed. Kevin On 2011, Apr 16, at 9:24 PM, Junfeng Li wrote: > Dear list, > > I am now wondering how to subjectively evaluate distance perception in virtual environments which might be synthesized using WFS or HOA (high-order ambisonics). In my experiments, the sounds were synthesized at different distances and presented to listeners for distance discrimination. However, the listener cannot easily perceive the difference in distance between these sounds. > > Anyone can share some ideas or experiences in distance perception experiments? or share some references on this issue? > > Thank you so much. > > Best regards, > Junfeng


This message came from the mail archive
/var/www/postings/2011/
maintained by:
DAn Ellis <dpwe@ee.columbia.edu>
Electrical Engineering Dept., Columbia University