Header image  
by Ralph Glasgal
line decor
Home Tutorials Tech
Kudos and
Demos Bio Free Ambio
Glossary Home
Rec Engineers
FAQ/Forum Links Contact us
line decor

Audio Engineering Society

Presented at the 99th Convention
October 6-9, 1995 New York, NY



The Synthesis of Concert-Hall Sound Fields in the Home

by Ralph Glasgal


The following areas must be addressed when converting a stereo or surround-sound system to Ambiophonics:

  1. The interaural crosstalk must be eliminated by the use of a central absorption barrier between the front speakers, extending to the listening position (Figs. 5 and 6).

    Fig. 5 - Ambiophonic main front channel listening arrangement eliminating crosstalk and mimicking microphone view.

    Fig. 6- Ideal barrier design to eliminate crosstalk and reduce room reflections.

  2. The reverberation time of the listening room should be reduced to .2 sec. or less at the listening positions.
  3. The front speakers should be moved to a position on either side of the barrier subtending an angle of 5° to l0° to the median plane.
  4. The loudspeakers chosen should be as directional as possible. The ideal front loudspeakers would be point sources that behave like wideband flashlights.
  5. Early proscenium reflections, side reverberation, and rear reverberation should be computer or DSP synthesized, with left and right channels processed separately and if possible with subtly different ambience parameters, so as to increase the hall IACC.
  6. Ambience loudspeakers should be both directional and large. Electrostatics or ribbons provide ambience diffusion without exacerbating room reflections as diffusion panels would (Fig. 7)

    Fig. 7 - This particular arrangement of tall, leaning speakers has produced an exceedingly realistic, Ambiophonic, "You-are-there" sound field.

  7. The proscenium early-refleciion loudspeakers should be placed at 55° to the left and right of the listening position, so as to minimize the IACC and synthesize the most interesting performance spaces, regardless of size (see Ando).


The only practical way of eliminating interaural crosstalk is to use a barrier between the speakers that extends to the listening position. A second listener can sit behind the first, as the first listener's head and body constitute an extension of the crosstalk barrier. While many electronic crosstalk-elimination methods have beon tried, none is as effective or inexpensive as the barrier. Once the barrier is in place, it is possible to move the loudspeakers to a position almost directly in front of the listener. This position has the effect of more closely duplicating the ORTF, or dummy-head, recording schemes in wide use today. It also eliminates the conflict between pinna response and interaural response that can occur when a centrally located sound source hits the ear from speakers that are located 30 or more degrees to the left and right. But by the same token instruments or voices coming from the extreme right or left (more than 120°) will be shifted slightly toward the center, because the intra-aural pinna directional effect overrides the interaural time/intensity effect.


The provision of an absorbing barrier between the main front speakers is something one has to get used to living with like the large screen in projection home video theaters. When stereo recordings first appeared, there were those who said it could never catch on because of the unsightly wire to the second loudspeaker and the necessity of sitting between them meant moving furniture. Now we routinely have six or eight loudspeakers in a surround-sound video room with wires running everywhere and interior decorators making a good living servicing those who want everything hidden or built in.

So let it be with the barrier. The barrier can come down from the ceiling, up from the floor, or out from the front wall, just like the screen and projector. A folding screen like that from MSB Technology, can be put away between listening sessions. Pillobaffles, from RPG Diffusor Systems, can be hung like a drape from hooks on the ceiling and easily removed when not wanted, since they weigh only six pounds. Other prefabricated barriers are available from Acoustic Sciences Corporation and System Analysis Company. Similar matching panels can be used to reduce the reverberation time of the listening room to less than .2 seconds. (Discussion of barrier and room materials can be found in the Glasgal-Yates reference in the bibliography.)


In general the Ambiophonic method is best served by speakers that have very narrow horizontal or vertical dispersion. Obviously, the less sound a loudspeaker sprays around the room, the easier it is to eliminate spurious reflections and room reverb. The side and rear ambient speakers also benefit from having a large radiating surface area such as found in electrostatics. Flat electrostatic speakers, or ribbons, provide diffusion aimed at the listening position, without stimulating unwanted room reflections. Multiple small satellite speakers can also be distributed randomly about the rear half of the room to achieve dispersion of the reverberant field.


Ambience synthesizers are still in their infancy, and with the emphasis on surround sound are getting harder to come by. By far the most intriguing work in this field has been done by JVC. They use an orthogonal array of six microphones, arranged in three head-spaced pairs sharing a common origin. This array can be placed at the best seat in any auditorium and used to record the impulse response of the hall. The outputs of the six microphones go to a computer, which calculates and stores the direction, intensity, and time of arrival of the early reflections and the direction, frequency response, intensity, and rate of decay of the reverberant field. These hall parameters can then be stored and used as is or modified by the listener. Two JVC XP-A1010s are now used in the experimental Ambiophonics system to good effect. Two Lexicons (CP-1 or CP-3) could also be used. Many surround-sound processors have ambience-generating modes, and even if not ideal, they will produce a noticeable improvement, if the barrier is in place and the room has been treated to eliminate irksome room reflections and shorten its reverberation time.


We can all agree that different types of music sound best in different types of halls. For instance?, symphony orchestras usually sound good in concert halls, string quartets sound better in salons or recital halls, and organs are more at home in churches or cathedrals. While one could construct a home listening room that could very accurately mimic Carnegie Hall, this would not be appropriate for a listener whose record collection also includes jazz, opera, madrigals, lieder and solo piano. Any home music theater must be capable of adapting quickly to each type of music played. Fortunately, the existing JVC XP-A1010 or Lexicon CP-3, and the software-based systems expected soon provide for rapid adjustment of the ambient field from the listening position. It is to be hoped that someday the CD itself will incorporate a digital preamble to control directly the ambience generator. In the meantime, it would be convenient if recording engineers were to include some measurements of the recording site in the program notes. Of particular interest would be the onset of the first and second early reflections, the high-frequency reverberation time, the low-frequency reverberation time, and the high-frequency roll-off of the ambient field in general.

Next Page >>