There are currently several ‘surround sound’ systems, most notably Dolby for Cinema and home. How is ambisonics different? In conventional ‘surround’ systems like Dolby 5.1, there is a ‘front’ and ‘centre’ from which music and dialog in films will mostly come. There are three other speakers. Two of them, at the back, concentrate on effects. The .1 of 5.1 is a sixth speaker, producing earthquake rumbles and other effects.
Ambisonics is not confined to the plane of the listener with
front, back and sides. It is a full-sphere surround sound technique: in addition to the horizontal plane, it covers sound sources above and below the listener. It is also ‘isotropic’. It has no preferred front direction, where all the important sounds happen and other directions carry only ‘effects’. All sounds are recorded and can be reproduced with the same quality, regardless of direction. It also means the sound can be rotated and transformed in many ways after the recording.
A little bit of history will help clarify.
First there was mono. Sound was recorded as if there was no direction, Music or birdsong or audience applause, they all came from the same direction. There were two problems with this. Sound had to be recorded from surprisingly close quarters for it to have clarity. If there were multiple musical instruments, it might be necessary to put one microphone in front of each and combine them. The second problem of course was was that there was no sense of envelopment – Everything came from one place, and at most there might be a pleasant amount of reverberation (not too much, not too little),
Stereo was invented to address the second issue. Good quality stereo recordings could create a sense of space, but it was still limited. Multiple microphones, mixed and panned to represent positions in the sound stage, did not make it better.
Alan Blumlein invented stereo, and his classic technique was to use two figure of eight microphones crossed. Variants were to use cardioid or more directional microphones, both on the same microphone stand, and angled at about 90 degrees from each other. I will not go into all the varients that were developed, but there was one significant method. It was called M-S which stood for mid side, The mid microphone could be a cardioid or Omni or any other pattern. The side microphone was always a figure of eight microphone. It was a satisfactory way of recording, and flexible in the way the signal could be processed after recording. One could for instance, narrow or widen the sound stage.
Four channel sound
The 70s saw a sudden influx of four loudspeaker systems, all calling themselves ‘surround’ . They all suffered from major problems. Most importantly there were no four channel playback systems. There were four track tape recorders, which could be altered to play all four tracks. There were heroic attempts at creating four channel Long Playing records. So most of the research and the technology went into squeezing four channels into a two channel recording, and unsqueezing it back at home to produce sound out of four loudspeakers, put at four corners of the room. Almost no effort went into understanding the psychoacoustics of human hearing. How do we accurately identify the source of a sound ?
Oxford mathematician and music lover Michael Gerzon developed a mathematical basis for four channel sound. It included height as one of the elements, and treated all directions of sound as of equal importance. He called the system ambisonics.
One can think of this as three dimensional equivalent of the M-S recording system. There are four channels, one represents the Omni signal called W, the other three X,Y and Z, represent three figure of eight signals.
The only problem here is, these have to be coincident microphones – all four microphones at the same physical space – which is not possible. The arrangement of four cardioid (or subcardioid) microphone capsules in a tetrahedron solves this problem. One can design delays to correct for non-coincidence, and a simple arithmetical addition and subtraction of outputs would generate the four wanted signals. The filters however had to be tailored for each microphone while measuring the results, and this made it an expensive process.
That was the situation till the 21st century began. Availabality of low cost multi channel digital equipment was the first important breakthrough, The second was the use of digital techniques for matching the capsules and creating a calibration file. Fairly low cost electret capsules could be simply matched. The resulting array could then be measured in different directions and a filter matrix created which corrected for remaining capsule errors, the non-coincidence and matrix of calculations to generate the W,X,Y and Z signals.
Brahma exploits all these advantages, and there is one more. As each array is measured against a calibration microphone, the resulting virtual microphones have a frequency response that matches that calibration microphone – flat over the spectrum in which the capsule operates.
There are several good and detailed descriptions of ambisonic theory on the web, including this Wikipedia article https://en.wikipedia.org/wiki/Ambis
A collection of articles, sources and links detailing the subject of Ambisonic surround-sound and digital audio technology, compiled by engineer/producer Richard Elen with over 25 years of experience of using the system. http://www.ambisonic.net/
This site is intended to accumulate my (Paul Hodges) understanding of sound recording, with special reference to ambisonics. https://ambisonic.info/
Maintained by Aaron Heller, and a major source of ambisonics information for those who want to delve deep. http://www.ai.sri.com/ajh/ambisonics/