This project aims at cataloguing, conserving, and making available in digital form as much information as possible about the surround sound framework known as Ambisonics (here, "ambisonic"). We've undertaken this project to ensure that the knowledge is preserved, and to make it easier for fresh implementers to get a hold of a technology we think is both underappreciated and uniquely powerful.
If you think you can help, suggestions, private reserves of information, code and relevant references should be submitted either to the primary editor, or to the sursound mailing list (no attachments, give a live link instead). Digitized material should be delivered as multipage, losslessly compressed TIFF or PDF. Optionally, but also preferably, a lossily encoded and especially an OCR'd version should be included.
Any exclusive rights should be divulged at the time when new materials are submitted, if at all possible. Similarly, complaints about our possible misuse of IP should be raised forthwith—the subject matter is both old and varied enough so that our small group of volunteers simply has no practical, economical means of ascertaining whether we can distribute the material or not. We have to rely solely on feedback.
This section contains an evolving collection of articles on the ambisonic framework. We will try to classify and prioritise them better and more glanularly as time goes by. The aim is to make this into an easily navigatable library of everything there is to know about ambisonic. In the meanwhile all suggestions for further development are welcome, and we sincerely hope you enjoy what we've come up with thus far.
Title | Author(s) | Publication information | Description |
---|---|---|---|
General Metatheory of Auditory Localisation | M. A. Gerzon | Presented at the 92nd Convention of the AES; 1992-03-24 to 1992-03-27; Vienna, Austria | The main theoretical paper underlying the psychoacoustics and the general mathematical formulation of ambisonic. |
Further Investigations of High Order Ambisonics and Wavefield Synthesis for Holophonic Sound Imaging | J. Daniel, R. Nicol and S. Moreau | Presented at the 114th convention of the AES; 2003-03-22 to 2003-03-25; Amsterdam, The Netherlands | A highly condensed, and theoretical, yet balanced presentation of the possibilities and the relative limitations of WFS and HOA. |
Title | Author(s) | Publication information | Description |
---|---|---|---|
Circular microphone array for discrete multichannel audio recording | E. Hulsebos, T. Schuurmans, D. de Vries and R. Boone | Presented at the 114th convention of the AES; 2003-03-22 to 2003-03-25; Amsterdam, The Netherlands | TEMP!!! |
A New Comprehensive Approach of Surround Sound Recording | A. Laborie, R. Bruno and S. Montoya | Presented at the 114th convention of the AES; 2003-03-22 to 2003-03-25; Amsterdam, The Netherlands | TEMP!!! |
Applications of Binaural Processing to Surround Sound Reproduction in Large Spaces | C. Landone and M. Sandler | Presented at IEEE ISCAS; 2000-05-28 to 2000-05-31; Geneva, Switzerland | TEMP!!! |
This is our working area. Unless you're a devoted ambisonic/surround fan, look no further. But if you are, do contact the primary editor of this collection: as soon as you read one of these files through, you're in a position to help us develop this project further. It's little cost to you, and much benefit to the public at large.
This is then the diehard section. Suggestions are again welcome, but there are limits: this is mostly for the few who've preserved hardcopies for the posterity, who otherwise have access to esoteric materials, who can deal out references-long-forgotten, or who've had restricted contacts with credible researchers which bind them from releasing the Beef.
Title | Description | Date | Comments | Status | Location/contact | Priority | Reference | Should include | Shouldn't include |
---|---|---|---|---|---|---|---|---|---|
Professional Ambisonic Decoder Design | Hardcopy, NRDC ring-binder, 12 pages | 1981-07 | Horizontal first order UHJ & B-format. Circuits & component lists. I am informed the Abacoid Decoder closely followed this design: I thought highly of the Abacoid when testing and reviewing for Studio Sound | Located | Peter Carbines | High | Box #1 | 1 | 0 |
System UHJ Encoding Standards | Hardcopy, NRDC bound document, 32 pages | 1977-11 | Located | Peter Carbines | High | Box #1 | 1 | 0 | |
Ambisonics & UHJ Multiplex FM Broadcasting | Hardcopy, NRDC bound document, 22 pages | 1979 | Located | Peter Carbines | Medium | Box #1 | 1 | 0 | |
NRDC Periphonic Decoder | Hardcopy, Photocopy of Geoff Barton's handwritten document | 1979-12 | Circuit and construction. Marked "Used for AES Demo Feb 1980" | Location kwn | Peter Carbines | Low | Box #1 | 1 | 0 |
Ambisonics: Part One: General System Description | Hardcopy, Photocopy of Peter Fellgett's article | 1975-08 | Studio Sound August 1975 | Located | Peter Carbines | High | Box #1 | 1 | 0 |
Ambisonics: Part Two: Studio Techniques | Hardcopy, Photocopy of MAG's article | 1975-08 | Studio Sound August 1975 | Located | Peter Carbines | High | Box #1 | 1 | 0 |
Practical Periphony: the Reproduction of Full-Sphere Sound | Hardcopy, Photocopy of typewritten draft by MAG | 1980-02 | Drafted for presentation at 65th AES Convention, London February 1980 | Located | Peter Carbines | High | Box #1 | 1 | 0 |
Periphony: With Height Sound Reproducer | Hardcopy, MAG paper | 1972-03 | JAES Vol 21 No 1 Jan/Feb 1973 | Located | Peter Carbines | Medium | Box #1 | 1 | 0 |
AMBISONIC TECHNOLOGY: Report #2: Studio Signal Handling Formats and Encoding | Hardcopy, Photocopy MAG handwritten report | 1975-08 | Located | Peter Carbines | High | Box #1 | 1 | 0 | |
AMBISONIC TECHNOLOGY: Report #3: Pan Pot and Sound Field Controls | Hardcopy, Photocopy MAG handwritten report | 1975-08 | Located | Peter Carbines | High | Box #1 | 1 | 0 | |
AMBISONIC TECHNOLOGY: Report #4: Artificial Reverberation and Spreader Devices | Hardcopy, Photocopy MAG handwritten report | 1975-08 | Includes 'With-height 2-channel spreader' and 'Spread Pan Pot' | Located | Peter Carbines | High | Box #1 | 1 | 0 |
AMBISONIC TECHNOLOGY: |
Hardcopy, Photocopy MAG handwritten report | 1978 to 1979 | Report #5: incomplete copy of 25 pages. This may NOT be Report 5 as date on documents makes this unlikely. It is included here as it fits with the sequence of reports to NRDC. Design equations and circuits for various manipulative devices: Quadrant Spread pan pot; Ambisonic/Stereo quadrant pan pot; Soundfield Blur Device; Triaxial Pan Pot theory; Quadraxial mixing; Suggested mixer for demonstrations; Multi-input 'wrap' device. | Located | Peter Carbines | Low | Box #1 | 1 | 0 |
AMBISONIC TECHNOLOGY: Report #6: Studio Decoders | Hardcopy, Photocopy MAG handwritten report | 1975-09 | Located | Peter Carbines | High | Box #1 | 1 | 0 | |
Market Research Report into Consumer Attitudes to Ambisonics | Hardcopy, BTG bound document, 70 pages | 1981-10 | Commissioned report. The Aiwa/Boots Micro-system was used for 'hall' testing. | Located | Peter Carbines | Low-/Historical | Box #1 | 0 | 1 |
REPRINTS OF ARTICLES AND REVIEWS | Hardcopy, Ambisonic Technology Centre bound document | 1992 | Prepared for the infamous "Abbey Road" demo. Includes the SS review on the Abacoid Decoder referred to above. | Located | Peter Carbines | Low | Box #1 | 1 | 0 |
AMBISONIC SURROUND SOUND SYSTEM | Hardcopy, NRDC printed 4-page introductory document | Located | Peter Carbines | Low-/Historical | Box #1 | 0 | 1 | ||
AMBISONIC TECHNOLOGY FOR SURROUND REPRODUCTION OF SOUND | Hardcopy, NRDC printed 2-sided A4 page | Located | Peter Carbines | Low-/Historical | Box #1 | 0 | 1 | ||
SOME PRODUCTION FACILITIES AVAILABLE IN AMBISONICS | Hardcopy, NRDC bound typescript, 14 pages | 1977-11 | Located | Peter Carbines | Low | Box #1 | 1 | 0 | |
Ambisonic and Wavefield Synthesis Theory and Limitations | File, presentation slideset, Christof Faller, lecture at Helsinki University of Technology | 2008-02-27 | Restricted by an NDA | Located | Sampo Syreeni | Medium | 2008_HUT_A_WFS.pdf | 1 | 0 |
Signal Processing for Audio and Acoustics | File, course notes, Christof Faller, at EPFL Lausanne, Switzerland | 2008-02-22 | Restricted by an NDA | Located | Sampo Syreeni | Low | lecture_spatialaudio.pdf | 1 | 0 |
It seems that it is impossible to describe all of the stuff in the lode using simple English. Thus, the specialised terms used have to be explained as well. This is why a kind of concept glossary is needed. Again, feel free to contribute to this part as well, in separation from the others. Suggestions for new terms/entries would especially be appreciated, because once you're privy to the literature, it is rather hard to spot the concepts that are the most confusing to a newcomer. Those terms then are clearly also the most worthy of detailed explanation.
Concept | Expansion | Detailed explanation |
---|---|---|
45J | TEMP!!! | TEMP!!! |
468-weighting | weighting function defined in ITU-R 468 | TEMP!!! |
A-format | raw microphone signals | A-format refers to the raw signal set captured directly from microphone array outputs. The term is in particular used with the output of the tetrahedral array within the SoundField mic. These signals are not usually accessible to the production environment, but are instead first transformed into the array independent B-format. |
A-weighting | TEMP!!! | TEMP!!! |
active matrix | TEMP!!! | TEMP!!! |
Ambisonic(s) | surround sound | A comprehensive framework of surround sound capture, processing and reproduction. Ambisonic is distinguished by its ability to reproduce soundfields isotropically even with few channels, by its extensive mathematical underpinnings grounded in the theory of spherical surface harmonic functions, by its tendency towards full holophony at high channel counts, by its use of a microphone and speaker agnostic intermediary format in transmission (B- and C-formats), by a highly developed and physical acoustics based effects infrastructure, and by the presence of a psychoacoustic decoder at the receiver side, which help optimize the final playback for different audiences. |
B-format | native ambisonic representation of a soundfield | This is the standard professional format for ambisonic. Originally it used four signals: W, X, Y and Z. W is the zeroth order spherical harmonic, which is omnidirectional. XYZ are the first order ones which give the dipole moments of the soundfield; that is, they represent fig-8 microphones aligned with the left-right, front-back and top-bottom axes. If only horizontal playback is wished, the Z component (tup-bottom) is discarded. |
B-weighting | TEMP!!! | TEMP!!! |
BBC | british broadcasting corporation | TEMP!!! |
BHJ | TEMP!!! | TEMP!!! |
BLaH | Benjamin, Lee and Heller | TEMP!!! |
Blumlein pair | crossed fig-8's towards L/R | TEMP!!! |
C-format | transmission channel optimized format | This is the format that is transmitted to the consumer. It is optimized for the specific transmission channel that is available. C-format also tries to be compatible with the transmission channel, so that ambisonic decoding isn't strictly necessary to yield satisfactory results. Most commonly the channel is a normal stereo one, and the C-format that is then used is the mono and stereo compatible variant of the UHJ hierarchy of encoding methods termed BHJ. Nowadays this could also be G-format, if the transmission channel is 5.0 or 5.1. |
C-weighting | TEMP!!! | TEMP!!! |
cardioid (pattern) | TEMP!!! | TEMP!!! |
classical | conforming to all of the POA knowledge | Quite a number of attempts at implementing ambisonic do not go deep enough into the theory or the tradition to catch even what is already known. That especially happens with ambisonic decoders, because it is far too easy to just go with the beauty of the underlying math, without paying attention to the messy, empirical facts of the field. Thus, "classical" has come to mean something that does pay attention to the tradition, especially on the decoder side. The BLaH series of papers represents a modern attempt to precisely quantify how classical your decoder then is, or in other words how well it follows the most basic psychoacoustics we know about. |
compatibility coding | coding that is directly listenable in another format | UHJ is mono and stereo compatible, Dolby MP is mostly stereo compatible, most quadraphonic systems aimed at the same. TEMP!!! |
crosstalk cancellation | TEMP!!! | TEMP!!! |
D-format | physical speaker feed signal set | This is what is fed to the speakers. In common surround sound and quadrifontal systems this would simply be the transmitted signal set. In ambisonic, however, the transmission format is mic and speaker agnostic. So the eventual speaker feeds have to be derived by an intelligent decoder, which then produces D- format for a specific speaker array. |
D-weighting | TEMP!!! | TEMP!!! |
dB | decibell | A comparative logarithmic scale which utilizes base ten logarithm. It is commonly used in conjunction with perceptually relevant and other measurements which have a wide dynamic range. There are different kinds of dB's for each different base unit, and each different base level used for comparison. The most widely known example is dB (SPL). |
dipole | TEMP!!! | TEMP!!! |
(Dolby) MP | (Dolby) matrix processing | TEMP!!! |
DirAC | directional audio coding | TEMP!!! |
distance compensation | first order NFC on the decoder side | meant to compensate for phase errors, but extends via NFC-HOA to wavefront deconstruction over wider areas. TEMP!!! |
discrete (encoding) | each channel represents a point/planewave source in the surrounding space | TEMP!!! |
energy vector | TEMP!!! | TEMP!!! |
figure of eight (fig-8) (pattern) | TEMP!!! | TEMP!!! |
(forward) dominance | a transform which amplifies a given direction and pushes sources towards it | The dominance, or zoom, transform, amplifies sound in one direction, diminishes it in the opposite one, and at the same time skews the angular distribution of sound sources so that sources are pushed towards the dominant direction. The effect is not unlike an optical zoom lense, only the source directions move opposite to the amplitude enhancement. This control has only been realized at first order, where it technically constitutes a time-invariant Lorentz transform of the first order B-format signal set. |
(forward) preference | preferential control of phasiness in a given direction, usually forwards | When decoding from underdetermined C-format (most commonly BHJ), optimum decoding always leads to some extra phasiness in some part of the decoded soundfield, above that of the best decode of the underlying ambisonic order and the number of speakers in use. The forward preference (sometimes called "focus") control then enables the user to reduce phasiness in the important forward direction at the expense of increasing it elsewhere. This control does not affect directionality at all, but only the distribution of the out-of-phase components that are being reproduced. Contrast with (forward) dominance. |
G-format | simple predecode of ambisonic for 5.0 | a subtype of C-format. TEMP!!! |
hierarchical coding | coding that extends by only adding more channels | UHJ has this, most others don't. MPEG's backwards compatible codecs attempt this, but fall short of the sophistication. TEMP!!! |
HJ | Matrix H/45J | HJ is the name given to the encoding specification jointly agreed to by the BBC and the NRDC, as a result of the fusion of BBC Matrix H and NRDC 45J. It is in the form of a tolerance zone defined on top of the Scheiber sphere, and as such only dictates how two-channel encoding is to be done. |
HOA | Higher Order Ambisonic | The recent development of ambisonic technology so that it can faithfully capture, store and render higher orders of spherical harmonics than just the first two. This yields higher directional accuracy, wider sweet spots, and lessened phasiness even outside the sweet spot. Plus it in theory enables more decoding options, because more data is available. The possibility of higher orders was acknowledged even in the earliest ambisonic papers, but was deemed impracticable at the time. Thus the distinction between POA and HOA might be misleading: this is the same system, it is just that we've only begun to explore its more demanding possibilities. At very high orders HOA converges to genuine holophony. |
holophony | isotropic surround sound in 3D/all of the space | TEMP!!! |
HRTF | head related transfer function | TEMP!!! |
hypercardioid (pattern) | TEMP!!! | TEMP!!! |
isotropic | TEMP!!! | TEMP!!! |
K-H (integral) | Kirchoff-Helmholtz integral | TEMP!!! |
kernel (encoding) | each channel represents a continuous 3D radiation pattern | TEMP!!! |
L (channel) | TEMP!!! | TEMP!!! |
M/S | mid/side; front facing cardioid plus a sideways fig-8 | TEMP!!! |
MAG | Michael A. Gerzon | TEMP!!! |
Makita theory | TEMP!!! | TEMP!!! |
Matrix H | eighth matrix tested by the BBC | TEMP!!! |
matrix(ing) | TEMP!!! | TEMP!!! |
MkIV, MkV | SoundField microphone, marks four and five | TEMP!!! |
MLP | Meridian lossless packing | TEMP!!! |
monopole | TEMP!!! | TEMP!!! |
multipole | TEMP!!! | TEMP!!! |
NFC | Near Field Correction | extended distance compensation, done at the encoding end. remember s/n ratio/headroom. TEMP!!! |
NRDC | National Research and Development Council | TEMP!!! |
omni(directional) (pattern) | TEMP!!! | TEMP!!! |
order | order of truncation of the spherical harmonic series | TEMP!!! |
pantophony | isotropic surround sound in 2D/in the horizontal plane | TEMP!!! |
pairwise mixing/panning (style) | simulating direction using only the nearest two speakers | Pairwise mixing is the practice of locating a mono signal panning between the nearest two loudspeaker feeds in the same general direction. Ambisonic on the other hand assigns mono sources a number of coefficients in B-format, which then translate in the decoding stage to psychoacoustically optimal D-format. The main difference is that all of the speakers are used in unison to reconstruct ambisonic, and their position will be fully taken into account. |
passive matrix | TEMP!!! | TEMP!!! |
periphony | isotropic surround sound in 3D | Periphony means 3D, or "full-sphere" surround sound which doesn't discriminate between any two directions. I.e. it is isotropic. Ambisonic's original 4-channel B-format framework is arguably most notable because it delivers periphony out-of- the-box. Something that other surround sound systems struggle with even today, while using far more channels. |
PHJ | TEMP!!! | TEMP!!! |
POA | Plain Old Ambisonic | This is a newly coined short which refers to the sum total of theoretical acoustics and electronic hardware development that once went simply by the name of Ambisonics (here, ambisonic). It has been introduced by the younger folks in the ambisonic scene, mainly because advancing technology has made HOA the center of interest and study. But at the same time, even the most avid enthusiast would still be introduced to ambisonic starting with a subset of POA, and would be hard pressed to acquire even a well functioning periphonic system. Thus, POA retains its place even know, and work continues to improve it, for example in the case of irregular speaker layouts. |
pressure field | TEMP!!! | TEMP!!! |
Q (channel) | TEMP!!! | TEMP!!! |
QS | TEMP!!! | TEMP!!! |
quadraphonic/quadrifontal | four channel/speaker | Quadraphonic is a term used to describe a system which attempts to produce sound images using four discrete transmission channels. Quadrifontal is a more specific term which also means that the four channels are being played back as four discrete speaker feeds around the listener. Usually at the corners of a square, which at a time seemed like a logical continuation of discrete two-channel stereo. But since the established transmission formats of the time did not permit four discrete channels, matrixing from four to two, and then back to four channels had to be used. When discreteness was then such an issue, all sorts of active decoders were utilized, quite without regard to the possibility of forgoing discreteness altogether, and simply utilizing the limited signal set to its best advantage. To a degree ambisonic was then born as the counter-reaction to this kind of limited thinking. |
R (channel) | TEMP!!! | TEMP!!! |
RACE | recursive ambiophonic crosstalk elimination | TEMP!!! |
RM | regular matrix | TEMP!!! |
Scheiber sphere | a spherical description of two channel matrix encodings | TEMP!!! |
SHJ | TEMP!!! | TEMP!!! |
SoundField (microphone) | the original ambisonic mic, and its derivatives | TEMP!!! |
soundfield | the whole acoustic environment around a listener | TEMP!!! |
spherical harmonic | TEMP!!! | TEMP!!! |
SPL | sound pressure level | A comparative scale of absolute sound pressure deviation, with the base unit of pascal, and a reference level of 20 micropascals deviation from standard atmospheric pressure, root mean square integrated over time. The most common use is in conjunction with the logarithmic decibel (dB) scale, which yields dB(SPL). To simulate the ear's sensitivity to different frequencies of sound, a weighting filter is often applied as well. That yields scales such as dB(A). |
SQ | stereo compatible quad | TEMP!!! |
steering | active matrixing | cf. active matrix |
subcardioid (pattern) | TEMP!!! | TEMP!!! |
super-stereo | ambisonically decoded stereo | TEMP!!! |
T (channel) | TEMP!!! | TEMP!!! |
TetraMic | a modern incarnation of the SoundField microphone | TEMP!!! |
THJ | TEMP!!! | TEMP!!! |
UHJ | the commonest ambisonic compatibility format | explain BHJ/SHJ/THJ/etc. hierarchical encoding. compatibility coding. TEMP!!! |
UMX | TEMP!!! | TEMP!!! |
variomatrix | active matrix | cf. active matrix |
VBAP | vector base amplitude panning | The extension of pairwise panning to three dimensions. This is a workable technique when absolute positions of sound sources are known, and the reproduction rig is dense. However, it does not work too well with recorded soundfields where the positions are not exact, nor does it adapt well to irregular reproduction rigs, since pairwise panning is only stable with closely spaced sources. |
velocity field | TEMP!!! | TEMP!!! |
velocity vector | TEMP!!! | TEMP!!! |
WFS | WaveField Synthesis | Ambisonic approaches soundfield synthesis by directional smoothing. Whenever channels are added, the directional response of the system becomes steadily more concentrated, accurate, and less spread out. In the end the system tends to holophony. So does WFS, but it starts differently and converges with different errors along the way. It starts with a large number of equidistant sound sources, usually placed along a line (cf. pantophony). It aims at directly realizing the so called "curtain of microphones", with some added processing to counteract the linear placement of the speakers. |
XY | crossed cardioids towards L/R | TEMP!!! |
This is a collaborative, non-profit effort, so we make it a point to acknowledge people's volunteerism as fully as possible. Everybody who has taken part in this project will be listed, but the form of listing is upto the individual.