Ambiosonic UHJ

 

Ambiosonic UHJ Audio

Mike Gerzon’s Ambiosonic UHJ encoding scheme is completely Stereo compatible, but also delivers height and depth and rear channels correctly for 6 channels. It was taken up by the UK NRC (National Research And Development Corporation) in 1977, and this effectively killed that exciting possibility before CDs had really been launched....

The Calrec Soundfield microphone was and another Gerzon idea, and used a tetrahedral array of microphones in a single unit. It produces excellent results even when the rich UHJ output is never exploited.. it is the sole microphone used in the University of Melbourne Music Schools’ Melba Hall, and the Senior Technical Officer (Les Craythorn) still has one of his own.. dedication to high quality sound capture in person!.

Here are some my own records and pieces of UHJ Equipment: Clicking on each title will open a full new page with the details

  1. 1.1977 1122 NRDC Ambiosonic UHJ Specification Document front page

  2. 2.Troy Car Ambiosonic UHJ Unit front view

  3. 3.Troy Car Ambiosonic UHJ Unit rear view

  4. 4.Troy Car Ambiosonic UHJ Connections detail view

  5. 5.Troy Car Ambiosonic UHJ Unit Information Leaflet front page

  6. 6.Troy Car Ambiosonic UHJ Unit Information Leaflet Inner pair of pages

  7. 7.Troy Car Ambiosonic UHJ Unit Information Leaflet second set of inner pages

  8. 8.Onkyo 909: the last major AV amplifer and control unit to have UHJ included - manual front page

  9. 9.Onkyo 909: the last major AV amplifer and control unit to have UHJ included - UHJ Setup page

  10. 10. MRW-built Integrex Ambiosonic UHJ and B matrix unit - front view

  11. 11. MRW-built Integrex Ambiosonic UHJ and B matrix unit - rear view

  12. 12. MRW-built Integrex Ambiosonic UHJ and B matrix unit - front page of the construction details

  13. 13. MRW-built Integrex Ambiosonic UHJ and B matrix unit - first page of Gerzon’s complete description of the unit

When the opportunity arises, full scans of these materials will be posted here, with further links on UHJ, Ambiosonics and Michael Gerzon, whom I had the privilege of knowing for many years, and (successfully) advising him (and supporting it indirectly with raid and forceful lobbying of the appropriate influential parties myself) to overcome his 3rd Class Honours Degree as a barrier to joining the Oxford University Mathematical Institute. We spent many happy hours (more like days) talking about audio, subjective perceptions of sound and music, coding standard interchanges and many other things that later came to fruition in our small flat in Stratfield Road, Summertown, in North Oxford in the 1960’s. We both often walked up to Geoffrey Horn’s excellent HiFi shop at the end of the road.. and both he and Mr Tandy showed quite remarkable patience with us both over quite some years. We even occasionally bought something in those poverty stricken years..

The loss of Ambiosonic UHJ encoding on all CDs was one of the earliest reversals for Ambio...and was due almost entirely to NRDC’s unhelpfulness as I recall ... others including the banning of mentioning in the Radio Times that specific BBC broadcasts were being done in UHJ... Nimbus Records (who by then held the copyright on the UHJ symbol) became the last repository of major recording use of UHJ, and these recordings were of extremely high stereo quality - sadly (as recently advised by Steven Dive) the UHJ icon did not betoken the full height, depth and surround sound encoding on their LPs and CDs, explaining why I could not get it out from my Integrex, Troy or Onkyo909, or any other decoder...although I have found that Nimbus mixed down to (excellent) stereo from their original B matrix recordings.. Steven also confirmed my assessment that full height and depth periphonic recording simply could not sensibly have been crammed into LPs. without unacceptable losses. I have a number of Nimbus LP, CD and cassette format recordings (virtually all from Nimbus). There is still an active market for Ambisonic UHJ equipment and recordings and also here.

The actual Nimbus recording processes used are cited by Steven Dive as follows (source, email 9-2-15):

Nimbus has only ever used their own Nimbus-Halliday pantophonic (XY only) microphone rather than the Calrec periphonic ones, so the Z component was never included. A long time ago Prof. Peter Fellgett sent me some papers that included a very poor photocopied photograph of the mic. It is a native B-format mic consisting of an omni (W) and two figure 8 (X & Y) capsules. You won’t find anything other than the two channel UHJ (BHJ) channels on the CD’s from Nimbus and, therefore, no height component, unfortunately. In any case, the AIFF file type on standard CD’s does not support more than 2 channels as far as I know. 

Any LP’s pressed with UHJ only have ever had the horizontal UHJ encoding on them (e.g. Nimbus records in the early 1970’s). The UHJ icon only ever betokened horizontal reconstruction for all commercially released material. I imagine in principle so-called quadrophonically pressed LP’s could have been used to cram the full 3D 4-channel UHJ on them but I can’t imagine how this could be done with sufficient accuracy to recover a satisfactory reconstruction of the soundfield (see Steven Dives comments confirming this above). It did not work well for quad and it is very unlikely it could have worked for ambisonic 3D.


Three channel UHJ only supports horizontal decoding but without the direction related phase errors of the 2 channel version. The forth is needed for height. Mathematically, any 2-channel encoding, thus including good ol’ stereo, has direction related phase errors built into it. Two channel UHJ design tackles the problem head-on by redistributing the errors progressively and less noticeably towards the rear. The Minim AD10 had an adjustment for forward dominance that used this feature.“ [Steven Dive (source, email 9-2-15)]


My Troy car system was very effective, and I still have mine (and the base extension unit).

The Integrex was a kit designed by Mike, and I built one (see above) I will add formal references to the papers that go into full detail when i can locate them again.

My Onkyo 909 was the last modern AV amplifier known to have a UHJ setting (please correct me if you read this and know of a later one)

This page will be expanded as resources and time permit, and add other dimensions... I clearly cannot match the University of York’s splendid website at http://www.ambisonic.net, or Martin Leese’s at http://members.tripod.com/martin_leese/Ambisonic/ , but nevertheless will try to bring useful key items together on this page in the hopes it gets more people interested in this form of encoding - on which I was advised in the early 2000‘s that the BBC is once again active in researching... on this aspect, I introduced Mike Gerzon to my father (Edmund Ramsay Wigan, the sound quality specialist at the BBC Research Labs in the 1960s, and an authority on psychoacoustics*), as his first link to the BBC ...

e.g. I hope to establish how much of the huge Gerzon recordings Archive held at the British Museum are recorded in UHJ. An excellent article explaining his approach is available, in a private open gerzon archive, reprinted from Wireless World, December 1974.

One aspect is that other forms of Quad recordings can be transcoded into Ambio/UHJ, and making UHJ files from ‘ambisonic’ recordings also requires some care.. a good summary is at http://www.radio.uqam.ca/ambisonic/b_uhj.html.

  1. E. R. WIGAN (1959) Unpleasantness of Distorted Sounds: A Criterion derived from the Distortion Spectrum. Nature 183, 1320-1320 doi:10.1038/1831320a0. Download this <Here> (permission to publish secured from Nature)


While commenting on the ground breaking work of my father at the BBC Research Labs, I must add this invaluable reference to the actual processes of sound and note detection that are still not fully appreciated, over 50 years later


  1. S. BLACK and E. R. WIGAN (1961) An investigation of selective deafness produced by direct suggestion under hypnosis. British Medical Journal  Sept 16, 736-641.  Download this paper <Here> (Formal permission secured from the BMJ )

Updated and Edited on 27 February 2016