Matrix H Articles

QuadraphonicQuad

Help Support QuadraphonicQuad:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Good on you Duncan. :)

The articles in my Matrix H book mention in the clippings that the BBC engineers and Mr Mackenzie preferred a 60 degree phase shift in the decoding of H , which differed over the encode which conformed to a 47.5 degree shift.
This was the end result with the BBC Engineers and Angus MacKenzie's preferred aural reception of the quadraphonic broadcasts.
When in doubt, always stick to the math. There were 2, maybe even 3 different decode equations that the BBC came up with for H during it's research and short run, then they went to different math for HJ when they punted on H. Also, there are most likely some small details about H and HJ that never made it to the tech papers, as by some accounts the BBC tech. folks were actively monitoring mix sessions when they were in progress. Thankfully they didn't leave us with too many recordings that need to be rescued.
 
Yes, at least two H variations. It seems the front shift changed late on from 47.5 to 48 as seen in one of the last 'H' papers 'Quadraphony: developments in Matrix H decoding' page 6, 1976. Rear phase required a shift of 90 degrees thus a mean shift of 69 degrees would have been optimal when using a standard QS decoder.
Around 1979 the BBC merged Matrix H with NRDC's Ambisonics System 45J. The new UHJ/BHJ system required a front shift of 30 and rear of 90. This resulted in the much quoted mean shift of 60 degrees used by the BBC on modified Sansui decoders used for test purposes.
 
Last edited:
Yes, at least two H variations. It seems the front shift changed late on from 47.5 to 48 as seen in one of the last 'H' papers 'Quadraphony: developments in Matrix H decoding' page 6, 1976. Rear phase required a shift of 90 degrees thus a mean shift of 69 degrees would have been optimal when using a standard QS decoder.
Around 1979 the BBC merged Matrix H with NRDC's Ambisonics System 45J. The new UHJ/BHJ system required a front shift of 30 and rear of 90. This resulted in the much quoted mean shift of 60 degrees used by the BBC on modified Sansui decoders used for test purposes.
It would be fantastic if someone who worked on the development of H/HJ at the Beeb were able to fill in some of the blanks for us. I expect if any were alive they would be quite old at this point. I still feel like there must be some pieces missing
 
It would be fantastic if someone who worked on the development of H/HJ at the Beeb were able to fill in some of the blanks for us. I expect if any were alive they would be quite old at this point. I still feel like there must be some pieces missing

This is the closest I've read so far as to thinking and work at the BBC re your blanks.

https://www.ambisonic.net/sfexp.html
And this Wiki page on Matrix H, which was simply the latest letter to be arrived at, presumably starting with Matrix A.

https://en.wikipedia.org/wiki/Matrix_H
I'd hazard a guess that various phase angle shiifs and speaker setups were tried out 'til H/HJ then finally UHJ.
 
I'd hazard a guess that various phase angle shiifs and speaker setups were tried out 'til H/HJ then finally UHJ.
Not really. The letters A through to H were the simply code letters used for different manufacturers' matrices in the BBB's subjective trials (see BBC report "The Subjective Performance of Various Quadraphonic Matrix Systems") and were as follows:
matrix table.JPG
 
I was being imprecise. The articles attached are useful to see what the BBC was trying to do, though the BBC really never said much directly. The author of the 'What were H and HJ?" article, Martin Leese, is particularly knowlegeable in this field and contributes in a surround list hosted at Virginia Tech, along with other 'gurus'. The list members concentrate on ambisonics (1st and higher order) and related surround techniques. The thing is, Matrix H cannot really be separated from UHJ as it was just a stop on the path to its development.
 
The articles attached are useful to see what the BBC was trying to do, though the BBC really never said much directly.
If you want to know what the BBC were thinking and doing then in addition to the matrix evaluation report I mentioned above ( BBC Research Dept. Ref. RD 1974/29) you should also read their detailed report "Quadraphony: developments in matrix H decoding" (BBC Research Dept. Ref. RD 1977/2).
 
If you want to know what the BBC were thinking and doing then in addition to the matrix evaluation report I mentioned above ( BBC Research Dept. Ref. RD 1974/29) you should also read their detailed report "Quadraphony: developments in matrix H decoding" (BBC Research Dept. Ref. RD 1977/2).
Yes, I've picked through this and the other docs. periodically. As hard as I try, my brain is not the greatest with math and as a result I have to bang my head against it over and over until eventually some things start to sink in. I know others that are way more comfortable with the tech. papers.
 
Yes, I've picked through this and the other docs. periodically. As hard as I try, my brain is not the greatest with math and as a result I have to bang my head against it over and over until eventually some things start to sink in. I know others that are way more comfortable with the tech. papers.

I'm with you there. As I started with ambisonics back in the early 80s, the general concepts are fine as I'm ok, more or less, with vector maths, wave propagation and spherical harmonic decomposition, which crop up in chemistry and physics quite a lot. Quad, to me, was always a bit of a dead end, so I never bothered with its maths.
 
To me quad was never a deadend , far more realistic than stereo . And that's why I bothered to support quad over stereo.
Don't get me wrong , I have listened to UHJ and do appreciate it's contribution to multi-channel. But for me 4 channels are impressive enough (at least for me).
Being that this thread is about quad matrices (H , HJ) I would hope we could stick to these particular quad formats and their overall performance and demontrative abilities in 4 channel.
Sorry I do not wish to offend anyone.
I'm well aware of the chronolgical development from H through to UHJ, but if we could possibly stick to the quad formats it may be beneficial , most especially to junior members of this forum.
Hope this does not rankle anyones feathers.
 
To me quad was never a deadend , far more realistic than stereo . And that's why I bothered to support quad over stereo.
Don't get me wrong , I have listened to UHJ and do appreciate it's contribution to multi-channel. But for me 4 channels are impressive enough (at least for me).
Being that this thread is about quad matrices (H , HJ) I would hope we could stick to these particular quad formats and their overall performance and demontrative abilities in 4 channel.
Sorry I do not wish to offend anyone.
I'm well aware of the chronolgical development from H through to UHJ, but if we could possibly stick to the quad formats it may be beneficial , most especially to junior members of this forum.
Hope this does not rankle anyones feathers.
Fizzy, I'm not offended in the least, and also welcome the idea of sticking with the 4 channel stuff, as it is hard enough for me to get a handle on the RM stuff. Ambisonics is way beyond my current level of comprehension, as much as I try to always be learning. The only reason for me that ambisonics comes into the conversation is because of HJ. It was a weird, messy transitional beast, and was supposedly able to be decoded with an ambisonic decoder as well as through a conventional matrix decoder like the one that the BBC offered. The problem is, it doesn't seem to work optimally with any of the approaches. Decoding may only be able to get to the "good enough" level.
 
Fizzy, I'm not offended in the least, and also welcome the idea of sticking with the 4 channel stuff, as it is hard enough for me to get a handle on the RM stuff. Ambisonics is way beyond my current level of comprehension, as much as I try to always be learning. The only reason for me that ambisonics comes into the conversation is because of HJ. It was a weird, messy transitional beast, and was supposedly able to be decoded with an ambisonic decoder as well as through a conventional matrix decoder like the one that the BBC offered. The problem is, it doesn't seem to work optimally with any of the approaches. Decoding may only be able to get to the "good enough" level.

My thoughts, too, more or less. It seems appropriate to place H & HJ in context.

(Now, if you ever want a mathematically really difficult job, just have a look at higher order ambisonics, which are now in use. I don't think anyone has seen 2.5, 3 nor 4 channel UHJ processors, let alone the maths. Current 1st order ambi processors only work at their best with UHJ (2 channel variety).)
 
All good points quadaholic.
It's just that........well , there is a totally separate section for :

UHJ.(Ambisonics)


I enjoy stevendive's and yours and soundfield's and Malcolm's comments regarding the progression from H - UHJ , but I was just kinda hoping there was more to share in the comments regarding four channel.
Particularly that of BBC H and HJ and the use and broadcasting of these quadraphonic formats.

Sorry but I think were just deviating away from the intention of this thread.(Four Channel Matrix)



Are there some quadraphonic(H /HJ) matrix experiences with the BBC Broadcasts everyone can share .
I know I have.
This could also include any commercial recordings.


My intention is to try and get this thread back on a quadraphonic information track.


:)
 
All good points quadaholic.
It's just that........well , there is a totally separate section for :

UHJ.(Ambisonics)


I enjoy stevendive's and yours and soundfield's and Malcolm's comments regarding the progression from H - UHJ , but I was just kinda hoping there was more to share in the comments regarding four channel.
Particularly that of BBC H and HJ and the use and broadcasting of these quadraphonic formats.

Sorry but I think were just deviating away from the intention of this thread.(Four Channel Matrix)



Are there some quadraphonic(H /HJ) matrix experiences with the BBC Broadcasts everyone can share .
I know I have.
This could also include any commercial recordings.


My intention is to try and get this thread back on a quadraphonic information track.


:)
got it, I'm muddling threads together, my bad.
 
These are the 6 graphs from the FCC Quadraphonic Tests .
Apart from Discrete 444 or 434 it shows how QS , SQ , and BBC Matrix H performed in 6 different listening situations.
I'm publishing it here as BBC Matrix H Actually performed quite well when compared to the other 2 matrices.
Both Matrix Encoders and Decoders were supplied by the various stakeholders , Sansui , CBS TATE , and BBC Matrix H. And were the latest and greatest Encoders/Decoders the Matrix backers had for early 1978.


20210825_105342.jpg
20210825_105352.jpg
20210825_105406.jpg



These FCC graphs are from : AUDIO MAGAZINE ...MAY 1978 !:giggle:

Needless to say the whole article is worth a good read .
 
https://worldradiohistory.com/Archive-All-Audio/Archive-Audio/70s/Audio-1978-05.pdf#page=28^^^
Monaural Compatibility
The rankings were identical to stereo: SQ, 4-4-4, H, and QS.


The major design objectives for SQ were mono and stereo compatibility, yet, Matrix H was developed years later with the same objectives (motivated by mono radio broadcasting needs).

(perhaps it's more of an NIH situation)


Was there ever a logic directed Matrix H decoder developed (I know a QS variomatrix decoder was adapted to decode H, but was it optimized for Matrix H decoding)?


Kirk Bayne
 
https://worldradiohistory.com/Archive-All-Audio/Archive-Audio/70s/Audio-1978-05.pdf#page=28^^^
Monaural Compatibility
The rankings were identical to stereo: SQ, 4-4-4, H, and QS.




Was there ever a logic directed Matrix H decoder developed (I know a QS variomatrix decoder was adapted to decode H, but was it optimized for Matrix H decoding)?


Kirk Bayne


I think Compcor incorporated QS Variomatrix with their BBC H . Perhaps even Lintek as well.(post #26)
And Sansui had a Matrix H adapted Decoder/Amp and Decoder with QS Vario as well . (See post # 45)
In addition to those Decoders , there were Decoders from Integrex and Catronics with BBC Matrix H .
(see posts #51 , and #41)

I saw a picture of the BBC Matrix H Decoder (the prototype) , and I don't think it had any logic circuitry, and I'm sure that would be what was supplied for FCC Testing Of BBC H .

I myself used a Sansui QSD-2 with a scratch built adapter that had BBC H , BBC HJ and a moderate UHJ .
Note the FCC did not test HJ , or so the quad matrix testing shows .
 
Back
Top