SQ Shadow Vector Soundfield Mapping

[Soundfield]

An update on progress ))
The decoding software is now working well and I'm very happy with a decode of several DSOTM tracks performed yesterday. The sound is nothing like a conventional single or 3 band decode with complete absence of pumping artifacts and solid positioning of instruments. It turned out best to filter vector modification across all 10 bands to around 5-7 positions/sec. Decoding accuracy is better than a few decimal places from 10Hz - 20kHz although that is with test tones and real life especially with vinyl is never that good. I've also implemented dynamic phase correction, which eliminates the traditional 180 degree 'hole' thus all sound sources between speakers are in phase.
Next job is to eliminate a very low level noise from the vector modification filters and implement QS followed by Matrix H.

I'm very impressed Malcolm! Is this all for you personal amusement or are you working towards something saleable?

 

[malcolmlear]

The aim is a saleable hardware decoder for enthusiasts. Clearly a small market, but I've been working on this out of love for the old quadraphonic systems and have always felt them to be underrated due mostly to the early terrible decoders.

 

[malcolmlear]

Here's a 1kHz test tone decode. Results are identical from <20Hz through to 20kHz.
Note the use of a new computer generated set of tones, the old set supplied by our friend OD were not accurate.

 

[Soundfield]

The aim is a saleable hardware decoder for enthusiasts. Clearly a small market, but I've been working on this out of love for the old quadraphonic systems and have always felt them to be underrated due mostly to the early terrible decoders.

Well there's certainly a few such enthusiasts here who are always looking for the best possible means of decoding their quad treasures in real time (none of that tedious mucking about with scripts!). The prospect of a Shadow Vector decoder would be a very welcome addition to the market. Any idea of a timescale? Good luck with the venture.

 

[Soundfield]

... the old set supplied by our friend OD were not accurate.

There's a surprise!

 

[malcolmlear]

I've tried to attach the new SQ test tones, but unfortunately zipped up the 1.5MB size exceeds the upload limit.

 

[malcolmlear]

Timescales are difficult, but I expect realistically we are looking at the early part of next year. I'm looking at a black anodized metal enclosure with a full colour OLED display which initially will just display the decode system logo's and later possibly a simple vector scope. The possibility of a vector scope obviously would be highly dependent on processing resources available after decoding which at the moment is pretty heavy.

 

[Soundfield]

I'm looking at a black anodized metal enclosure with a full colour OLED display which initially will just display the decode system logo's and later possibly a simple vector scope.

Appetite suitably whetted! I guess postage and packing to a fellow Essex resident won't be too expensive!

 

[fredblue]

hi malcolmlear
this is all sounding very exciting!
will your new SQ decoder remove/cancel out vocals that have bled from Front to Rear?
many thanks and best of luck with the whole endeavour,
Adam

 

[malcolmlear]

Hi Adam,
Very interesting question. During development I kept hearing vocals in the back channels, at least in the few albums I had available and considered it a problem. However a colleague did a Fourier analysis of these tracks and found they were actually encoded off the cardinal axis. In other words the encoding placed the vocals away from the periphery and somewhere in the centre of the soundfield. I can only guess whether this was intentional or an artifact of early encoders. I really need more decoding material to determine whether this was normal mixing or just bad encoding.

 

[sjcorne]

Hi Adam,
Very interesting question. During development I kept hearing vocals in the back channels, at least in the few albums I had available and considered it a problem. However a colleague did a Fourier analysis of these tracks and found they were actually encoded off the cardinal axis. In other words the encoding placed the vocals away from the periphery and somewhere in the centre of the soundfield. I can only guess whether this was intentional or an artifact of early encoders. I really need more decoding material to determine whether this was normal mixing or just bad encoding.

What albums are you using? Aside from a few exceptions, almost every single quad mix ever done by Columbia has the lead vocal center front with no leakage to the rear whatsoever.

That’s so interesting that the SQ encode may have inadvertently messed up the vocal position- it would explain a lot about why cancelling vocal bleed to the rears is so difficult.

 

[malcolmlear]

The problem may be due to the various encoding methods used. The best in my opinion was the 16 channel positional encoder. Here up to sixteen totally independent audio sources could be placed anywhere in the periphery, perhaps even off periphery ? However I think many albums were created by feeding premixed disceet 4 channels into a 'forward' or 'backward' orientated encoder. This was considered acceptable at the time but was a really bad way of doing things. The problem is that the instrumental/vocal raw sources had already been mixed into the 4 dicreet channels and when fed into the basic SQ encoder had issues due to common phase into multiple inputs.

 

[Soundfield]

I’ve just listened to Lynn Olson’s 1975 NYPR radio interview (sadly not very technical) about the potential Shadow Vector decoder:
https://www.wnyc.org/story/the-shadow-vector-decoder/
Interesting that he felt that advanced matrix decoding would soon match or exceed the perceived performance of discrete tape! It is clear that he felt that any decoder that achieved improved separation at the expense of the accurate reproduction of reverberation was a poor decoder (by which single measure alone most decoders of the day were very poor decoders!)

He also talks about the importance of psycho-acoustics to the improvement of quad but I’ve read his US patent 4,018,992 and I don’t see that there’s any specific psycho-acoustic processing going on over and above the clever dynamic multiple axis decoding technique, perhaps I’m missing something.

A lot of the interview concentrated on the interest in quad synthesis and what a good synthesiser the Shadow Vector Decoder would be, with Lynn saying it would out perform the Sansui, having continuously variable dimensional control. Again I don’t see any such functionality in the patent. Will your decoder provide a synthesiser mode Malcolm?

It is a shame that this technology was killed off when it was on the verge of transforming the performance of quad in the home.

 

[malcolmlear]

I think the basic function of the shadow matrix could be described as having continuous dimensional control and the QS version would no doubt provide an excellent stereo to quad synthesis mode. It's such a shame that Lynn's smart solution that could have saved matrixed quad was pushed aside in favour of the Tate method, although to be fare it my have just been too late make a difference.

 

[malcolmlear]

Finally obtained the great Chase-Chase album in SQ format. There was about 10% out of phase crosstalk between the stereo channels. Fortunately the decoder can deal with that, albeit a manual setting. It could be the copy I have is taken from a poorly set up vinyl, very bad if its a CD created from the SQ master tapes. Anyway here is the amazing trumpet intro:

 

[malcolmlear]

Also dynamic phase correction is now perfected. Note how the LF(top), LB(3rd down) and RB(bottom) form a common in-phase sound, whereas LF and RF(2nd) share a lower frequency in-phase signal. This helps greatly in producing clear well defined images, something previously limited to discrete sound systems.

 

[fredblue]

Finally obtained the great Chase-Chase album in SQ format. There was about 10% out of phase crosstalk between the stereo channels. Fortunately the decoder can deal with that, albeit a manual setting. It could be the copy I have is taken from a poorly set up vinyl, very bad if its a CD created from the SQ master tapes. Anyway here is the amazing trumpet intro:

separation looks amazing!!!!
if you would like any SQ encoded samples for testing purposes do let me know.

 

[sjcorne]

separation looks amazing!!!!
if you would like any SQ encoded samples for testing purposes do let me know.

Agreed- wow!

It might be a good idea to use some “torture test” SQ LPs to really test the separation, such as:

• “2001” from Deodato's Prelude has one of the rear channels is dead for most of the track and occasionally bursts in
• “Night Game” from Paul Simon's Still Crazy also has very subtle use of the rears, one channel is not used for much of the song
• The entire Aerosmith Toys In The Attic album: it's a really strange mix for Columbia, there's a lot reverb and often silence in the rears
• "That Lady" from the Isley Brother's 3+3 has the lead guitar positioned diagonally

Also, do you have the SQT-1100 test LP? It has some handy test tones for checking rear center, dead center, and side center positions, which are generally difficult for the matrix to handle.

 

[malcolmlear]

Many thanks for the offer fredblue. There still seems to be a very low level noise issue from the vector filters so maybe good quality classical tracks could help track down the problem. I could also send back decoded files in wave or DTS for comment.

 

[fredblue]

Many thanks for the offer fredblue. There still seems to be a very low level noise issue from the vector filters so maybe good quality classical tracks could help track down the problem. I could also send back decoded files in wave or DTS for comment.

maybe one of the Masterworks SQ LPs that contain Quad material Vocalion have reissued on Surround SACD?
something like the Pulcinella Suite could be useful to have the discrete masters to compare to the decodes?
pretty sure i have the SQ LP of that one.