INVOLVE SQ - IS COMING

QuadraphonicQuad

Help Support QuadraphonicQuad:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Hi Soundfield

Still wallowing in a swamp of poo.

At present I have not looked too deeply into the variants you listed - frankly Dolby PL2 and circle surround (by SRS now DTS) are also SQ variants. As I mentioned we are trying a few "known" techniques of logic decoding SQ but truthfully I am not impressed yet. It is hard to get an even surround from an inherently non even encode such as SQ.

Having said that in frustration I went deeply into "my happy place" last night and stared blankly into a bunch of equations and vector diagrams for a few hours and walked out "talking in tongues" with a possible concept to get a better decode. We will trial it this week.

Regards

Chucky

Thanks for the update Chucky. Whilst not wishing to add unnecessarily to the depth of poo.... have you given any consideration to the various 'flavours' of SQ encoding that (albeit briefly) existed? I'm thinking of the different soundstage emphasis that arose from the the fairly arbritrary adoption by different producers of the so called "Forward Oriented", "Backward Oriented" or "London Box" encoding regimes. Having said that, I'm not sure if any decoder is actually capable of detecting which particular type of encoder may have been used, let alone whether any diffferent decoding strategies might be called for to optimise uniquely for any of them, so I'd be interested in your thoughts.
Cheers.
 
Chucky - I recognise that SQ is inherently (and at CBS' request, deliberately) flawed from a true surround perspective, and any performance improvements will be a real struggle (I'm thinking you are in silk purse and pig's ear territory here) so good luck with your experiments. Please keep us informed. Cheers.
 
Chucky - I recognise that SQ is inherently (and at CBS' request, deliberately) flawed from a true surround perspective, and any performance improvements will be a real struggle (I'm thinking you are in silk purse and pig's ear territory here) so good luck with your experiments. Please keep us informed. Cheers.

Well, as far as matrix systems go, SQ is better than QS(better LF-RF separation).
I have personally achieved outstanding results from software decoding (about 30 dB channel separation)...e.g., PF , WYWH which I prefer my SQ LP conversion to the Immersion box discs, which are edited and sound...lifeless...
Also, tha 1st Santana LP DVD-A is incredible!!!


BTW, love your avatar....
 
Chucky- please don't forget to look at/incorporate techniques used in the tate units. Some of the original designers of that unit occasionaly show up here on Quadraphonicquad.
 
Thanks for the update Chucky. Whilst not wishing to add unnecessarily to the depth of poo.... have you given any consideration to the various 'flavors' of SQ encoding that (albeit briefly) existed? I'm thinking of the different soundstage emphasis that arose from the the fairly arbitrary adoption by different producers of the so called "Forward Oriented", "Backward Oriented" or "London Box" encoding regimes. Having said that, I'm not sure if any decoder is actually capable of detecting which particular type of encoder may have been used, let alone whether any different decoding strategies might be called for to optimize uniquely for any of them, so I'd be interested in your thoughts.
Cheers.

You'd probably have to contact Martin Willcocks who designed the original Tate ICs to get into that level of SQ paramatrix decoding. He was the one who raised the possibility that future SQ decoders could be tweaked by recognizing and then adapting their paramatrix decoding based on the encoder type used as you suggest. Not sure anyone else has explored that aspect of SQ processing.
 
Hi All

I thank you all for the information on prior attempts on SQ decoding, I am reading as much as I can about it. It looks like the approach we will attempt this week is different to Tate and Sony and we hope it will be successful.

As mentioned previously, when we are satisfied that the system is working well we will issue the proto sample to Ron (Rustyandie) and to any other QQ member that wants to be our "test bunny" for comment. Meanwhile I really would appreciate any detailed information on SQ decoding as I do not want to waste time " reinventing the wheel" although my preference is finding a new approach!

Regards

Chucky
 
Thanks for the great updates! I know SQ is complex, but you guys seem to have great ideas! I'm so excited. I bought 7 more SQ records yesterday. My SQ collection is starting to get pretty impressive.
 
Soundfield said:
Thanks for the update Chucky. Whilst not wishing to add unnecessarily to the depth of poo.... have you given any consideration to the various 'flavours' of SQ encoding that (albeit briefly) existed? I'm thinking of the different soundstage emphasis that arose from the the fairly arbritrary adoption by different producers of the so called "Forward Oriented", "Backward Oriented" or "London Box" encoding regimes. Having said that, I'm not sure if any decoder is actually capable of detecting which particular type of encoder may have been used, let alone whether any diffferent decoding strategies might be called for to optimise uniquely for any of them, so I'd be interested in your thoughts.
Cheers.

Are those really different types of encoding or recording styles that attempt to maximize the desired effect on the poor quality decoders of the time? If so, there is no real need to change much for those recordings. If they sound funky through a modern decoder, the levels could be adjusted by the listener. Regardless of what things the engineer did to make it sound better, I always thought that the actual encoding to a stereo matrix was done through pretty much the same process (allowing for hardware variances).
 
Are those really different types of encoding or recording styles ?

Nothing to to with recording styles Q8, these were different encoding regimes (i.e different matrix parameter values were employed in the encoder) in an attempt to variously improve stereo separation, improve mono compatibility for broadcasting purposes or, the case of the "London Box", both.
 
Last edited:
Thanks ! I didn't know about that. Was there a large number of recordings that used different encodings? Also, even though they fudged the encoding perameters, I'm guessing they didn't expect the listener to be using anything but a standard SQ decoder did they? Do the recordings sound good through a tate or the scripts?
 
Thanks ! I didn't know about that. Was there a large number of recordings that used different encodings? Also, even though they fudged the encoding perameters, I'm guessing they didn't expect the listener to be using anything but a standard SQ decoder did they? Do the recordings sound good through a tate or the scripts?

These are all excellent questions Q8, but sadly ones to which I have no answers!
 
Dear All

Just for my interest and knowledge does anyone know if the original Sony logic SQ decoders were tri band like the Sansui QSD-1? Also I believe Tate used tri band - am I correct.

Pardon my ignorance but we really prefer to do our own thing!

Hope to hear your answers.

Regards

Chucky
 
chucky3042 said:
Dear All

Just for my interest and knowledge does anyone know if the original Sony logic SQ decoders were tri band like the Sansui QSD-1? Also I believe Tate used tri band - am I correct.

Pardon my ignorance but we really prefer to do our own thing!

Hope to hear your answers.

Regards

Chucky

I'm not familiar with tri band. But do your own thing. Beat the Tate. Make a video scope comparison with the Tate vs. The Surround Master like you did with the QSD-1 if rusty or someone has a Tate. I know you will blow our minds. Thanks guys.
 
Dear All

Just for my interest and knowledge does anyone know if the original Sony logic SQ decoders were tri band like the Sansui QSD-1? Also I believe Tate used tri band - am I correct.

Pardon my ignorance but we really prefer to do our own thing!

Hope to hear your answers.

Regards

Chucky

Hi Chucky -
No Sony did not produce a tri-band SQ decoder. The SQD2020 (mine is pictured in my avatar) is about as sophisticated as Sony got and it is resolutely a single band device. In reality it is nothing more that a simple SQ matrix decoder which sits in an block of voltage controlled amplifiers operated by 'wave matching' and 'front to back' logic circuits in an attempt to dynamically change channel gains in response to the predominant signal direction (with success that ranges from poor to moderate!). It is tremendously complicated for what it does and a tri-band version would have been out of the question on cost grounds alone (apart from the fact it would have probably have been all but impossible to set up and keep all those analogue circuits aligned). I'd be interested to know if tri-band processing actually has any mathematical basis for improved performance in any type of decoder - I'm struggling to understand why it would.
 
Hi Chucky -
No Sony did not produce a tri-band SQ decoder. The SQD2020 (mine is pictured in my avatar) is about as sophisticated as Sony got and it is resolutely a single band device. In reality it is nothing more that a simple SQ matrix decoder which sits in an block of voltage controlled amplifiers operated by 'wave matching' and 'front to back' logic circuits in an attempt to dynamically change channel gains in response to the predominant signal direction (with success that ranges from poor to moderate!). It is tremendously complicated for what it does and a tri-band version would have been out of the question on cost grounds alone (apart from the fact it would have probably have been all but impossible to set up and keep all those analogue circuits aligned). I'd be interested to know if tri-band processing actually has any mathematical basis for improved performance in any type of decoder - I'm struggling to understand why it would.

I do not know why but
the QSD1 and the Surround Master has sharper decoding
than the QSD1000 a single band unit
or the QSD2 also single band
 
Thanks Soundfield and Rustyandie

Still would like absolute confirmation if Tate used multi band.

I answer to your good question of

I'd be interested to know if tri-band processing actually has any mathematical basis for improved performance in any type of decoder - I'm struggling to understand why it would.

I will slightly lift our skirts by saying the encode process for SQ and QS is a single band process and is very valid across the full frequency spectrum. The big problem is how you unscramble the encode. With both SQ and QS you only have a few dB of separation to start working with. The main trick used by most clever decoders is to pick direction (or magnitude) dominance as the human ear has trouble deciphering two loud signals simultaneously (we are simple creatures - more like frightened bunnies only able to directionalize on the first loudest sound - one at a time).

One big problem of simple picking magnitude dominance is that you can get simultaneous events at different parts of the frequency spectrum, in addition sometimes the human ear perceives the quieter sound as being louder (refer the Fletcher Munsen curve) dependent on frequency. In addition if you look at a magnitude plot of music amplitude vs frequency you will see it mimics pink noise with a huge magnitude dominance of lower frequencies (translation is that the higher frequency separation will be incorrectly steered by lower frequency bands due to their magnitude dominance).

As you can see its all a complicated mess and the Surround Master has incorporated multiple strategies to allow for all the above issues (and a whole bunch of other issues). One of the big weapons we have is the use of 3 separate processor paths in different frequency bands 20 - 300Hz, 300 - 3 Khz and 3 - 20kHz. partially ensuring (with other tricks) that each band is treated on its own merits without amplitude corruption from other bands.

Rustyandie's observation of

QSD1 and the Surround Master has sharper decoding
than the QSD1000 a single band unit
or the QSD2 also single band


Is quite accurate and I hope what I have discussed shines some light on the issue (I am still keeping a bunch of secrets !).

Regards

Chucky
 
Oops, forgot to add.

In one of our earlier test reports on the QS decoder (prototype level) one of the tests was separation of simultaneous tones in different frequency bands. If you look at pages 4 to 10 you will see a lot of hard to understand graphs that show say 20 - 30 dB separation for simultaneous inputs at different frequencies. If a single band decoder the best you would get would be around 3 - 4 dB.
 

Attachments

  • DSP implementation test results V2.pdf
    167.7 KB · Views: 163
Hi All

Just a quick update. Work stopped on the SQ about a week ago just as we were circling the solution! Yes I am frustrated but we had to swap projects for a week or two to our new "Surround Cube woofer/ 6 channel amplifier/ preamp/ Involve" Woofer. I will publish some concept shots for you all shortly.

I hope we will be back on the job in a week or two.

Regards

Chucky
 
OK Chucky , I guess we all understand this isn't your top priority - we look further to further updates when work resumes!
 
Back
Top