Using the Precision MS Matrix for Mono Crossover LF Blending

Where we discuss new analog design ideas for Pro Audio and modern spins on vintage ones.
User avatar
mediatechnology
Posts: 5442
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: Using the Precision MS Matrix for Mono Crossover LF Blen

Post by mediatechnology »

I'll post links to the full schematics.
I want to re-tile the E77.

I'm finding out what a "Swiss Army knife" the MS board is.

It looks like everything up to the VAB84 was 6 dB/octave and by virtue of blending cancelled LF Side information rather than fold it into Mono/Mid like the VAB84.
I need to find some VAB84 schematics to see what the filter order was.
I think there's a case for 12 or 18 dB per octave and folding to mono particularly when high vertical/lateral crossover frequencies are needed.
User avatar
mediatechnology
Posts: 5442
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: Using the Precision MS Matrix for Mono Crossover LF Blen

Post by mediatechnology »

VAB84 Description and Schematics

I found the English technical description and schematics for the VAB84 and it confirms the different method it uses over prior generations of "Elliptical Equalizers."

Image
Nuemann VAB84 Block Diagram Showing Low Pass-Filtered "Side" Converted to "Mid" By Steering Into Left and Right Using Summation.


Image
Nuemann VAB84 Description. The sliding filter is 6 dB per octave.
"Flank signal components are simultaneously transferred to the other channel."


Just what I had thought. Side gets converted to Mid via steering.

This reduces phase cancellation effects and provides a greater degree of mono compatibility.

Image
Nuemann VAB84 Description. The sliding filter ranges from 0.1 to 1000 Hz.

The VAB84 moves the filter into the subsonic to get it out of the way when it is not needed.
Apparently, with a 6 dB per octave slope, Neumann engineers saw the need to get it out of the mid-band when not needed.
No re-matrixing of the Left and Right signal is necessary.



Image
Nuemann VAB84 Description. Limiter Specifications.

The pdf with schematics are here: http://www.waynekirkwood.com/images/pdf ... nglish.pdf
File courtesy of: http://audio.kubarth.com/

I've found at least two ways to emulate a VAB84 fixed frequency filter with the MS Matrix board.
I'm trying to find the best configuration before showing it.
I don't know if Neumann explored higher-order filters.
With this topology however higher-order is possible and quite likely desirable to the ME since it does not compromise mid-band separation, may not require a dynamic sliding filter just a "static" adjustable one, and is mono compatible.
Things that require heavy low frequency blending won't drop out of the mix because instead of being cancelled they get folded down to mono.
The EE70 and EE77 don't do that: The VAB84 does.
User avatar
mediatechnology
Posts: 5442
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: Using the Precision MS Matrix for Mono Crossover LF Blen

Post by mediatechnology »

I found a second way to do this externally as an insert to the MS matrix.
The circuit I'm testing right now is second order with a high cutoff.
There is virtually no separation below 200 Hz; at 1 kHz its about -25 dB.
On a lot of material - already mastered - I can't hear it working which I suppose is a good thing.
One tune I found, "Carry On" by Crosby Still Nash and Young, was one I had transferred from vinyl.
You may recall that the intro to that song has acoustic guitar panned hard right with no elements panned left.
That song was always a good test of separation.
The transfer has a huge low frequency warp (vertical) component that bounces the vector at LF visually left and right into quadrants two and four.
On headphones you hear the LF mud, some of it outside the head.
Using the "elliptical" second-order equalizer it disappears without sonic penalty and the warp wiggles in the vectorscope "elliptical" image disappear.
It seems the tool has application both for cutting transfer as well as playback transfers.

Here's a video of the before and after vectorscope patterns: https://proaudiodesignforum.com/content ... rscope.wmv (3.7 MB)

I realize that this is a playback transfer but it illustrates how low-frequency out-of-phase material is processed to eliminate the vertical component.

Image
VAB84 Emulator MS Matrix Crosstalk Curve 12dB Per Octave

Image
VAB84 Emulator Frequency Response
User avatar
mediatechnology
Posts: 5442
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Low Frequency to Mono Using Side to Mid Transformation

Post by mediatechnology »

The schematic below shows how the decode section of the Precision MS Matrix board can be used to emulate the Neumann VAB84 with a fixed frequency filter and optional increased slope.

The VAB84 and this implementation are both different from the EE70 and E77 which eliminate low frequency difference components by subtraction.
The VAB84 and the circuit below provide Side to Mid transformation using addition to increase mono compatibility.

Image
Neumann VAB-84 Fixed Frequency Emulator Using the Precision MS Matrix

A large printable version of the drawing shown above is here: https://proaudiodesignforum.com/images/ ... Matrix.JPG
The MS Matrix Schematic is here: http://www.ka-electronics.com/Images/jp ... _Large.JPG

The decode section is used because it has a dual op amp summation stage following the decode matrix.

In this implementation Left and Right are never broken down and are passed through to the output.
The Mid output is not used.
In the schematic shown above, the inputs, outputs and inserts inside the block correspond to the PC board's silkscreen designators.
The designators outside the box are the actual signals "Left, Right, Side Out" etc.

Op amp A is a 12 dB/octave second-order low pass filter for the side output.
1/2 The output of op amp A - the low frequency "vertical" or difference signal - is summed with Right.
The output of A is inverted by op amp B.
1/2 the output of op amp B is summed with Left.

You can calculate other filter values here: http://sim.okawa-denshi.jp/en/OPstool.php
And round the resistors to E96 values here: http://www.logwell.com/tech/components/ ... alues.html


The result is a crossover to mono below the cutoff frequency of the filter with a 12 dB per octave slope transitioning to stereo.
Because the difference signal is summed into Left and Right it is transformed from Side to Mid and remains in the mix as a mono element.


Note that polarity differences are very important.
If a higher-order filter is used with a topology that has overall signal inversion, then the polarities of the signal feeding left and right need to be reversed.
If they are not, comb filtering will occur at the cutoff frequency.

The EE70 and EE70 elliptical equalizers filter out the difference signal by blending and the LF stereo elements do not remain in the mix at all.
The EE70 and EE77 also have a fixed 6 dB/octave slope.


An MS Matrix having it's Side high pass filtered will emulate an EE70 and EE77.
Its slope will be 6 dB/octave regardless of the filter order because the resulting filtering is subtractively derived when re-matrixed to reconstruct Left and Right.


To emulate an EE70 or EE77 with the MS matrix HP filter side.
To emulate a VAB84 with a fixed filter and increased slope use the circuit posted above.
Two things contribute to the VAB84's transparency: Many people credit the sliding filter but IMHO a big difference is the Side to Mid transformation.
Gold
Posts: 677
Joined: Sat Dec 01, 2007 5:20 pm

Re: Using the Precision MS Matrix for Mono Crossover LF Blen

Post by Gold »

Thanks for alerting me to this thread Wayne. I'm late to the party so forgive me if I've missed something. The reason it is called an elliptical filter is because of the groove shape it makes not the filter topology. It narrows the groove for uncorrelated low end.

You could use the M/S boards for this but I think the Neumann circuit is more elegant. You don't have to go through the M/S conversion.

The VAB 84 is a dynamically controlled HPF. I haven't studied the schematic mostly because I've heard it. It's some fancy circuit building but I think it is more obtrusive than a simple EE. I have the full document package for the VAB84. It's 20 pages so I'll send it to Wayne via email and maybe he can post relavent bits. You circuit guys should have a blast as this is a beast.

Another random comment is that although the Neumann EE's are the most common others made them in a different way. I saw a Neve mastering console that let you choose the slope of the roll off as well as frequency. I think others attempted to add low end to the center channel to make up for the side channel loss. These weren't as successful in the real world.
Gold
Posts: 677
Joined: Sat Dec 01, 2007 5:20 pm

Re: Using the Precision MS Matrix for Mono Crossover LF Blen

Post by Gold »

BTW I still have two extra EE boards that Roger laid out. If anyone wants a unit let me know. I'm not selling PCB's and I'm not ordering anymore.

Here are some pictures of the finished unit. The boards work but I still have to test the final unit and tack a couple of wires in. The console has separate power supplies for preview and modulation paths. If I was thinking ahead I would have had the power done separately for each stereo pair and had jumpers to join or separate. I had to use four boards for the A and B paths.

Here are some pictures of the finished unit.
Image
Image
Image
User avatar
mediatechnology
Posts: 5442
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: Using the Precision MS Matrix for Mono Crossover LF Blen

Post by mediatechnology »

Thanks Paul for checking in.
I'll pore over those docs.

What I think you may have missed in this thread as it eventually evolved is that the technique I'm showing here doesn't actually use MS.
It just uses the MS board. In fact Mid isn't used at all.
The first experiment I ran was to HPF Side in MS.
The VAB84 technique worked better.

As Neumann wrote about the VAB84 ""Flank signal components are simultaneously transferred to the other channel."
So if the Bass elements that are in Side have some artistic value to the Mix they're still heard, not cancelled by blending.
Side info in right gets transferred to left and vice-versa.
The EE-series simply cancel the side component by allowing it to self-cancel.

Left and Right stay in their respective domains: Only Side is derived so it can be folded back into Left and Right without the Bass loss.
The Bass in Side gets put back into L and R or more specifically L+R.
I think the VAB84 sliding filter may have been chosen as a technique because they used a low order filter.

Using the VAB84 technique at 12 dB/octave I could move the filter up to 500 Hz - way higher than it needed to be - without a huge sonic penalty.
At more sane frequencies I really couldn't hear it.

The other explanation I've heard for "elliptical" was the 'scope pattern from monitoring in X-Y mode.

That's a really nice build.
Gold
Posts: 677
Joined: Sat Dec 01, 2007 5:20 pm

Re: Using the Precision MS Matrix for Mono Crossover LF Blen

Post by Gold »

mediatechnology wrote: What I think you may have missed in this thread as it eventually evolved is that the technique I'm showing here doesn't actually use MS.
It just uses the MS board.
I did miss that. I'll read the thread more carefully.
As Neumann wrote about the VAB84 ""Flank signal components are simultaneously transferred to the other channel."
So if the Bass elements that are in Side have some artistic value to the Mix they're still heard, not cancelled by blending.
Side info in right gets transferred to left and vice-versa.
The EE-series simply cancel the side component by allowing it to self-cancel.
The dynamic nature of the filter makes it more audible than a simple EE IMO. Conceptually it seems better, like it would only do what is necessary and get out of the way when it's not necessary. In practice having the filter cutoff move around mostly serves to make it more audible. That is of course my opinion but I have heard similar things from a few people I respect.
Left and Right stay in their respective domains: Only Side is derived so it can be folded back into Left and Right without the Bass loss.
The Bass in Side gets put back into L and R or more specifically L+R.
I think the VAB84 sliding filter may have been chosen as a technique because they used a low order filter.
I have heard some other attempts to do this (like the Haeco) and I think in practice they are less successful than a simple EE. They do work in some cases but in many cases they just make the mix seem bloated.

The other explanation I've heard for "elliptical" was the 'scope pattern from monitoring in X-Y mode.
It's really the same thing as a vectorscope is a 2D analog of the groove shape.
That's a really nice build.
Thanks. I'm finally getting faceplates I can live with. The bones of the console is getting pretty close to installation. The console wiring should be done in a couple of weeks. Then it's a month or two of finishing the transfer path and monitor/meters and testing. Then install. I think it can happen over a long weekend as I can test it with all the source and record machines connected with extender cables. So I can really test it in situ before installing. Then finish up the rest of it and drop pieces in as they are finished.

The biggest problem with the install is that it's much bigger than the Neumann SP75. I will sit a few inches higher. The speakers will need to be raised and the acoustics tweaked.
Last edited by Gold on Thu Feb 13, 2014 7:34 pm, edited 1 time in total.
User avatar
mediatechnology
Posts: 5442
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: Using the Precision MS Matrix for Mono Crossover LF Blen

Post by mediatechnology »

The dynamic nature of the filter makes it more audible than a simple EE IMO. Conceptually it seems better, like it would only do what is necessary and get out of the way when it's not necessary. In practice having the filter cutoff move around mostly serves to make it more audible. That is of course my opinion but I have heard similar things from a few people I respect.
I don't think I buy into the sliding filter either. A switch goes a long way.

What makes the VAB84 worth exploring is: "Flank signal components are simultaneously transferred to the other channel."

What I found with using the VAB84's topology was that I could switch the LF blend in and out without any perceptual LF loss or build-up on most well-behaved stuff.
The vinyl to wav transfer that had large amounts of LF Side information that sounded incredibly weird (CSNY Carry On) in 'phones sounded more focused with it in circuit.

So I think it depends on what the source material hands you:

Some LF out-of-phase components may not have artistic value regardless of whether not they can be cut to vinyl.
Other LF out-of-phase components may have actual artistic value.

Some things might benefit from an EE70/77 elimination of out-of-phase of Side components.
Other things might benefit from folding Side into Left and Right using the VAB84's "L-[1/2(L-R)]" steering.
Gold
Posts: 677
Joined: Sat Dec 01, 2007 5:20 pm

Re: Using the Precision MS Matrix for Mono Crossover LF Blen

Post by Gold »

mediatechnology wrote: So I think it depends on what the source material hands you:

Some LF out-of-phase components may not have artistic value regardless of whether not they can be cut to vinyl.
Other LF out-of-phase components may have actual artistic value.
This is undoubtably true. I however am a sausage maker. If I auditioned each side and tweaked the LF control to whipped and airy perfection I'd be broke. If there is a problem, an EE is the best overall solution to keep the record from skipping. If the client doesn't like an EE I am happy to run the clock. If I'm mastering the record I can take care of that stuff with EQ and compression.
Post Reply