can someone explain Sum and diffirence mode and give some nice examples of using it?

[perx]

I never used this function and just trying to understand it and maybe use it once I learn more about it.

I have the feature in my neve 8816 mixer. (skip the following part if you dont want to know what the manual has to say about this)

The manual says this about it:

"If the mix level control is pressed then the illumination will change to red indicating that the INSERT POINT has changed to a SUM AND DIFFIRENCE INSERT.

In SUM AND DIFFIRENCE MODE the standard stereo signal is converted into a mono common or sum signal on the left output and a mono diffirence signal on the right. This enables the engineer to apply diffirent processes to the common (or centre) image and the diffirence (or stereo) image.

At the insert return, the signal is re-converted to a standard stereo signal. The technique is frequently used in mastering recordings and really handy for such things as:

Using an overall stereo EQ where the l&r image will not shift with any tracking errors within the EQ unti , or EQ the mono to sublety change the mix balance post mixing.

Compress and limit the sum and diffirence separately to negate tracking errors or create a diffirent stereo image.

Add a sub bass synth to the mono program for a really deep mono bass that will not make the cutter head jump if mastering to vinyl.

a vocal could be brightened by EQing the sum signal or cymbals could be brightened EQing the diffirence signal."

have I understood this the right way that what this process is doing is that it separates "stereo signals" and sends them to the right insert output and whatever "mono signals" to the left insert, and then sums the returns together?

In a way, the way they describe it makes sense, but lets say I want to use two diffirent compressors or one eq and one compressor.. How do you match the output volumes, etc?? Right now I have an api compressor on the main inserts - using it for the final main mix compression.

Can anyone describe how this sum and diffirence works - AND if possible, give some good examples on how to connect diffirent outboard and use this to add some nice touches when mastering or producing a track etc..

Thanks in advance..

[nlc201]

Think of the theory this way. Understanding the basics is crucial to using this technique properly and can also inspire some creative uses of it as well.


You start with a stereo signal: The two components we'll label "L" for left channel and "R" for right channel. When you add both together, you have L+R. This is essentially taking both sides and summing them together into one signal. Just like the mono button does on the console or putting each channel's pan to the center. This signal is known as the SUM (=L+R). Now, take those same two signals and flip the polarity of one of them and add them together. This is known as the DIFFERENCE signal (=L+(-R)=L-R). The difference signal represents all elements of the stereo signal that differ from each other. If you did this to a true mono signal, you would cancel it out and have nothing. In a stereo signal, the difference between the right and left sides is what gives the program the properties of a 2 channel stereophonic sound, IE, sounds "existing" on one side rather than the other or a sense that certain elements are "wide" and coming from a whole region of space rather than a pinpoint location.

Before you get into how to process these signals individually, it's important that we know how this is all going to translate back into the real world of stereo information. If we simply summed the two back together, we would end up with one discrete side of our original stereo signal. IE, SUM + DIFFERENCE=(L+R)+(L-R)=2L=just the left side. Now, if we flip the polarity of the difference signal and add it to the sum, we get the other side. SUM-DIFFERENCE=(L+R)-(L-R)=2R=just the right side. If we do both of these processes at the same time and pan each result back to the appropriate Left or Right channel, we end up back with our original stereo signal. Get it?


Now the fun begins!


One of the simplest and most effective uses of this technique is just a simple alteration of overall level of one or both of the channels. If the program is a little too centered in the middle and needs some width to it, a good way to get it is to use Sum and Difference and simply turn up the difference channel (or turn down the sum channel). Likewise, program that is a little too wide and lacks definition and punch can be reined in a little by simply turning down the difference channel. The mono information comes up, presto.


Of course, there are a whole myriad of tricks we can do by processing sum and difference. Use EQ to brighten a vocal without affecting the stereo acoustic guitar and getting nasty string and pick noise. Give some weight to the bass guitar and kick drum without muddying up the synth pads. Make the vocal reverb sound brighter but leave the vocal untouched, etc.

Try compressing the sum signal and see if you can get the center of your program sitting a little bit more even in the stereo soundscape. Likewise, compress the difference to make your extreme left and right a little more even. Or, compress the crap out of the difference for a cool, pumping ambience effect.

Add a little mono reverb to the difference signal only to really increse the width of the ambience.


Like, I said, there's a myriad of interesting techniques. Mastering engineers use this technique a lot to problem solve because it can make subtle to drastic changes to some elements of a stere program without making drastic or subtle changes to other parts.


This technique can be done easily, even if you don't have a Sum and Difference box. You just need three channels on a mixing console.

Channel 1=Left side of program, panned center
Channel 2=Right side of program, panned hard left
Channel 3=Right side or program, panned hard right, polarity reversed.


The stereo output should now be the sum and difference signals. The left channel being the Sum and the right channel being the difference.

Here's how to re-encode it from sum and difference back into stereo left and right. Again, 3 channels.

Channel 1=Sum signal, panned center
Channel 2=Difference signal, panned hard left
Channel 3=Difference signal, panned hard right, polarity reversed



That's all there is to it. Any sum and difference encoding/decoding box or circuit is essentially just doing this.

That's all there is to it! Well......not really.....but enough to get you started. Try messing around with it. It's kinda fun! Good luck

[perx]

Originally Posted by nlc201

This technique can be done easily, even if you don't have a Sum and Difference box. You just need three channels on a mixing console.

Channel 1=Left side of program, panned center
Channel 2=Right side of program, panned hard left
Channel 3=Right side or program, panned hard right, polarity reversed.


The stereo output should now be the sum and difference signals. The left channel being the Sum and the right channel being the difference.

Here's how to re-encode it from sum and difference back into stereo left and right. Again, 3 channels.

Channel 1=Sum signal, panned center
Channel 2=Difference signal, panned hard left
Channel 3=Difference signal, panned hard right, polarity reversed



That's all there is to it. Any sum and difference encoding/decoding box or circuit is essentially just doing this.

That's all there is to it! Well......not really.....but enough to get you started. Try messing around with it. It's kinda fun! Good luck

Awesome! thanks for taking the time to write all that down.. Really appreciate it.

Just not 100% about it yet.. (slowly getting there though.. LOL..)

If you read the text excerpt from the 8816 manual - it says it converts the signal to sum and diffirence on the insert outs, and then on the returns it re converts back to a standard stereo signal. So with the 8816 doing that for me, I guess I do not have to patch like the example you gave above.? I should not need to use the polarity reversed at all if Im right?

So, with the 8816 doing the split - this basically mean I should use a processor (EQ, compressor, whatever), on each of the two channels? Lets say I use a compressor on the sum (mono´s) to get my drums and all really hard. And then something else on the diffirence signal. Would that be "correct" in the terms of connecting the ins and outs of the output? How do you go about having it all come together - my guess is that I need to watch the outputs of all units and make sure the levels on both are relatively the same? Because once I get back the stereo signal - its going to be recorded into the computer..

Those are the first questions I come to think about - the 2nd being about the polarity reversed you mention in your connections example. Should I check that button inside cubase on one of the panned signals? What if the analog mixer im using do not have that on the returns back? I guess the Neve has since it converts things back and forth - but Im just thinking if I would not use the never, but another analog mixing desk instead.

Hope the questions are´nt too hard.. ;-)

thanks again!!!

[nlc201]

Not at all. Happy to help.


The Neve works by doing exactly what I wrote down except it has a decicated circuit. Thus, when you engage this circuit, it goes through that process for you and spits out sum and difference on the insert outs instead of left and right. The insert returns conveniently re-encode the sum and differnce back into stereo for you. I went through the trouble of explaining how the process works in the hopes that grasping the theory behind this technique will lead you to a more effective mastery of its use. I didn't understand it at all at first and I really had to think about it for awhile in order to finally get what's going on. Be patient, you'll get it.


Now, as to how exactly to process sum and difference, you've just got to experiment and really learn what that kind of processing does and what it sounds like. Yes you should be very concious of your relative levels back into the Neve as even small balance changes between the sum and difference signal can have drastic and unforseen effects when the signal is re-encoded into stereo. The key is to listen very closely (which you should be doing anyway!) to the effects after the signal is re-encoded back to stereo to be printed.


In regards to setting this up in cubase, I'm not familiar with that particular DAW but as long as you acount for latency, the technique should work just the same. If there is no polarity flip, try using a simple plugin that has that feature. BE VERY CAREFUL THAT LATENCY IS ACCOUNTED FOR IF YOU DO THIS!!!! Any time difference between the signals (even 1 sample!) incurred through digital processing will SCREW UP the process as none of the phase information will be accurate. A nice comb filtered sound should let you know if this is happening. I work in pro tools and set up aux inputs in my PT mixer exactly how I describe. Since there is no polarity button in PT, I place a simple "trim" plugin on each input. I don't actually use the trim feature, I just use the polarity flip on the channel that requires it. The reason I insert the plugin on all three is to match the latency incurred accross all channels, keeping the latency the same thereby keeping the signals phase accurate. But, only one of the three plugs actually does anything to the signal (flip polarity). The other two are left untouched at their default, do-nothing dettings simply to match the latency of those channels to the latency of the channel have its polarity flipped. In the analog world, this is not necessary as there really are not latency considerations for analog circuits like this.

On an analog board with no polarity switches, you can try using a polarity flipped cable, that is a cable with one of the connectors wired backwards (hot to cold, cold to hot). Some Y-cables are premade to have one branch of the split polarity reversed. Or, easier yet, get an adapter that flips the polarity (they make them for all types of connectors).


I think that's about it. My advice remains that you should simply try out this technique. Play around with it and see what it does and doesen't do. Like many things, done right it can work wonders. Done wrong it can screw up everything up royally. Good luck!