Readers who are familiar with my reviews know that I do a lot of sum-difference processing. In other words, I convert stereo Left/Right signals to Mid/Side, process the M/S, and reconvert back to L/R. (Or sometimes I start with M/S, process, then convert to L/R.) My Sony DMX-R100 digital console (Tape Op #25) has M/S encoders available on all fader pairs, and my Peter Engh Sonic Orbit (Tape Op #56) is a passive, transformer-based, stereo matrix for converting between M/S and L/R. The "math" for converting between L/RandM/Sisquitestraightforward:M=L+R,andS=L-R. With a bit of seventh-grade algebra, you can derive the converse:L=M+S,andR=M-S.Ifyourereadthesetwo pairs of equations, you'll notice that the first of each pair is a sum and the second of each pair is a difference, and you can use a sum-difference matrix to go either way-L/R to M/S, or M/S back to L/R. (Astute geeks will realize that this is actually a minor simplification; the equations require the square root of 2 as a coefficient in order for the conversion to work both ways without a change in gain.) You can implement a one-way, sum-difference matrix using three faders of a mixer. Send Track 1 to Fader 1 and assign it to Bus 1 and Bus 2, panned center. Send Track 2 to Fader 2 and assign to Bus 1 only. Mult (send a copy of) Track 2 to Fader 3, flip polarity, and assign to Bus 2 only. Then group Faders 2 and 3 so that they stay at equal levels. To convert back, create a second, identical matrix with Faders 4-6, and this time, feed Bus 1 to Fader 4, panned center; Bus 2 to grouped Faders 5 and 6, hard-panned left and right with Fader 6 polarity flipped. By changing the level, the EQ, or even the timing (delay) of the faders, or by inserting external effects on Bus 1 and 2, you can vary the stereo character of the two recorded tracks. Generally speaking, you can do things like change the level of the center image in a stereo mix; widen the stereo field; or change the dynamics or EQ of the ambience. Specifically, you can do things like increase the "crack" of a centered snare drum while simultaneously reducing the "sizzle" of hard-panned cymbals in a stereo drum mix. Anyway, the list of sum-difference trickery goes on and on. But admittedly, sum-difference processing requires a penchant for tweaking things to the nth degree, and unless you have a dedicated sum-difference processor or plug-in, you'll also need a mixer that tracks well (with signal levels and EQs that are extremely consistent between faders). The latter is no problem if you're working "in the box" or on a digital mixer. But if you're analog, even the most expensive analog desks suffer from some inconsistencies between faders, and you could end up hurting your stereo image even before you insert any processing. Furthermore, setting all this up correctly and calibrating everything would really slow down your mix session; it's a lot of extra work.
There are outboard processors that implement the sum/difference networks for you. I've played with many of them. There's the aforementioned Sonic Orbit; it's only a one-way matrix, so you need two of them to convert each way, straddling your outboard processors in the middle. The Groove Tubes MS (Tape Op #56) is another one-way box, but this one is active, and I prefer the sound of the transformer-based Sonic Orbit. There are high-end digital processors from Weiss Engineering (EQ1 in Tape Op #31); these have a higher learning-curve and are really designed for mastering applications. Dangerous Music's S&M (Tape Op #54) is an active, analog two-way matrix with inserts between its two sum-difference networks; its only built-in control is a knob to adjust stereo width. And then there's the Portico 5014 Stereo Field Editor. The 5014 SFE is designed to take a stereo L/R signal, convert it to M/S, allow you to affect the M/S signal with built-in controls as well as an external effects loop, and then reconvert the signal back to L/R.
I've had the 5014 SFE in my rack for a couple months now. Like its siblings in the Portico line, the SFE utilizes input and output transformers (RN8960 and RND-2041) custom-designed by Rupert Neve, with internal circuitry incorporating well-regarded IC op-amps (NE5534, NE5532, as well as OP27G for the meter card). And like my Portico 5033 parametric EQs (Tape Op #57), the SFE imparts a little bit of midrange "magic" (with a teeny bit of girth) to whatever I send through it, so even before twiddling any of the controls, I was liking it. Controls include left and right input level pots with a range of -6 dB to +12 dB of gain. It's nice that you're given an area of very fine trim within +/-2 dB of zero, but I would rather have had a single stereo knob for level alongside a balance pot. Eight-step LED meters above each pot can be switched to show either input or output level. Because the SFE sports a textured plastic fascia inset into the heavy-gauge, aircraft-grade aluminum front panel, the meters look kind of fuzzy, and the light from each individual LED bleeds to its neighboring LEDs, so it's difficult to get a precise reading. There are also controls to enable and vary Width and Depth, as well as a simple, semi-parametric midrange EQ for the difference channel. After careful listening tests, I slapped a signal generator and a digital scope onto the 5014 to verify what I heard.
The Width knob changes the level of the difference signal, allowing you to take it out completely or add more into the matrix before reconversion to L/R. Personally, I wish the knob also changed the level of the sum in inverse, so that the final L/R output wouldn't change in level so drastically as you turn it; as it's designed now, it's easy to overload the L/R output if you turn the Width up too much without turning down the L/R inputs correspondingly. The Depth knob is kind of a continuously-variable crossfade between the sum and difference signals, swapping the two as you turn the knob from left to right. In other words, with the Depth knob at full counter-clockwise rotation, when you push the Depth enable button, the Mid channel has mostly sum in it while the Side channel has mostly difference. At full clockwise rotation, the Mid channel has mostly difference in it while the Side has mostly sum; what's centered becomes stereo ambience and vice-versa. Unfortunately, there isn't a complete swapping, even at extreme settings of the knobs. Furthermore, both the insert (external effects loop) and EQ are single-channel and normally affect only the Side. I say "normally" because you can cheat by flipping the polarity of one of the L/R inputs on the SFE and flipping the polarity on the corresponding output (using your mixer or whatever the SFE is feeding). This will swap the Mid and Side within the sum and difference matrix such that the Width knob will control the level of Mid; the EQ and insert will affect the Mid; and the Depth crossfade will work in reverse. (Doing the sum-difference "math" will verify this.) But really this cheat is nowhere near as useful as it would have been to have simultaneous control of both the Mid and Side, and herein lies the SFE's biggest limitation. To fit into the physical design constraints of Portico line-both in terms of layout and bussing architecture for connection to other Portico units (as well as to the new Rupert Neve 5088 console)-there's no room for a second insert channel. For advanced
sum-difference work, I think the SFE would be many times more useful if it had inserts for both Mid and Side to process the two independently.
For example, using the SFE to implement the sum-difference trick I mention earlier of adding "crack" to the snare while reducing cymbal splash in a stereo drum mix would require workarounds. To bring out the crack of the snare, you can add highs to the L/R drum mix using your DAW or an external EQ while simultaneously using the SFE to reduce highs in just the Side, not only to counteract the highs added overall but also to reduce the cymbal "sizzle". Or you can do two passes thru the SFE-one pass to reduce the highs in the Side followed by a second pass with the polarity-reverse trick to increase the highs in the Mid. But at this point, you might as well implement your own sum-difference network using six faders and a couple busses. It's clear that the SFE is not for the supergeek who thinks in sum-difference notation. Instead, the SFE is designed for the engineer who wants to change the "feel" and "placement" of stereo tracks-and no pocket protectors or RPN calculators are necessary.
The 5014's Width control can really turn up the "stereo" effect in a stereo signal, and you can EQ that stereo effect too. But what's really powerful is the Depth control (in conjunction with the Width and EQ) and its ability to affect the perceived ambience-by centering or further dispersing the room sound or reverb without having to change its level-so that you have an extra dimension (other than pan, level, or EQ) at your disposal for positioning instruments in a mix. For example, I had a lot of success-and fun-placing stereo background vocals in a dense mix using the SFE. Normally, I would have pulled out a bunch of the "normal" mix tools to make sure the background vocals sat in the right place-EQs, compressors, pan and level knobs, different delays and reverbs... and a lot of tweaking. On the other hand, with the SFE patched in, it was easy to dial in the position of the background vocals-"behind" and just "surrounding" the main vocals, but "closer" than the guitars-with less destructive EQ and compression, and without getting things all muddy with additional delay or reverb. Same goes for acoustic guitar (which I almost always record with two mics). To make sure the acoustic doesn't get all soupy with the electrics or stomp all over the vocals, I normally use a frequency-selective compressor to significantly reduce the lower midrange. Now I can be a little less blunt with the compressor, because I have that additional dimensional control that the SFE affords me in positioning the acoustic guitar relative to the other instruments in the mix. By now, I'm sure you can imagine other uses that the SFE would excel at.
It's worth noting that all of the SFE's processing capabilities can be emulated using the multi-fader sum-difference network I describe earlier. Even Depth-like control can be implemented by using an intermediate set of faders on the M/S busses to allow cross-panning of the M/S signals-that's super-advanced sum-difference geeking. But that just further proves my point. All that geeking and potential for signal degradation would really take the fun and immediacy out of mixing. The 5014 SFE, on the other hand, is an exceptionally easy-to-use and great-sounding tool that gives you dimensional control over a stereo signal with just a few friendly knobs. For all the engineers and mixers out there that can't be bothered with advanced algebra, it's a powerful and fun box to have in your arsenal. ($1795 MSRP; www.rupertneve.com)
Tape Op is a bi-monthly magazine devoted to the art of record making.