Information from the Tannoy Near-Fields Handbook

1.0 Introduction 2.0 The Basics 2.1 Unpacking your speakers 2.2 Connecting to your speakers 2.3 In Polarity 2.4 Amplifiers 3.0 Placement of the Speakers 3.1 Orientation 3.2 Setting the Toe-in 3.3 Console Reflections 3.4 Speaker Mounting 3.5 Bass Ports 3.6 Equalization 4.0 Hearing Safety Issues

1.0 Introduction

2.0 The Basics

2.1 Unpacking your speakers

2.2 Connecting your speakers

With the PBM 8, there's another worthwhile connection option to consider, and that's something we call "bi-wiring," first cousin to bi-amplifying. By separating the high and low pass filters in the passive crossover, and allowing a separate input to each, we can reduce the effects caused by the EMF (or generator effect) of the woofer voice coil on the signal, and improve the damping factor further. The woofer voice coil moves quite a lot, and generates quite a large amount of voltage under braking to a stop (that's the damping factor). A more direct connection to the brakes (the amplifier output terminals) improves the bass response, and by freeing the tweeter and its crossover filter circuit from these large excursions of reverse voltage, it is possible to clean up the midrange and high frequency signal. These are two ways to implement bi-wiring, the first is the simplest. Remove the jumper between the high and low frequency sections of the bi-wire terminals on the speaker system. You can connect two runs of #10AWG speaker wire together at each amplifier channel. Make sure you maintain the correct polarity for the low and high frequency sections. All this will cost you is double the cost in speaker wire.

The second more expensive option involves two identical amplifiers, with the two left channel inputs paralleled together, and the two right channel inputs paralleled together. Connect one amplifier's outputs to the low frequency bi-wire inputs on the speakers. Connect the second amplifier's outputs to the high frequency bi-wire inputs on the speakers. Ensure that correct and matching polarity is maintained. You now have a passively bi-amplified system, with the added option (and responsibility) of balancing the output levels of the high and low frequencies. This does require identical amplifiers to ensure that the phase response, time response and polarity is correct at the crossover point. This allows the low frequency amplifier to handle the woofer movement effectively while eliminating any audible artifacts with the passive crossover filters.

If you're using PBM-8 LM monitors equipped with the Limpet amplifier mounted on the back panel, all of the above concerns have been taken care of by having the amplifier really close to the loudspeaker (you can't get a piece of paper between them). This direct coupling improves the bass response by having very short speaker cable between the amplifier and loudspeaker crossover. We don't need to bi-wire this wire run of a couple of inches.

2.3 In Polarity

2.4 Amplifiers

3.0 Placement of the Speakers

3.1 Orientation

All two way component systems have to live with some listening position dependent compromises at the crossover point. The crossover frequency of all these small systems fall into the center of the midband (2.0kHz to 3.0kHz), where we are most capable of discerning frequency/phase response aberrations.

In the diagrams below is a representation of the speaker systems operating at the crossover point where both high and low frequency drivers produce the same output level. The first one shows a pair of two way loudspeakers laying on their side. Not that each driver is producing sound, and because there is a physical distance separating them on the baffle, there is also a time difference separating the drivers, and the result is what you see here. Around the crossover point, the speaker will produce numerous lobes, giving you variable midrange sound character as you move across the horizontal listening plane.

With the monitors laying horizontally, you will move through the largest number of variations caused by the physical/time offsets between the drivers. If you think this is hard to look at, imagine listening to it!

Because stereo happens left to right, that is the listening plane in which we try to minimize the changes in (physical/time) offset between the woofers and tweeters. And we gotta be honest, it's not perfect, that driver offset is still there, but by stacking the woofer and tweeter vertically on the baffle, we can give the mix engineer the widest range of movement in the horizontal plane. You can roll the chair across the length of your mixing console and not change the relationship between the order of the woofer and tweeter (just don't bob your head up and down while you do it).

With the woofer and tweeter stacked vertically, you experience the least variation as you move across the horizontal plane of the console work surface.

Another interesting note, the Tannoy true point-source Dual Concentric(c) monitors are free of the physical offsets previously described, which means that the diagram below is representative of the behavior of the Dual Concentric(c) monitors is that no matter how close you get to them you can't get the image to fragment into separate high and low frequency sources.

With dual Concentric monitors, there are no changes in the relationship between the woofer and tweeter, you have complete vertical and horizontal freedom of movement.

Now, if you were to follow the all too common practice of laying your two way monitors on their side to give you better sight lines over your meter bridge, you can see (and hear) what will happen. With the monitor on its side, moving your head horizontally means that you are now moving through all those rays, or lobes, where the wavefront from the woofers and tweeters interfere with each other. The midrange frequency response will be different for each head position. All two way component monitors, no matter who manufactures them, need to be used with the multi-driver axis vertical (that's just the way it has to be when you're in the near-field). Our Dual Concentric(c) monitors work equally well when placed horizontally or vertically, as the relationship between the woofer and tweeter doesn't change either way. And if you're wondering how three-way near-fields work with a whole bunch of speakers stuck all over the baffle, well... you'll quickly realize why we stopped at two way speaker systems.

3.2 Setting the Toe-in

The speaker axis (shown on the diagram) should be aimed at the half-way point between your furthest forward and the furthest rearward listening positions (as indicated by the two heads on the diagram). This is typically a range of about 24" (600mm). If you can, you also want to try to get your ears lined up with the vertical speaker axis (half way between the woofer and the normal listening position lined up in the best spot possible. If this would have you resting your chin on the console, you could tilt the monitor back slightly. This keeps your head in the sweet spot whether you're leaning forward adjusting level or EQ, or leaning back and listening to the mix. Don't go crazy trying to get this exact to three decimal places, if you are within an inch or two, that gets you into the game. Your Tannoy monitors have a wide sweet spot both horizontally and vertically to reduce the variations in sound quality as you move around doing your recording engineer stuff. Turning the monitors in like this has an added benefit of keeping the high frequencies from splashing off the walls and outboard gear.

3.3 Console Reflections

So you doubt our admonitions about this reflection. OK, dig out your pink noise generator, it's time for an experiment. You'll need an assistant from the audience for this. Have your assistant hold your monitor up in the air about 600mm (24") from the top of the console while feeding pink noise through it. Make sure you're listening right on the speaker axis (that line between the woofer and tweeter), and have your assistant slowly lower the monitor onto its perch on the meter bridge. Hear that change in midrange character as it gets within the last foot of the console? How much different does it sound than when it was up in the air? There's two significant things happening here. The first problem is comb-filtering (interference) caused by the reflection from the console top taking a fraction of a second (and a darn tiny fraction!) longer to get to your ear than the sound directly from the speaker. And you though you couldn't afford a Flanger for your studio. You may have noticed that in the last foot prior to touchdown on the console the comb filters did a big sweep through the midband. There is a real cruelty to Murphy's Law, the speed of sound in air gives us wavelengths in the midrange that just seem to coincide with the dimensions of everything we mount speakers on, maximizing the problems created by these reflections. The previous diagrams (horizontal and vertical) that show the interaction between a separate woofer and tweeter also demonstrate the problem generated by the reflection, substitute the reflection for one of the drivers in the diagram, and you've got a pretty good picture of what happens when the original wavefront and reflected wavefront meet. The big problem with this comb-filtering is that you can't fix it with any terrestrial equalizer, once the waves cancel there is nothing left to boost. You may also notice that as you move your head, the filtering changes (Cool, it really is a flanger!), so any equalization you attempted to apply for one position would really screw up all the other listening positions.

We mentioned that there are two things happening here, the second thing is a change in directivity caused by the addition of a boundary (the console top). When you want to shout at someone outdoors you cup your hand around your mouth to increase the directivity of your voice. By placing the monitor on the meter bridge, you've done the equivalent of putting one hand to your mouth. In the lower midrange (200-800Hz), where the wavelength is long enough at 1400mm-400mm(55"-16") that you don't get cancellation, you get the wavefronts adding together (Man, if it's not one thing it's another!). This causes the level in the lower midrange to get bumped up at the same time that holes show up in the midrange from comb-filtering.

Now that you've begun to believe it is impossible to use near-fields on a console top, we'll talk about what you can do to help mitigate these problems. The first thing you need to do is be able to identify the surfaces that are close enough to do serious harm. You can do this using a time domain analyser worth several thousand dollars, and spend a few months learning to use it, OR, you can grab a length of string, some gaff tape, and a mic stand, and get set for another experiment.

First the theory. For unblemished stereo imaging and frequency response you would want to listen in a completely reflection free environment, like an anechoic chamber, where all you would hear is the image and sound produced by the speakers, no nasty reflections anywhere. For most people this is impractical. Next best thing, if you can establish a listening position free of reflections arriving within 2 milliseconds after the direct sound (that's the time it takes sound to travel about 24" or 600mm or less which represents all frequencies from 500 Hz and up), and minimize reflections arriving within 10 milliseconds of the direct sound, you can maintain a remarkably stable stereo image, and uniform response throughout the mix area. The "direct sound" is just that, it is the shortest straight line path that sound can take from the speaker to your ear, no bounces, no reflections. the 2 millisecond reflection window really affects the character of the sound at the mix position, drastically altering the response of the speaker in the critical audio bands of 500Hz and above (imagine a maniac getting loose with a multiband parametric in cut mode). The 10 millisecond reflection window does some more subtle things to the speaker's response (spacing the comb filtering from reflections actually), but because the ear/brain reads reflections arriving within 10 milliseconds of the direct sound as being part of the speaker's response, these reflections can pull the stereo image around in different directions at different frequencies. Enough theory, now the lab.

Here's where we have fun with science. Take the microphone stand and place it at the mix position. Attach one end of the string to the top of the mic stand, and stretch the string out to the front of one of the monitors. This is the direct sound path from the speaker to your mix position. To locate all the surfaces that will contribute reflections within that magic 2 millisecond window, add 600mm(24") to the string you have stretched out. Take a small piece of gaff tape and attach the string to the baffle (NOT the drivers!). Now, every surface you can touch with any part of that string can contribute a reflection to your mix position.

This same concern about reflecting surfaces applies to all nearby widgets such as computer monitors, outboard gear, coffee cups, rolls of tape, the producer's wooden leg and all the other usual paraphernalia found in a studio environment. Cleaning up this short sound path between the speaker and your listening position is like using a Q-Tip on your ear.

You can repeat the process with an extra 3000mm (10 feet) of string to see which surfaces will influence the timbre and imaging of your mix. This exercise is not just about finding places to stick fuzzy or foamy absorbers to, the last thing you want to do is make your room completely dead. What this process will show you is which surfaces you should try to angle to redirect reflections away from your mix position. If there are surfaces that you cannot move or shift, you can apply a small amount of absorbent material to specific surfaces, rather than covering the entire room in absorbers.

You may want to consider a speaker placement other than the console meter bridge, perhaps on an elevated mounting arm attached to the wall, or on a mid-field monitor stand, just behind the console. These positions can help clean up that 2 millisecond window. Keep your ears open for other problem reflections, like between the sloped connector panel on the back of the console and the baffle of the speaker, which will find its way back to the mix position a bit too late to be useful. The string trick works fine for locating these reflecting surfaces.

It's important that you listen to the effect of speaker placement on the sound character of your speakers, and understand what those changes are going to do to your mix. To make the point here's some more stuff to consider.

3.4 Speaker Mounting

3.5 Bass Ports

If you're using a separate passive or active subwoofer, you may want to consider plugging the port tubes on your near-fields with a closed cell foam-rubber plug, friction fit for a full seal. Because the ports aren't needed if the monitor speakers are being used with a high pass filter, you won't be losing any bass performance and you can improve the mid-bass response by plugging the ports.

3.6 Equalization

An important thing to keep in mind if you are equalizing these compact playback monitors is that they are not intended to substitute for really large loudspeaker systems you might use for tracking synth lines, or drum samples. Don't try to equalize them to sound the same as the 15" three-way you were listening to in the synth department in the music store. The equalization of your near-fields should be undertaken with great subtlety, with only the best intentions of correcting some of the placement related characteristics we've talked about here, a subtle bit of reduction in the mid bass where the console or table top will give you a rise, perhaps a bit of tailoring to taste here or there. As a general rule for this application, you should never need more than 6dB of control range, and you should never see two adjacent filters with more than 6dB of difference between them. Equalizers cause shifts in phase response in the speaker system that's causing the problem, then ll this shiftiness cancels out. As we mentioned before, you can't fix a dip in the response caused by a reflection, that phase shift is waaaaaayy too big for an equalizer to fix, all you do is mess up the phase response and the frequency response around the problem.

4.0 Hearing Safety Issues