Hearing yourself on stage. Ben Duncan forms his thinking into a tight wedge. Part two next month.

The alleged purpose of monitoring is to help musicians hear themselves on stage. A good monitoring system allows everyone to hear exactly what they need to hear, and little else. For the perfect monitoring system, read "...and nothing else".

Alas, a real tour of 'the public lavatories' is so very different. Basic monitor speakers, no names mentioned, are apt to issue forth with a shrieky, British telecomesque quality. Musicians often think they need to hear everyone else's voice/instrument at full level to play successfully. Then someone always wants more level on the flugelhorn at all frequencies. So EQ knobs go up and down, as the monitor guy sweats to retain a faithful sound while duelling with the dual-shriekbacks of monitor howlround (eeeeeEEEEEE! 140dB at 3kHz plus confused or irate musos).

Meanwhile, kilowatts of out-of-balance, muddy, nauseous sound exit from each monitor, leap down the mikes and louse up the outfront mix.

The monitor guy jettisons the shrieking monitor cans to rescue the vocalist's mike stand, which has just collapsed, thanks to some over-enthusiastic fans. With so much love, praise and appreciation for hard work on all fronts at once, it's no wonder monitor engineers get paranoid and end up pretending they're stone deaf.

On a superficial examination, the outfront PA needs to sound as good as it possibly can, even if this is at the expense of the monitors. But this approach is a trifle myoptic when we consider:

1) The music will itself suffer if the monitoring fails to keep the musicians informed.

2) An improved monitor set-up reflects directly on the sound outfront, the more so, the more mikes and monitors there are on stage.

3) The musician's direct experience and consequent assessment of how the PA performs at each gig is solely down to the success (or otherwise) of the monitors. The outfront sound may well be audible, but cannot be judged on stage.

Even when all the amps are switched off, monitoring is a brain damage topic. Perhaps the first hurdle is psychological. This means establishing not so much what musicians think they need to hear, but defining the absolute content of the mix, the bare essentials required to keep time and perform the show effectively. Getting by with less is an arch discipline. Ergo, musicians often find themselves needing less as they become more experienced.

To wit, original American R'n'B musicians will play, and play damned well, even if the monitoring is rudimentary to the point of near non-existence. Regardless of whether the monitor set-up is a lonely 12in speaker in a wedge-shaped box, or a 20kW triamped system, with 999 band graphic EQ and (gasp) an intergalactic time-screech co-ordinator.

The principle holds that the fewer the instruments put through each monitor, the easier it is to make possible the impossible. Rock'n'Roll monitoring is like this because musicians need 128 decibel sound levels to hear themselves above the adjacent backline cabs (=124dB), and even the spill from the outfront PA (itself typically 105 to 115dB).

At the same time, players are naturally close to their mikes, so whatever the SPL at ear level, at the mike it will be much the same. At working stage levels, this is a sure recipe for an effect variously known as feedback/howlround/shriekback. The mechanism at work here is a signal chasing its own tail. The event is triggered by spurious aggregations of sound energy at specific frequencies and points in space. In other words, it's all the fault of inconsistencies in the directivity (dispersion) and even-ness of response in both the mikes, and the monitor speakers. To counter this, the engineer EQ's the mix. Alas, cutting out large segments of the audible spectrum has a disturbing effect on the sound of some, if not all the instruments. Even with a 1/3rd octave graphic (this has 27 knobs to permit fine tuning) the monitor engineer may have a hard time EQing to balance instrument tonality against the entirely different EQ setting necessary to keep howls and squeals at bay. And the higher the sound levels demanded, and the more and more varied the instrument's tonality, the more unlikely it is that the poor monitor engineer will reconcile these conflicts before the end of the gig/the tour/the band/the world.

The 'all instruments in all mixes' philosophy also carries with it the disadvantage that a greater number of mikes are now operating in the circle of latent howlround; while more-and-louder-monitors mean yet more honky sound feeding through to the first 1/10/100 rows of seats.

MONITORING ON THE SMALL £

Here are some shortform tips for beginners, garnered from the zippy Andy Salmon, better known to many as Midland Sound & Light.

Without the luxury of a purpose-built monitor console, the average small PA has to rely on the valuable auxiliary sends provided courtesy of the outfront mixer. For monitoring, the signal needs to be derived before each channel's fader; that's a pre-fader auxiliary. So even if your mixer is well provided for, with 3 or 4 sends, and some pre-post switching, practical monitoring is more often limited to 2 or 3 mixes at best. Especially so if you need echo and maybe another FX.

Actually, it's possible to use a postfader send, if you're desperate, but watch the polarity! More often than not, there's a polarity (or 'phase') transposition between the pre- and post-sends. Meaning that when wedges 1 and 2 are blowing, wedge 3 is sucking.

Not only is this a gross offence to public decency but, much worse, it can lead to sporadic cancellation, the same sort of effect as having two adjacent mikes 'out-of-phase'. So when the vocalist steps back, a chunk of the monitoring disappears — literally — into thin air. Or maybe the air gets fatter — a summation. Then the system gets into howlround. For this reason, post-fader auxiliaries are best sent out to the drummer, who can be relied upon to sit still. Put the wedge behind him, on the floor.

Some mixers below the £1k have overall EQ on the send(s) output(s). This may be primitive (eg: "TONE") but mostly it helps. In fact on a low budget, any EQ is better than nothing. This can be taken as far as feeding the monitor send(s) into an Accessit equaliser or any similar gizmo you care to try. Shure's 'feedback equaliser', long since obsolete, is a Godsend if you can acquire a secondhand one cheaply.

Usable power for basic monitoring is around 100 to 200 watts per wedge. You can connect up as many megawatts as you like, but it's a waste of time unless and until those little howlrounds have been tuned out of the system as much as possible before applying EQ, in other words, by your choice of mikes and wedges, in particular.

Numero 1 aggravateur de la howle in elementary setups is usually the speakers' drive units, their cabinets and or their crossover network. Many smaller PA companies build their own monitor cabinets, but it's a waste of good loot to fit up-market components, like the time-honoured ATC 9in or RCF 15in drivers. Unless, that is, you can find time to tune the prototype enclosure and crossover for a smooth response. Other than checking for a good vocal sound, this means hooking up the proposed cabinet to a mike plus amp. Then we vary the physical volume of the enclosure (and bass vent, if applicable), trying different combinations of wadding inside, and tweaking the crossover components. Fane and Celestion have both published wedge monitor designs in their respective cabinet handbooks.

After you've spent a few Sunday afternoons playing about with the highest possible sound level without EQ, you'll begin to suss why pro wedges are so expensive, and why megawatt big-stage systems are invariably bi-amped. Why? Because with active crossovers it's much easier to balance the relative contributions of the individual drivers, and thus smooth out the response.

MONITOR TUNE-UP TIPS

For best articulation (definition), punch and attack, keep the monitor mix as sparse as possible. Avoid guitars ('blox out vox'), and true bass frequencies, the ones you can feel physically. Over-mixing lessens the maximum sound level before howlround so if there's a big problem be ruthless and cut out that flugelhorn.

EQ tuning is done with a mike clasped in your right hand, while you chant the great litany "ONE! TWO! TESTING! — ONE! TWO! TEST! TEST!". When adjusting the EQ, a brief listen to the tone of the incipient howl/screech/shriek will soon guide you to the correct frequency. But the talent of knowing 'which knob?' on a 27-band graphic will take a little longer to perfect.

In general, shriekbacks are found between 1kHz and 5kHz. As one shriekback is stifled, a fresh wail will take its place, as the level is progressively inched up. You are allowed 1,000 chants to get it right. Don't forget the intermediate listening and tuning, at each stage to keep the vocals sounding tasty. For example, we've got as far as "ONE TTTWO TSSSTTING", so let's cut the mid a bit: "UNN UOOO ESSTIN" and now some more at 8kHz: "ONESSST TOOESSST TEESTINGSSSST", and so on, until you arrive back at the crisp, clipped, articulate sound you know to be like your own voice. But hopefully twice as loud as when you started... no marks if it isn't any louder after six minutes.

When tuning, try different positions. No, we don't mean sitting in a lotus posture on the console. Get someone to walk around stage with a mike, listening out for the ringing sound that signals the threshold of instability. Then put some sound up the wedges and check the audibility and articulation around the stage.

When the monitors howlround, the channel overload LED(s), where fitted, will tell you which channel auxiliary send(s) to turn down. Only hit the main send if you loose control of the situation. Most humans can handle but two knobs at a time when a simultaneous rotary motion is called for. Any other adjustments (like anticipating a howl from hearing an occasional ringing sound) should be done more slowly, thereby making them less obvious.

HOW TO REACT UNDER EXTREME BARLOCKS

The PA company remains nameless, but they turned up with two wedges for an 8-piece band. The band's long-suffering monitor engineer calmly pointed the hapless wedges in the general direction of the band, wound the gain up, and promptly blew the lot. Unalarmed, he made a mental note to ask for a 'substantial' reduction on the hire fee, then walked over to the PA stax, and angled the cabs slightly inwards, so the band could hear the outfront sound.

Fane's Loudspeaker Enclosure book costs £1.50, as does Celestion's Cabinet handbook. If you can't get them at your local music shop, go direct: Fane Acoustics, (Contact Details). Celestion, (Contact Details).