Setting sound system level controls: The most expensive system set up wrongnever performs as well as an inexpensive system set up correctly.
Jul 1, 1997 12:00 PM, Dennis Bohn
Correctly setting a sound system's gain structure is one of the mostimpor-tant contributors to cre-ating a sound system that sounds good.Conversely, an improperly set gain structure is one of the leadingcontributors to systems that sound bad. The cost of the system is secondaryto proper setup. The most expensive system set up wrong never performs upto the level of a correctly set, inexpensive system. Setting all thevarious level controls is not difficult; however, it remains an often misunderstood topic.The key to setting level controls lies in the simple understanding of what you are trying to do. A few minutes spent in mastering thisconcept makes most setups intuitive. A little common sense goes a long wayin gain setting.
A dozen possible procedures exist for correctly setting the gain structureof any system. What follows is but one of these, and is meant todemonstrate the principles involved. Once you master the fundamentalprinciples, you will know what to do when confronted with different systemconfigurations.
Mastering the lingoAudio-speak is full of jargon, but none so pervasive as the decibel. Thoseunfamiliar or rusty with decibel jargon are directed to the sidebar,"Decibel Notation and Its Many Reference Levels."
Mastering gain or level control settings also requires an understanding ofdynamic range and headroom. Dynamic range is the ratio of the loudestundistorted signal to that of the quietest discernible signal in a piece ofequipment or a complete system, expressed in decibels (dB). For signalprocessing equipment, the maximum output signal is ultimately restricted bythe size of the power supplies: It cannot swing more voltage than isavailable. The minimum output signal is determined by the noise floor ofthe unit: It cannot put out a discernible signal smaller than the noise(generally speaking). Professional-grade analog signal processing equipmentcan output maximum levels of +26 dBu, with the best noise floors being downaround -94 dBu. This gives a maximum unit dynamic range of 120 dB - apretty impressive number coinciding nicely with the 120 dB dynamic range ofnormal human hearing (from just audible to painfully loud).
For sound systems, the maximum loudness level is what is achievable beforeacoustic feedback or system squeal begins. The minimum level is determinedby the overall background noise. It is significant that the audio equipmentnoise is usually swamped by the HVAC plus audience noise. Typical minimumnoise levels are 35 dB to 45 dB-SPL, with typical loudest sounds being inthe 100 dB to 105 dB-SPL area. (Sounds louder than this start being veryuncomfortable.) This yields a typical usable system dynamic range on theorder of only 55 dB to 70 dB - quite different than unit dynamic ranges.
Note that the dynamic range of the system is largely out of your hands. Thelower limit is set by the HVAC and audience noise; the upper end isdetermined by the comfort level of the audience. This usable dynamic rangeonly averages about 65 dB. Anything more doesn't hurt, but it doesn't helpeither.
Headroom is the ratio of the largest undistorted signal possible through aunit or system to that of the average signal level. For example, if theaverage level is +4 dBu and the largest level is +26 dBu, then there is 22dB of headroom.
Because you cannot do anything about the system's dynamic range, your jobactually becomes easier. All you need worry about is maximizing unitheadroom. Fine. But, how much is enough?
An examination of all audio signals reveals music as being the most dynamic(big surprise) with a crest factor of 4 to 10. Crest factor is the termused to represent the ratio of the peak (crest) value to the rms (root meansquare - think average) value of a waveform. For example, a sine wave has acrest factor of 1.4 (or 3 dB) because the peak value equals 1.414 times therms value.
Music's wide crest factor of 4 to 10 translates into 12 dB to 20 dB. Thismeans that musical peaks occur 12 dB to 20 dB higher than the "average"value. This is why headroom is so important. You need 12 dB to 20 dB ofheadroom in each unit to avoid clipping.
Preset all controlsAfter all equipment is hooked up, verify system operation by sending anaudio signal through it. Do this first before trying to set any gain andlevel controls. The audio signal will ensure that all wiring has been donecorrectly, that no cables are bad and that no audible hum or buzz is pickedup by improperly grounded interconnections. Once you are sure the system isoperating quietly and correctly, then you are ready to proceed.
Turn down all power amplifier level and sensitivity controls.
Turn off all power amplifiers. (This allows you to set the maximum signallevel through the system without making yourself and others stark ravingmad.)
Position all gain and level controls to their off or minimum settings.
Defeat all dynamic controllers, such as compressor-limiters, gate-expandersand enhancers, by setting the ratio controls to 1:1 or turning thethreshold controls way up (or down for gate-expanders).
Leave all equalization until after you have correctly set the gain structure.
Gain settingsA detailed discussion of how to run a mixing console lies outside the rangeof this article, but a few observations are relevant. Think about thetypical mixer signal path. At its most basic, each input channel consistsof a mic stage, some EQ, routing assign switches and level controls, alongwith a channel master fader. All of these input channels are then mixedtogether to form various outputs, each with its own level control or fader.To set the proper mixer gain structure, you want to maximize the overallsignal-to-noise (S/N) ratio. Now think about that a little: Because of thephysics behind analog electronics, each stage contributes noise as thesignal travels through it. (Digital is a bit different; let's leave that toanother day.) Therefore each stage works to degrade the overall S/N ratio.Here's the important part: The amount of noise contributed by each stage isrelatively independent of the signal level passing through it. So, ingeneral, the bigger the input signal, the better the output S/N ratio.
The rule here is to take as much gain as necessary to bring the signal upto the desired average level, say, +4 dBu, as soon as possible. If you need60 dB of gain to bring up a mic input, you don't want to do it with 20 dBhere, 20 dB there and 20 dB some other place. You want to do it all at onceat the input mic stage. For most applications, the entire system S/N (moreor less) gets fixed at the mic stage. Therefore set it for as much gain aspossible without excessive clipping.
Note the wording excessive clipping. A little clipping is not audible inthe overall scheme of things. Test the source for its expected maximuminput level. This means, one at a time, having the singers sing and theplayers play as loud as they expect to sing or play during the performance.Or, if the source is recorded or off-the-air, turn it up as loud as everexpected. Set the input mic gain trim so the mic OL (overload) light justoccasionally flickers. This is as much gain as can be taken with thisstage. Any more and it will clip all the time; any less and you are hurtingyour best possible S/N.
Once you've set all the input gains and then created the overall desiredmix (involving all sorts of art and science I'm not going to get into),then you must set the output level controls in a similar manner: Advancethe output control until the output OL light begins to flicker. This is themaximum output level.
(Note that a simple single microphone pre-amp is set up in the same manneras a whole mixing console.)
Setting outboard gear I/O level controlsAll outboard unit level controls (except active crossovers) exist primarilyfor two reasons:
They provide the flexibility to operate with all signal sizes. If the inputsignal is too small, a gain control brings it up to the desired averagelevel; if the signal is too large, an attenuator reduces it back to thedesired average.
They act as level controls for equalizers, providing make-up gain in thecase where significant cutting of the signal makes it too small, or theopposite case, where a lot of boosting makes the overall signal too large,requiring attenuation.
Many outboard units operate at unity gain and do not have any levelcontrols - what comes in (magnitude-wise) is what comes out. For a perfectsystem, all outboard gear would operate in a unity gain fashion. It is themain console's (or pre-amp's) job to add whatever gain is required to allinput signals. After that, all outboard compressors, limiters, equalizers,enhancers, effects or what have you need not provide gain beyond thatrequired to offset the amplification or attenuation the box provides.
With that said, you can now move ahead with setting whatever level controlsdo exist in the system.
Whether the system contains one piece of outboard gear or a dozen, gainsare all set the same way. Again, the rule is to maximize the S/N ratiothrough each piece of equipment, thereby maximizing the S/N ratio of thewhole system. And that means setting limits such that your maximum systemsignal goes straight through every box without clipping. Choose between oneof the three methods: OL light, oscilloscope or AC voltmeter. With theconsole or pre-amp set up as described, you now need a convenient soundsource. Use an oscillator (built-in or external) and feed in a tone around1 kHz. (Personally, I hate the sound of 1 kHz, so I prefer something lower,say, around 400 Hz.) Or you can substitute pink noise for the OL light oroscilloscope methods, but pink noise will not work for the AC voltmetermethod because the ACVM will not respond fast enough to catch the peaks.
OL light method. Set the level of the oscillator (NOT the console's orpre-amp's output level; it has already been set!) by turning up its owncontrol, if existing, or by using a spare channel on the console. Turn itup such that the output OL indicator just begins to light. It's a largesignal, on the order of +20 dBu, and would be very loud if you had notalready turned off the amps. What you have now is the maximum expectedsignal level running through the system. From here on, everything will beset so it does not clip with this signal. Once this is done, operators canrun the system as loud as they want without fear of feedback or distortion.
Oscilloscope method. Using the OL light is a fast and convenient way to setthis level. However, a better alternative is to use an oscilloscope andactually measure the output to see where excessive clipping really begins.This method gets around the many different ways that OL points are detectedand displayed by manufacturers. There is no standard for OL detection. Ifyou want the absolute largest signal possible before real clipping, youmust use an oscilloscope. And, of course, if the unit or console does nothave an OL indicator, then an oscilloscope is mandatory to establish theactual clipping point. For a really clever alternative to an oscilloscope,see "Piezo Magic" in the Spring 1996 issue of The Syn-Aud-Con Newsletter.
AC voltmeter method: If an oscilloscope is out of the question, anotheralternative is to use an AC voltmeter (preferably with a "dB" scale).Instead of relying on the OL indicator to tell you when you have a maximumsignal, you choose a very large output level, say, +20 dBu (7.75 Vrms) anddefine that as your maximum level. Now set everything so nothing clips atthis level. This is a reasonable and accurate way to do it, but is it anappropriate maximum? Well, you already know that you need 12 dB to 20 dB ofheadroom above your average signal. It is normal pro-audio practice to setyour average level at +4 dBu (which, incidentally, registers as 0 dB on atrue VU meter). Because all high-quality pro-audio equipment can handle +20dBu in and out, then this value becomes a safe maximum level for settinggains, giving you 16 dB of headroom - plenty for most systems.
Outboard gear falls into three categories regarding gain and levelcontrols: no controls; one control, either input or output; or both inputand output controls.
Obviously, the first category is not a problem. If there is only one levelcontrol, regardless of its location, set it to give you the maximum outputlevel either by observing the OL light or the oscilloscope, or by settingan output level of +20 dBu as shown on your AC voltmeter.
With two controls it is important to set the input control first by turningup the output control just enough to observe the signal. Set the inputcontrol so that it barely lights the OL indicator, then back it down ahair, or set it just below clipping using your oscilloscope. Now set theoutput control also to just light the OL indicator, or just at clippingusing the scope. No good way exists for optimally setting an input controlon a unit with two level controls using only an AC voltmeter.
For Rane digital audio products, such as the RPM 26 DSP multiprocessor, inwhich input A-to-D metering is provided with the RW 232 software, settingthe input level gain is particularly easy and extremely important: Usingthe maximum system signal as the input, open up the input trim box andsimply slide the control until the 0 dBFS indicator begins lighting. Thisindicates the onset of digital clipping and is definitely something youwant to avoid, so this is the maximum gain point.
Setting power amplifiersIf your system uses active crossovers, for the moment, set all thecrossover output level controls to maximum.
Much confusion surrounds power amplifier controls. First, let's establishthat power amplifier level-volume-gain controls are input sensitivitycontrols, no matter how they are calibrated. They are not power controls.They have absolutely nothing to do with output power. These controlsdetermine exactly what input level will cause the amplifier to produce fullpower. Or, if you prefer, they determine just how sensitive the amplifieris. For example, they might be set such that an input level of +4 dBucauses full power, or such that an input level of +20 dBu causes fullpower, or such that whatever input level your system may require causesfull power.
They do not change the available output power. They only change therequired input level to produce full output power. OK, I feel better.
Clearly understanding that point makes setting these controls elementary.You want the maximum system signal to cause full power; therefore, set theamplifier controls to give full power with your maximum input signal usingthe following procedure:
Turn the sensitivity controls all the way down (least sensitive; fully CCW;off).
Make sure the device driving the amp is delivering max (unclipped) signal.
Warn everyone you are about to make a lot of noise!
Put on hearing protectors and turn on the first power amplifier.
Slowly rotate the control until clipping just begins. Stop! This is themaximum possible power output using the maximum system input signal. Ingeneral, if there is never a bigger input signal, this setting guaranteesthe amplifier cannot clip. (If this much power causes the loudspeaker tobottom out or distort in any manner, then you have a mismatch between youramplifier and your loudspeaker. Matching loudspeakers and amplifiers isanother subject beyond this article.)
Repeat this process for each power amplifier.
Turn the test signal off.
Active crossover output level controlsSetting the output attenuators on active crossovers differs from otheroutboard gear in that they serve a different purpose. These attenuatorsallow you to set different output levels to each driver to correct forefficiency differences. This means that the same voltage applied todifferent drivers results in different loudness levels. This is theloudspeaker sensitivity specification, usually stated as so many dB-SPL ata distance of 1 m when driven with 1 W. Ergo, you want to set thesecontrols for equal maximum loudness in each driver section. Try thisapproach:
Turn down all the crossover outputs except for the lowest frequency band,typically labeled "low-out." (Set one channel at a time for stereo systems.)
If available, use pink noise as a source for these settings; otherwise usea frequency tone that falls mid-band for each section. Turn up the sourceuntil you verify the console is putting out the maximum system signallevel, which should be somewhere around the console clipping point. Usingan SPL meter (turn off all weighting filters; the SPL meter must have aflat response mode) turn down this one output level control until themaximum desired loudness level is reached, typically around 100 dB to 105dB-SPL. Very loud, but not harmful. (Note that one to two hours is thepermissible noise exposure allowed by the U.S. Dept. of Labor NoiseRegulations for 100 to 105 dB-SPL, A-weighted levels.)
OK. You have established that with this maximum system signal this driverwill not exceed your desired maximum loudness level (at the location pickedfor measurement). Now, do the same for the other output sections as follows:
Mute this output section. Do not turn down the level control; you just setit! If a mute button is not provided on your crossover, then disconnect thecable going to the power amplifier.
Turn up the next output section: either "high-out" for two-way systems or"mid-out" for three-way systems, until the same maximum loudness level isreached. Stop and mute this output.
Continue this procedure until all output level controls are set.
Un-mute all sections, and turn off the test source.
Congratulations! You have finished correctly setting the gain structure foryour system.
Now you are ready to adjust equalization and set all dynamic controllers.After equalizing, remember that you must always reset the EQ level controlsfor unity gain as required. Use the bypass (or engage) pushbuttons to "A/B"between equalized and un-equalized sound, adjusting the overall levelcontrols as required for equal loudness in both positions.
Optimum performance requires correctly setting the gain structure of soundsystems. It makes the difference between excellent sounding systems andmediocre ones. The proper method is straightforward: take all necessarygain in the console or pre-amp, set power amplifier sensitivity controlsfor a level appropriate to pass the maximum system signal without excessiveclipping, and set active crossover output controls to correct forloudspeaker efficiency differences. Once you understand what the terms meanand how those knobs and lights function, you're on the road to a soundsystem that sounds great.
decibel Abbr. dB. Equal to one-tenth of a bel. (After Alexander GrahamBell.) The preferred method and term for representing the ratio ofdifferent audio levels. It is a mathematical shorthand that uses logarithms(a shortcut using the powers of 10 to represent the actual number) toreduce the size of the number. For example, instead of saying the dynamicrange is 32,000 to 1, we say it is 90 dB. (The answer in dB equals 20 logx/y, where x and y are the different signal levels.) Because the decibel isa ratio, it has no unit. Everything is relative, therefore it must berelative to some 0 dB reference point. To distinguish between referencepoints, a suffix letter is added as follows:
0 dBu A voltage reference point equal to 0.775 Vrms (u = unterminated,i.e., the impedance is irrelevant).
+4 dBu Standard pro audio voltage reference level equal to 1.23 Vrms.
0 dBV A voltage reference point equal to 1.0 Vrms.
-10 dBV Standard voltage reference level for consumer and some pro audiouse (e.g. TASCAM), equal to 0.316 Vrms. (Tip: RCA connectors are a goodindicator of units operating at -10 dBV levels.)
0 dBm A power reference point equal to 1 mW. To convert into an equivalentvoltage level, the impedance must be specified. For example, 0 dBm into 600V gives an equivalent voltage level of 0.775 V, or 0 dBu; however, 0 dBminto 50 V, for instance, yields an equivalent voltage of 0.224 V--somethingquite different. Because modern audio engineering is concerned with voltagelevels, as opposed to power levels of yore, the convention of using areference level of 0 dBm is obsolete. The reference levels of 0 dBu or -10dBV are the preferred units.
0 dBr An arbitrary reference level (r = re, or reference) that must bespecified. For example, a signal-to-noise graph may be calibrated in dBr,where 0 dBr is specified to be equal to 1.23 Vrms (+4 dBu); commonly statedas "dB re +4," that is, "0 dBr is defined to be equal to +4 dBu."
0 dBFS A digital audio reference level equal to full scale. Used inspecifying A-to-D and D-to-A audio data converters. Full scale refers tothe maximum peak voltage level possible before digital clipping or digitaloverload ("overs") of the data converter. The full scale value is fixed bythe internal data converter design and varies from model to model.