Sound Reinforcement, II - System Optimization


The ultimate goal of a sound system is to provide even sound coverage to the entire audience. In a perfect world, the ideal sound system would be a point-source loudspeaker with the ability to overcome the inverse-square law of sound propagation and whose coverage could be strictly controlled to avoid any interaction with nearby surfaces. This is not, as you may have noticed, a perfect world. Thus, today's sound system is comprised of many smaller sub-systems, each in turn comprised of multiple loudspeakers. To paraphrase Bob McCarthy in the Meyer Sound Design Reference book, our goal is to use global solutions to achieve the Best Possible Scenario given many constraints. We can achieve close-to-perfect results using a single loudspeaker for a given listening area; as we add more loudspeakers into the system and account for room acoustics as well as interaction between speakers and sub-systems, we leisurely stroll further and further away from the close-to-perfect situation. But we must try.


Beginning with a single loudspeaker cabinet...

Speaker System - polarity, coverage, placement, selection, gain, frequency response.

Sub-Systems - polarity, coverage, placement, gain optimization, balance with other systems, frequency response with other systems.

Whole System - polarity, coverage, gain optimization, balance, frequency response, transitions.

Tools: SIM, SMAART, TEF, etc. RTA (2D) versus dBSPL (1D) versus FFT (3D).

EQUALIZATIONIn sound reinforcement, equalization plays a considerable part. There are two sections of equalization that may be found in a large sound system. The first stage of equalization is on the console itself-- the channel eq. This eq should be tailored to each individual microphone and will be discussed a little bit more in detail a little later. The second stage of equalization usually occurs after the console and before the amplifiers-- to equalize the sound system to better match the room's acoustics. There are a couple of ways to do this. The first way is to get a pink-noise generator, and patch it in to the system. Pink noise is electronically generated noise that has equal energy per octave. It essentially has the same amount of energy per frequency in relation to what the human ear can hear. As this noise is playing through the system and annoying anyone in close vicinity, a real-time analyzer is used. A specially-calibrated test microphone is plugged into the real-time analyzer and the analyzer will display the levels of each frequency (usually every 1/3 octave) from a given location, dictated by where one puts it. This display will give a graphical representation of what frequencies are accentuated in the house. By taking several analyzers, or at least the data from several different locations, the sound engineer can boost or cut the corresponding frequencies on the equalizers. When the analyzer displays a "flat," or relatively flat house curve, the system is equalized to compensate for speaker imperfections and house imperfections, yielding, ostensibly, a flat frequency response.

There is another way to do this, which doesn't require the use of real-time analyzers (which can be very expensive). This other principle is known as ringing out the system.

RINGING OUT THE SYSTEMThe process is generally done for foldback monitors on stage in large, big, super-loud, professional sound systems, but it also applies to house systems to some degree.

The principle behind ringing out a sound system lies in the fact that feedback is caused simply by frequencies that are accentuated by room acoustics and are fed round and round the sound system. So, if there are many accentuated frequencies, and we de-accentuate them, we'll have a relatively even frequency response and be less prone to feedback.

What one does is to first turn on the EQ unit and set a perfectly flat response curve on the eq. Next, open the mics in question (if wireless mics are a potential feedback threat, use them). Turn up the master faders until the system starts to feed back. Using the EQ, figure out what frequencies are feeding back, and cut those frequencies. Do this for all offending frequencies. At some point, you will be able to boost the overall master gain of the mics. [If the EQ has a "In/Out" switch, try switching it and see just how much feedback you've eliminated, and just how much more system gain you've obtained. It's cool.]

From here we can now taylor the system to sound like what the sound designer wants it to sound like. Remember: a perfectly flat frequency response on paper (or analyzer) is nice and all, but it can also sound like absolute crap. Use your ears. Using a CD of pre-recorded music that the designer is familiar with, we can boost frequencies that we want accentuated. Try to use a variety of music, but concentrate on those types which represent the type of program material that the system will be used for-- i.e. if one is doing Jesus Christ Superstar, look for something rock-ish. Chess is a good choice. (Get the London version. It's better.) One can also use the actual recording of the show in question, but this may become a bit tedious after one hears the show fifteen times in succession during rehearsals and techs. Remember, though, not to overdo it. Save really specific eq-ing for the individual mics.

[n.b. there is a third way of eq-ing a system, utilizing a PC-based computer system and various test-microphones located around the house (Meyer's SIM System or Apogee's Correqt System). The computer will automatically scan all the microphones and find offending frequencies and will control the equalizers themselves utilizing a data network. If you can afford it, go with it. Behringer Specialized Studio Equipment, Ltd., also manufacturers a totally digital 31-band 2-channel EQ/RTA. It's really cool, and not terribly expensive. It has a function on it called "Search and Destroy," which automatically will cut resonant frequencies. Cool. "Search and Destroy." Behringer products are supposed to be very good, but it's hard to tell 'cos Ulrich Behringer's ego keeps getting in the way.]

From here, we are ready to play with our system. Now is the time when we would equalize each wireless mic or stage mic to the sound of the respective performer's voice. Characteristics such as mic placement, mic type, performer location, and performer type will play into this equalization. Try to make the microphone sound natural and well-balanced. Good luck.

DELAY UNITS AND APPLICATIONSDelay units can play a very significant and important role in sound reinforcement systems-- especially for theatrical sound systems. Remember that the goal of the theatrical reinforcement system is to not be heard. Judicious use of delay will aid in not being heard. Concentrating on vocal clusters, this is the theory behind using a delay unit: In 1947, a man named Helmut Haas performed some tests into delayed sound and how we hear it. Haas found that if a human was listening to two sound sources of the same material, and one was slightly delayed with respect to the other and also slightly louder, the ear would associate the direction of the sound source to the sound it heard first, not the sound it heard loudest. However, this only works within a window of 10dB. If the second, delayed speaker is 10dB louder than the first, non-delayed speaker, the effect starts to wane. This is known as the Haas Effect, and is the basis for such image-positioning systems as AKG's Deltastereophony. Thus, the goal is to arrange the system so that the delayed, amplified sound arrives after the direct sound, thereby focusing the audience's attention on the real source. See the Sound Glossary for a more complete explanation of the Haas Effect.

By inserting a delay line before the amplifier to a vocal cluster, we can try to achieve what Helmut Haas proved. By delaying the cluster from 10ms to 30ms, we can shift the audience's perception of the sound source to the stage itself, not to the vocal cluster. This effect is limited by the sound coming from the stage; often if actors are miced well and sound natural while facing front and suddenly turn around to face the back wall (upstage), a sudden shift in perception is noted as the original sound (their voice) is not being heard clearly.

Another application for the delay unit is in venues with balconies. Underneath the balcony, intelligbility from the center vocal cluster will diminish, since there is a distinct shadow caused by the balcony overhead. Thus, insertion of small speakers with good high- to mid-frequency response is necessary to improve intelligbility. However, without the use of a delay line, sound will come from these small underbalcony speakers first-- which will confuse if not irritate the audience. Insertion of a delay line will fix this problem-- by delaying the underbalcony fill speakers, sound will seem to "appear" from the stage first, and the fill speakers will provide for improve intelligibility. Cool, no?

Return to the Sound Index.

Comments, Questions, and Additions should be addressed via e-mail to Kai Harada. Not responsible for typographical errors. - © 1999 - 2003 Kai Harada. 16.11.03

click site click site