02/04/07

2.1 Basics: What is sound? Frequency? Wavelength?

If you are not familiar with how sound works, a brief refresher course may help. Don't be put off by occasional technical jargon. Most of the FAQ includes analogies and examples to illustrate ideas in plain language. (The author apologizes to acousticians everywhere for presuming to summarize their craft in a few paragraphs.)

Sound is a pressure wave traveling in air or water. A sound wave resembles the surface wave made when you throw a stone into a calm pool of water, except that

Sound is usually generated by vibration of an object or surface such as a speaker cone, a violin body, or human vocal cords. The vibrating surface "radiates" pressure waves into the adjoining air or water as sound. (Obviously, sound can also be generated by turbulent airflow, by bubbles collapsing, or by many other phenomena.)

The frequency (number of wave crests per unit time that pass a fixed location) measures the tone or pitch of a sound. For example, a bass guitar plays lower frequencies than a violin. The wavelength, or distance between wave crests, is related to frequency: lower frequencies have longer wavelengths.

In some respects, sound and vibration are quite similar. You might find it useful to think of sound as a vibration traveling through air. Many of the same concepts apply for both sound and vibration, but there are certain significant differences. For example, when sound travels through air, all frequencies of sound travel at the same speed (340 meters per second). By contrast, for some types of vibration traveling through a structure such as a wall or floor, low frequencies travel faster than high frequencies.

How is noise different from sound? Noise is simply unwanted sound. Philosophers wonder: "If a tree falls in the forest and nobody is there to hear it, does it make any noise?" When they phrase the question in precisely that way, the answer is NO for this reason: "sound" is not really "noise" unless someone hears it AND finds it offensive.

2.2 What is active noise control?

The question is usually posed like this: "I heard about a new noise control technology called Active Something-Or-Other. Can I use it to make my house quiet when my neighbor’s son plays 'Black Sabbath' on his electric guitar?" Another variant is “Can I create a silent paradise in my back yard next to a major highway?”

The technology in question is "active noise control," also known as "active noise cancellation" or "anti-noise," and it has been a topic of intense scientific research for several decades. Let's jump straight to the bottom line: yes, active noise control works in the proper circumstances, but no, you cannot use it to noise-proof an entire house.

Active noise control is sound field modification, particularly sound field cancellation, by electro-acoustical means.

In its simplest form, a control system drives a speaker to produce a sound field that is an exact mirror-image the offending sound (the "disturbance"). The speaker thus "cancels" the disturbance, and the net result is no sound at all. In practice, of course, active control is somewhat more complicated.

The name differentiates "active control" from traditional "passive" methods for controlling unwanted sound and vibration. Passive noise control treatments include "insulation", silencers, vibration mounts, damping treatments, absorptive treatments such as ceiling tiles, and conventional mufflers like the ones used on today's automobiles. Passive techniques work best at middle and high frequencies, and are important to nearly all products in today's increasingly noise-sensitive world. But passive treatments can be bulky and heavy when used for low frequencies. The size and mass of passive treatments usually depend on the acoustic wavelength, making them thicker and more massive for lower frequencies. The light weight and small size of active systems can be a critically important benefit; see later sections for other benefits.

In control systems parlance, the four major parts of an active control system are:

Analog controllers may also be used, although they are somewhat less flexible and more difficult to use.

2.3 Is active control new?

The idea of active noise control was actually conceived in the 1930's (see the Lueg patent mentioned below), and more development was done in the 1950's. However, it was not until the advent of modern digital computers that active control became truly practical. Active control became a "mainstream" research topic in the 1970's and 1980's. In recent years, researchers have published technical articles at the rate of several hundred per year. There are now dozens of companies that specialize in active control products, and the topic is widely studied in universities and government research laboratories.

2.4 Are there different kinds of active control?

There are two basic approaches for active noise control: active noise cancellation (ANC) and active structural-acoustic control (ASAC). In ANC, the actuators are acoustic sources (speakers) which produce an out-of-phase signal to "cancel" the disturbance. Most people think of ANC when they think of active noise control; some examples are mentioned below. On the other hand, if the noise is caused by the vibration of a flexible structure, then ASAC may be more appropriate than ANC. In ASAC, the actuators are vibration sources (shakers, piezoceramic patches, etc.) which can modify how the structure vibrates, thereby altering the way it radiates noise. (ASAC is distinguished from ANC only in how it is applied, since in either case you have a controller using actuators to control the response of a plant.)

Active vibration control is a related technique that resembles active noise control. In either case, electromechanical actuators control the response of an elastic medium. In active noise control, the elastic medium is air or water through which sound waves are traveling. In active vibration control, the elastic medium is a flexible structure such a satellite truss or a piece of vibrating machinery. The critical difference, however, is that active vibration control seeks to reduce vibration without regard to acoustics. Although vibration and noise are closely related, reducing vibration does not necessarily reduce noise.

Actually, you can generate your own catchy phrases with the following handy buzzword generator. Simply choose one word from each column, string them all together without commas, and paste the result or its acronym into your document or conversation.

ANC Buzzword Generator

2.5 Is active noise control like noise masking?

Active noise control is quite different from noise masking. Acoustic masking is the practice of intentionally adding low-level background sounds to either make noises less distracting, or reduce the chance of overhearing conversations in adjoining rooms. In active noise control, the system seeks not to mask offending sound, but to eliminate it. Masking increases the overall noise level; active control decreases it -- at least, in some locations if not all.

2.6 How can adding sound make a system quieter?

It may seem counter-intuitive to say that adding more sound to a system can reduce noise levels, but the method can and does work. Active noise control usually occurs by one, or sometimes both, of two physical mechanisms: "destructive interference" and "impedance coupling". Here is how they work:

On one hand, you can say that the control system creates an inverse or "anti-noise" field that "cancels" the disturbance sound field. The principle is called "destructive interference." A sound wave is a moving series of compressions (high pressure) and rarefactions (low pressure). If the high-pressure part of one wave lines up with the low-pressure of another wave, the two waves interfere destructively and there is no more pressure fluctuation (no more sound). Note that the matching must occur in both space and time -- a tricky problem indeed.

On the other hand, you can say that the control system changes the way the system "looks" to the disturbance, i.e., changes its input impedance. Consider the following analogy:

Picture a spring-loaded door - one that opens a few centimeters when you push on it, but swings shut when you stop pushing. A person on the other side is repeatedly pushing on the door so that it repeatedly opens and closes at a low frequency, say, twice per second. Now suppose that whenever the other person pushes on the door, you push back just as hard. Your muscles are heating up from the exertion of pushing on the door, but end result is that the door moves less. Now, you could say that the door opens and that you "anti-open" it to "cancel" the opening. But that wouldn't be very realistic; at least, you would not actually see the door opening and anti-opening. You would be more accurate to say that you change the "input impedance" seen on the other side of the door: when the other person pushes, the door just doesn't open.

(The spring-loaded door is supposed to represent the spring effect of compressing air in a sound wave. This is not a perfect analogy, but it helps illustrate impedance coupling.)

In some cases, destructive interference and impedance coupling can be two sides of the same coin; in other cases destructive interference occurs without impedance coupling. The difference is related to whether the acoustic waves decay with distance traveled:

Sound from a speaker hanging in the middle of a stadium decays (is less loud) at a distance because of "spherical spreading." As you get farther away, the sound energy is spread out over an increasingly large area. Go far enough away and, for all intents and purposes, the sound decays completely down to nothing. On the other hand, sound in a "waveguide" such as a duct can travel long distances without significant decay. There are many situations in which walls, ducts, buildings, roadways, or other surfaces can act as waveguides for sound.

If a control system actuator is close to the disturbance source, destructive interference and impedance coupling can both occur. But what about when the actuator is far away from the disturbance, so far away that any wave it creates decays completely down to nothing before reaching the disturbance? There can still be destructive interference near the actuator, even though the actuator cannot possibly affect the impedance seen by the disturbance. Example: the tiny speaker in an active control headphone will not affect the impedance seen by a cannon firing a mile away, but it can create destructive interference within the headphone.

In some cases, an active control system can actually absorb acoustic energy from a system. Of course, the amount of energy absorbed by the system is usually tiny compared to mechanical losses or other losses in the system, but absorption is one possible mechanism for active systems.

2.7 When does active control work best?

Active noise control works best for sound fields that are spatially simple. The classic example is low-frequency sound waves traveling through a duct, an essentially one-dimensional problem. The spatial character of a sound field depends on wavelength, and therefore on frequency. Active control works best when the wavelength is long compared to the dimensions of its surroundings, i.e., low frequencies. Fortunately, as mentioned above, passive methods tend to work best at high frequencies. Most active noise control systems combine passive and active techniques to cover a range of frequencies. For example, many active mufflers include a low-back-pressure "glass-pack" muffler for mid and high frequencies, with active control used only for the lowest frequencies.

Controlling a spatially complicated sound field is beyond today's technology. The sound field surrounding your house when the neighbor's kid plays his electric guitar is hopelessly complex because of the high frequencies involved and the complicated geometry of the house and its surroundings. On the other hand, it is somewhat easier to control noise in an enclosed space such as a vehicle cabin at low frequencies where the wavelength is similar to (or longer than) one or more of the cabin dimensions. Easier still is controlling low-frequency noise in a duct, where two dimensions of the enclosed space are small with respect to wavelength. The extreme case would be low-frequency noise in a small box, where the enclosed space appears small in all directions compared to the acoustic wavelength.

Often, reducing noise in specific localized regions has the unwanted side effect of amplifying noise elsewhere. The system reduces noise locally rather than globally. Generally, one obtains global reductions only for simple sound fields where the primary mechanism is impedance coupling. As the sound field becomes more complicated, more actuators are needed to obtain global reductions. As frequency increases, sound fields quickly become so complicated that tens or hundreds of actuators would be required for global control. Directional cancellation, however, is possible even at fairly high frequencies if the actuators and control system can accurately match the phase of the disturbance.

Aside from the spatial complexity of the disturbance field, the most important factor is whether or not the disturbance can be measured before it reaches the area where you want to reduce noise. If you can measure the disturbance early enough, for example with an "upstream" detection sensor in a duct, you can use the measurement to compute the actuator signal (feedforward control). If there is no way to measure an upstream disturbance signal, the actuator signal must be computed solely from error sensor measurements (feedback control). Under many circumstances feedback control is inherently less stable than feedforward control, and tends to be less effective at high frequencies.

Bandwidth is also important. Broadband noise, that is, noise that contains a wide range of frequencies, is significantly harder to control than narrowband (tonal or periodic) noise or a tone plus harmonics (integer multiples of the original frequency). For example, the broadband noise of wind flowing over an aircraft fuselage is much more difficult to control than the tonal noise caused by the propellers moving past the fuselage at constant rotational speed.

Finally, lightly damped systems are easier to control than heavily damped ones. (Note: Damping refers to how quickly the sound or vibration dies out. Damping should not be confused with "dampening", which happens when you throw water on something.)

2.8 What is adaptive active control?

Adaptive control is a special type of active control. Usually the controller employs some sort of mathematical model of the plant dynamics, and possibly of the actuators and sensors. Unfortunately, the plant can change over time because of changes in temperature or other operating conditions. If the plant changes too much, controller performance suffers because the plant behaves differently from what the controller expects. An adaptive controller is one that monitors the plant and continually or periodically updates its internal model of the plant dynamics.

Copyright (c) 1994-2007 by Christopher E. Ruckman. All Rights Reserved.