Designing Scientifically
Feb 1, 2005 12:00 PM,
By Tony Warner
Returning to scientific roots can help systems integrators solve tricky design issues.
We’ve all seen it, and if we’re honest, we’ve probably all done it at some point: cutting and pasting elements from a previous design to easily create a new one. At times it’s because we know it’s the correct solution. More often than not, however, we do it simply to save time and expense. Whether it’s caused by a fear of repeating past failures, a lack of understanding of the technology, or just out of ease or cost savings, we have adopted standard ways of solving certain problems and are hesitant to deviate from them.
Finding the best scientific solution to a design problem sometimes means diving back into the textbooks we studied in school.
For instance, we might have already designed an AV system for a theater or a boardroom that we know worked well. It is very tempting to readily assume that putting the same system in another theater or boardroom will be the best solution. We can realize significant cost savings by using the work we’ve already done and pushing out the same design. Although this approach has some merit, I suggest that it has also caused us to become somewhat cavalier and to stray from the scientific principles that should be guiding our designs.
A good parallel that is familiar to many of us is audio mixing. To the amateur facing an audio board for the first time, the sundry knobs and buttons can be overwhelming at best. Consequently, many who take up sound mixing are content simply to learn which buttons to push. This may be passable in forgiving environments when everything is functioning correctly, but the moment anything goes awry, the typical reaction is panic. When the skills of the trade aren’t formed around the science of what’s going on behind the scenes, the results can be catastrophic. In much the same way, when we approach designing without knowledge of the underlying science, we are setting ourselves up for disappointing results.
In most designs that have problems, it is fair to assume that a certain amount of guesswork was in play without consideration of the science behind the technology. When we adhere to the laws of science and don’t try to make equipment and spaces perform contrary to their nature, we can much better predict and control a system’s performance. Implementing the systems within these guidelines is where the engineering or art form comes into play.
As an example of this, I’ve taken a design challenge familiar to all of us: mic placement in audio conferencing systems. These systems can be tricky to design and implement successfully. Few systems are less forgiving, and if done poorly, they can perform so terribly that they are rendered useless and seldom used. How best to design these systems has been debated in numerous ways by countless individuals over the years; yet, they remain among the most problematic systems being designed. Recognizing that, we are quick to look for a standard solution that will work in every situation.
As we take a closer look at the scientific guidelines for mic placement, several factors come into play that should influence our designs. Very good intelligibility is key for conferencing and often defines the success of the system. When we are in a noisy room, we can often decipher what is being said by combining aural cues with visual ones. Our brains interpolate any information that gets lost in the noise of the room. With conferencing, that luxury is lost. For that reason, intelligibility is perhaps the one element most critically affected by mic placement. Getting our arms around this problem is significantly easier when we understand a few fundamental laws of science.
Figure 1: The Inverse Square Law states that when the distance is doubled, the surface area quadruples. Applied to audio, it refers to the distance between a sound source and a listener.
The ever-popular Inverse Square Law is responsible for several of our design guidelines. Simplified, the law states the relationship between two powers or voltages. For our purposes, that is the acoustical power or sound pressure in free space, where every doubling of distance between a sound source and a listener results in a reduction of 6dB of level. Figure 1 shows the total surface area of a sphere quadrupling as the radius is doubled. Therefore, each doubling of distance results in the intensity dropping to one-fourth of its previous value. When plugged into the Inverse Square Law, this results in a reduction of 6dB. This holds true only in free space. When reflective surfaces are added to the equation, the reflections add to the original signal and typically cause a net loss less than 6dB.
Comb filtering is an often-overlooked phenomenon that distorts frequency response through phase cancellation. The phase cancellation is caused by time-of-arrival variations of the same signal. These variations or delays can be caused by multiple mics picking up the same source or by a reflective surface near a sound source, among other things. A slightly delayed signal (less than a wavelength) gets summed with the original source. The resultant sound wave is a compilation of boosted frequencies and reduced frequencies, and it takes on an appearance of a comb (see Figure 2). If this effect is significant, it can cause voices to lose intelligibility and begin to sound somewhat muddy.
Combining the Inverse Square Law with what we know about comb filtering has produced two fundamental relationships that should govern our mic placement. The Inverse Square Law determines the point at which the delayed signal is reduced in level enough to eliminate any significant comb filtering. The first relationship is a 2:1 Rule, which states that the distance between a sound source and any reflective surface must be equal to or exceed two times that of the distance between the source and the listener (or mic). This typically comes into play with mics mounted near, but not directly on, a hard surface. If a handheld mic is placed in a mic clip on a lectern or table, the 2:1 Rule will more than likely be violated. The resultant frequency response will not be a true representation of the sound source. Gooseneck mics typically adhere to this guideline, because they shorten the distance between the source and the mic’s diaphragm. Likewise, boundary mics are not susceptible to this phenomenon. They are mounted directly on the reflective surface, so reflections are essentially eliminated.
The 3:1 Rule is similar but pertains to multiple mics rather than a reflective surface. The distance between a sound source and the nearest mic (or listener) should be no more than one-third of the distance from that mic to another. As the distance between a mic and the source increases, maintaining this relationship often becomes more difficult. Typically, mics mounted on ceilings present more challenges to this rule than those mounted on a table. Even pendant-style or hanging ceiling mics are typically 3ft. to 4ft. above participants’ heads. The direct distance to the next hanging mic down the table is likely to be around 5ft. to 6ft. This is definitely in violation of the 3:1 Rule and will cause a significant amount of comb filtering unless we implement electronic mic gating.
Reverberation is an aspect of sound that both helps and hurts us. The ratio of the original sound to its reverberation is known as the direct-to-reverberant sound. The ratio is dependent on how far the mic, or listener, is positioned between the source and a reflective surface. Critical distance is that point where the reverberation level equals the level of the original source. For good intelligibility, it is recommended that a mic or listener be kept within one-third of that distance. It can be challenging to adhere to this guideline when using ceiling mics instead of table mics.
Figure 2: A typical effect of comb filtering.
Applying these laws to audio conferencing, we can reach several general conclusions: (1) The closer the mics can be placed to the participants, the easier it is to stay within the guidelines for minimal comb filtering and good direct-to-reverberant sound. Placing mics on a ceiling makes this very difficult, if not impossible, to accomplish. (2) Using gooseneck mics for multiple people often results in a violation of the 2:1 Rule governing reflective surfaces. (3) Surface-mounted mics essentially eliminate the possibility of reflective sounds from the table, but increase the likelihood of a 3:1 Rule violation due to the distance increase from the sound source. Many modern auto mixers successfully gate all but one mic at any given time, so this may not be a factor in many instances. Nevertheless, it’s wise to take the rule into account.
Many times, maintaining aesthetically pleasing and ergonomic room conditions makes it challenging to achieve ideal audio equipment performance. This is an important factor and one that can’t be dismissed. While sometimes we have the luxury of dictating these room parameters, more often than not we must take what we are given and scientifically determine the best possible solution in the given environment.
GOING BACK TO SCIENCE
These conclusions reiterate much of what many have argued for years. This is far from a comprehensive list of the factors that should influence us; it is meant only to underscore the importance of incorporating more science into audio designs. These fundamental laws of sound do not change. In certain environments other factors complicate the formulas and laws, but the fundamentals stay the same. We are wise to supplement knowledge of scientific theory with the collection of actual scientific data through measurements. We should maintain a strong arsenal of test equipment and know how to use it. Rather than simply adopting the conclusions of others, we can be stronger designers by gathering the evidence, analyzing it ourselves, and then using it to steer our designing. If we do this, we will find that the pieces often tend to fall into place. Many times simplicity can be obtained by returning to the basics and fundamentals of our trade.
While we have focused only on the science of mic placement, an equal number of guidelines should steer us in designing other systems also. Calculating variables such as projector and image brightness, good viewing areas, loudspeaker coverage patterns, frequency response, and sound pressure level all provide opportunities to brush up on the more technical aspects of our trade. The textbooks on our shelves should be as worn as the keys on our computer.
Too often., we view one particular approach as always being correct. If systems always had a given solution, we would all be out of work. In reality, every single situation and environment is indeed unique. The beauty of our craft is that it’s essentially an art form. There are usually several legitimate solutions for each situation, rather than just one right answer. Two firms may come up with completely different designs, but they both may be completely valid approaches. The key is to not get caught up in the exact details, but to get our arms around the big scientific picture and then be creative. Each situation provides an opportunity to solve a puzzle. When we forgo “cutting and pasting” and instead utilize the scientific principles that provided the foundation for our education, we are then truly functioning as engineers.
Tony Warner, CTS-D, CDT, manages Audio-Visual Design as part of the Special Systems Design Group for RTKL Associates, a worldwide planning, architecture, design, and creative services organization. Warner has worked on projects for the U.S. House of Representatives, the U.S. State Department, and the U.S. Secret Service.
