To reduce the effects of standing waves and the resulting "dead zones" at different frequencies and locations in the reverb room, dozens of tests are performed in succession while the measuring microphone moves around in all three planes (Figure 1). The absorption in sabins for each band is derived from the amount of change in reverb decay time, and the tests are averaged to yield a single value for each band. Finally, the results are converted to absorption coefficients by dividing sabins by the front surface area of the sample.

In this article I will detail what I feel are limitations of the standard lab tests currently used to measure the performance of acoustic treatment products and materials. Please understand I am not dismissing the value of the current testing standards! The audiophile world is full of products based on "new physics" and other dubious science, and standard lab tests can identify incompetent products. However, I believe a different approach is needed - in addition to the standard tests - for many of the products currently sold for use in smaller rooms.

One problem with the standard tests is that they assume the absorbers will be mounted flat on a wall or ceiling, with only the front surface area exposed to the room and thus actively absorbing. But in practice, flat absorber panels are often mounted spaced away from a room surface because they work better with an air gap (Photo 1). When a panel is mounted with an air gap, its absorption extends to lower frequencies, which is always welcome, and its effective surface area is also increased. That is, sound striking the wall or ceiling at an angle near the panel is reflected into the rear of the panel, which can also absorb. So with an air gap of a few inches or more, the panel behaves as if it's larger than it really is, which is also welcome.

When the rear surface of a panel is exposed, the notion of an absorption coefficient no longer applies because coefficients consider only the front surface area. Similarly, when a flat panel is mounted straddling a corner, which is typical for bass trapping, a triangle shaped opening forms above and below letting sound reach the rear of the panel to be absorbed. Likewise, thicker panels can absorb a substantial amount via their edge surface as well as their front, and the edge surface is not considered either when converting from sabins. Top

Figure 2 shows that a 2 by 4 foot panel that's 4 inches thick has 50 percent more exposed surface area due to the edges. The E mounting standard of ASTM C 423 allows for testing panels mounted with an air gap - for example, ceiling tiles meant to be placed in a suspended grid - but the standard requires that during testing, the edges and rear be blocked from absorbing by applying a skirt around the panels to simulate how the panels will actually perform when installed. Further, when testing with an air gap, the panels are required to be placed adjacent, to minimize the contribution of edge absorption. But in practice, acoustic panels often are not mounted adjacent but rather placed independently at key places such as first reflection points.

Another problem with the current test method is there's no standard for measuring absorbers intended for mounting in corners as bass traps. A panel that measures, say, 5 sabins at 100 Hz when placed near the center of the lab floor typically exhibits two to four times more absorption when placed straddling a corner of the same room. Further, bass traps meant for corner placement are not always flat panels. Some are cylinders, others are wedges, and yet others are flat panels in the shape of triangles for mounting in tri-corners where two walls meet the floor or ceiling. Figure 3 shows several commercial bass traps that are not rectangular panels and thus cannot be compared fairly using the standard methods, nor can their performance be expressed properly using absorption coefficients.

Cylinders have no front surface at all, so it's impossible to derive an absorption coefficient. And while wedges - a popular shape for bass traps made from acoustic foam - do have front surfaces, other surfaces may be exposed during testing but not exposed when installed. Figure 4 shows how a triangle-shaped column has different surface areas exposed during testing versus when installed. Because there's no standard for grouping and placing wedge shaped foam in a test lab, it's anyone's guess as to what surfaces were exposed during testing. Even if the vendor states clearly how the devices were placed and grouped, it's impossible to compare what's measured to other devices that were placed and grouped differently in other labs. Indeed, it seems pointless to measure a device designed for corner placement anywhere other than in a corner! Top

Performance below 100 Hz is what separates the men from the boys with acoustic treatment products. It's easy to build an absorber that works well at mid and high frequencies, and rigid fiberglass or acoustic foam one or two inches thick are common materials that do a fine job above 500 Hz. However, it's much more difficult to design an absorber that is highly effective below 100 Hz, yet these very low frequencies are usually the main problem in small or poorly proportioned rooms. With bass traps, performance below 100 Hz is what matters most!

This reveals another limitation with the standard tests. Most US labs are not certified to measure below 100 Hz because they're not large enough to develop the reverberant field on which these tests rely. Rather, at very low frequencies the reverb room's modes dominate, and those modes may or may not align with the standard third octave test frequencies. Further, lab results can vary as much as 50 percent at 125 Hz, even though 125 Hz is within the range of certified frequencies. Results vary even more below 125 Hz, and I've even seen negative sabins reported due to the inherent inaccuracy of the reverberation room method at very low frequencies. Top

As a designer and manufacturer of bass traps and other acoustic treatment, I needed a more reliable way to assess low frequency absorbers and compare proposed trap designs. The solution I devised is to use room analysis software in the small "lab" room at my company's factory (Photo 2). I use ETF and R+D, which are Windows programs that cost $150 for both from www.etfacoustic.com. For Mac users there's FuzzMeasure, which is equally capable and costs $150 from www.fuzzmeasure.com. These programs are intended mainly for consumers to measure the frequency response of their listening rooms, but they also offer waterfall plots to display modal ringing. When bass traps are added to a room, the peaks are reduced and the nulls are raised, but the modal ringing decay times and the peak Qs are also reduced. A waterfall display is the key feature that makes it possible to accurately assess the performance of low frequency absorbers in a typical bedroom-size space.

In my experience, using ETF in a small room is more useful below 100 Hz than reverb room lab tests. However, it is important to understand that "homemade" tests like these are useful only for experimentation, not for publishing official performance data! One reason is that these tests do not give any numbers. All analysis is done visually, looking at how the decay times are reduced (the "mountains" come forward over time) at each mode frequency, and also by seeing how the Q of each mode is lowered, thereby making the peaks broader. Figures 5a and 5b show ETF screen shots taken in the RealTraps lab with the room empty and with 12 rigid fiberglass panels placed in wall-wall and wall-floor corners.

To assess absorption at, say, 50 Hz, either true reverberation or a natural room resonance is needed. Something has to decay for a change in the decay time to be observed. Fortunately, even a medium-sized bedroom has measurable resonances at very low frequencies, and the change in decay time after adding bass traps is large enough to be assessed reliably. This also reveals a limitation of using ETF and similar software to measure absorber performance. You can assess absorption reliably but only at the room's resonant frequencies. For example, in the RealTraps lab, I can test reliably at 27 Hz, 42, 57, 70, 85, 97, 114, 141, 155, 171, and 183 Hz. To test other frequencies, I'd need to use a different room. But in practice, this approach has proven quite adequate for testing the types of broadband bass traps that I develop. Top

In theory it's possible to determine sabins of absorption from these waterfall plots. But the results would not be truly reliable due to a lack of climate control and isolation from outside noise, such as the rumble of passing traffic. However, the relative difference between the room empty and trapped, or between samples of different proposed trap designs, is valid.

Note that the graphs in Figures 5a and 5b are taken from a series of tests that I did to help me understand how the density of rigid fiberglass affects its absorption, and also what effect the paper membrane has when applied to FRK-type fiberglass. For readers who may be interested, the complete article showing data for 12 different sets of samples is in the Density Report on my personal web site.

Testing panels with an air gap, and not adjacent in a cluster, can be done using the current methods if the size of the air gap and the spacing between panels were standardized. Because this is the way panels are often installed in listening rooms, it makes sense that they be tested this way too. It would also be easy for acoustic product vendors to agree to state absorption for corner bass traps as sabins instead of absorption coefficients. That will put an end to the current practice of proclaiming ever-higher - and physically impossible - amounts of absorption. Agreeing to use sabins only instead of coefficients also makes sense for products meant to be mounted away from a wall or ceiling. This gives a truer measure of a product's effectiveness because the size of the device, plus any potential absorption from the rear surface, is factored into the results.

Devising tests for bass traps mounted straddling corners or tri-corners is more difficult, and measuring accurately below 100 Hz is more difficult still because participating labs would need larger rooms or, at least, identical mode frequencies. But I can wish, can't I? Top