September 4, 2015
by Niklas Moeller
From their early uses in commercial offices to relatively newer applications such as patient rooms in hospitals, sound masking systems are becoming a more common component of interior design.
This technology distributes an engineered background sound throughout a facility, raising its ambient level in a controlled fashion. The principle is simple—any noises below the new engineered level are covered up, while the impact of those still above is lessened because the degree of change between baseline and peak volumes is smaller. Similarly, conversations are either entirely masked or their intelligibility is reduced, improving speech privacy and decreasing various disruptions to occupants’ concentration.
Most people have experienced this effect when washing dishes at their kitchen sink while trying to talk to someone in the next room. They can tell the other person is speaking, but it is difficult to understand exactly what is being said because the running water has raised the ambient level in their area.
When introducing a sound to a workplace, it is vital to ensure it is as comfortable and unobtrusive as possible. Otherwise, it risks becoming a source of irritation rather than a way to help solve an acoustic problem, as was the case with the original masking systems developed in the late 1960s, which used white noise generators.
White and pink noise
Though the term ‘white noise’ still tends to be used interchangeably with ‘sound masking,’ it is a very different type of sound than what is produced by modern masking technologies.
White noise is a random broadband sound—meaning it includes a wide range of frequencies—that typically spans the audible range of 20 to 20,000 hertz (Hz). Graphical representations of this type of noise vary depending on the horizontal axis. If it shows individual frequencies, volume is constant. However, if the scale is in octaves, each octave’s volume increases by three decibels (dB) because it contains double the number of frequencies than the one before it. As a general rule, the combined volume of any two sounds of equal volume is three dB higher. Thus, a graph depicting white noise shows either flat or increasing volume.
Most people describe white noise as ‘static’ with an uncomfortable, hissing quality. Those old enough to remember analog televisions compare it to the ‘snow’ broadcast when the antenna lost the transmission signal and instead picked up electromagnetic noise. Unsurprisingly, these early masking systems were typically turned down or shut off soon after they were installed.
‘Pink noise’ is another term often inaccurately substituted for ‘sound masking.’ This is also a random broadband sound, but instead of being equal in volume at each frequency, volume decreases at a rate of three dB per octave as frequency increases. However, because these decreases are offset by the increases created by the doubling of frequencies in each octave, pink noise is constant in volume per octave. Subjectively speaking, this sound is less hissy than white noise. On the other hand, the relatively louder low frequencies give it a rumbling quality, prompting comparison to the sound of a waterfall.
Given these descriptions, it is understandable why modern masking systems do not emit white or pink noise, or in fact any of the other colors (e.g. brown, blue, or violet).
A sound masking spectrum
A sound masking spectrum—often called a ‘curve’—is engineered to balance effective acoustic control and comfort. It is usually provided by an acoustician or an independent party such as the National Research Council (NRC), rather than by the masking vendor. Though a masking curve includes a wide range of randomly generated frequencies, it is narrower than the full audible range—typically from at least 100 to 5000 Hz, though sometimes as high as 10,000 Hz. Further, the volume of masking frequencies is not equal, and does not decrease at a constant rate as frequency increases.
It is important to understand the curve defines what the system’s measured output should be within the space. Regardless of how the system is designed, its ‘out-of-the-box’ settings, the size of its zones, or the orientation of its loudspeakers (i.e. upward- or downward-facing, sometimes called ‘direct-field’), the sound is influenced as it interacts with elements of the workplace interior, such as the layout, furnishings, and other variables. Therefore, in order for the sound to actually meet the desired masking curve, the system’s volume and frequency settings have to be adjusted. In other words, it must be tuned for the particular environment in which it is installed.
Tuning is handled by a qualified technician after the ceilings and all furnishings are in place, and with mechanical systems operating at normal daytime levels. Since conversations and activities can prevent accurate measurement, it is done prior to occupation or after hours. The technician uses a sound level meter to measure the masking sound at ear height. They analyze the results and adjust the system’s volume and equalizer controls accordingly. This process is repeated as often as needed until they meet the curve at each tuning location.
Achieving spatial uniformity
Most people compare the sound of a professionally tuned masking system to that of softly blowing air. However, there is much more significance to the tuning process than simply providing a pleasant auditory experience.
A sound masking system’s effectiveness is directly related to its ability to closely match the specified curve. Variations in the masking sound can profoundly impact performance, so a specification not only provides a target curve, but also a tolerance that indicates by how much the sound is allowed to deviate from that curve across the space. Achieving consistency is also important for comfort—a uniform sound fades into the background more easily and occupants come to consider it a natural part of their space.
Historically, tolerance has often been set to ±2 dBA (i.e. plus or minus two A-weighted decibels), giving an overall range of 4 dBA. However, advances in masking technology over the last few decades (described below) allow it to be as low as ±0.5 dBA, or an overall range of 1 dBA.
Measuring the impact
Articulation Index (AI) tests conducted between two workstations illustrate the importance of keeping tolerance as low as possible and consistently meeting the specified sound masking curve throughout the facility.
In this example, two occupants sit approximately 4.7 m (15.5 ft) apart within an open plan. The partitions are 1.6 m (64 in.) in height and the ceiling tile is highly absorptive (Figure 1). However, without sound masking, the ambient level is only 40.6 dBA, and the listener can understand 85 percent of the other person’s conversation. When masking is applied, comprehension quickly declines. In fact, for each decibel of increase in masking volume, comprehension drops by an average of 10 percent. With the masking set to 48 dBA (i.e. the typical maximum level for comfort) with a narrow tolerance of ±0.5 dBA, the listener can understand just 14 to 25 percent. When a broader tolerance of ±2 dBA is applied, he or she can understand up to 59 percent—barely an improvement over the unmasked conditions (Figure 2).
Though this example focuses solely on volume, variations in frequencies can similarly impact masking performance.
The importance of spatial uniformity (i.e. achieving tight tuning tolerances throughout the space) is also emphasized by the evolution of sound masking architecture. Since the technology was first introduced in the 1960s, numerous advancements have been made in order to make tuning a more precise and efficient exercise.
Centralized sound masking
The earliest sound masking systems employed a centralized architecture. The name derives from the electronic components used to generate the masking sound (and provide volume, frequency control, and amplification), which are located within an equipment room or closet. The settings established at this central point are broadcast over a large number of loudspeakers—sometimes as many as hundreds. While most offer limited analog volume control at each loudspeaker (usually 4 to 5 settings, in 3-dBA steps), their centralized design means large areas of the facility are nonetheless served by a single set of output settings that offer little or no option for local adjustment.
Technicians have to set each large zone to a level that is best ‘on average,’ because they cannot make precise volume changes in specific areas. Variations in the acoustic conditions across the space and the impact of interior elements cause the masking sound to be too low in some areas and too high in others. If the technician raises the volume to address a performance deficiency in one area, they simultaneously increase it in others because of the sheer size of the zone, which affects occupant comfort. If the technician lowers the volume to boost comfort, they sacrifice speech privacy and noise control. This pattern repeats at unpredictable points across the space, which is why central system specifications typically allow large variations in overall masking volume. Tolerance is typically 4 to 6 decibels (i.e. ±2 to 3 dBA). Further, centralized architecture only provides a global frequency control for each large zone.
Centralized architecture is also prone to a phenomenon called ‘phasing,’ or noticeable variations in the masking level. To try to avoid this problem, technicians employ a dual-channel, interlaced design, ensuring adjacent loudspeakers do not emit the same masking signal. However, it requires two sets of centrally located electronic equipment per zone, which raises the costs.
Decentralized sound masking
Decentralized architecture emerged in the mid-1970s to address the problem of large zone size. Rather than place sound generation, volume, and frequency control in a central location, the electronics required for these functions are integrated into ‘master’ loudspeakers, which are distributed throughout the facility—hence the ‘decentralized’ name.
Each ‘master’ is connected to up to two ‘satellite’ loudspeakers, which repeat its settings. Therefore,
a decentralized system’s zones are only one to three loudspeakers in size (i.e. 30 to 62 m2 [225 to 675 sf]). This distributed design inherently controls ‘phasing.’ Additionally, because each small zone offers fine volume control, local variations can be addressed, allowing more consistent and effective masking levels to be achieved across a facility. However, there are still limits to the adjustments that can be made with respect to frequency, which, in turn, impacts performance.
Further, a technician must make changes directly at each ‘master’ loudspeaker, using either a screwdriver (i.e. with analog controls) or an infrared remote (i.e. with digital controls), making future adjustments challenging. It is advisable to measure performance and modify a sound masking system’s settings when changes are made to the physical characteristics of the space (e.g. furnishings, partitions, ceiling, flooring) or to occupancy (e.g. relocating a call center or human resource functions into an area formerly occupied by accounting staff). The likelihood these types of changes will occur during a sound masking system’s 10- to 20-year lifespan is almost certain. Therefore, a ‘set-it-and-forget-it’ approach is neither practical nor advised. Engineers had to develop a more practical way of adjusting the masking sound.
Networked sound masking
The first networked sound masking system was introduced more than a decade ago. The technology leverages the benefits of decentralized electronics, but networks the system’s components together throughout the facility—or across multiple facilities—to provide centralized control of all functions via a control panel and/or software. Changes can be made quickly following renovations, moving furniture or personnel, maintaining masking performance within the space without disrupting operations.
When designed with small zones of one to three loudspeakers offering fine volume (i.e. 0.5 dBA) and frequency (i.e. 1/3 octave) control, networked architecture can provide consistency in the overall masking volume not exceeding ±0.5 dBA. It also offers highly consistent masking spectrums, yielding much better tuning results than possible with previous architectures. Some networked sound masking systems can also be automatically tuned using a computer, which first measures the sound and then rapidly adjusts the masking output to match the specified curve.
Today, there are several different product offerings within the centralized, decentralized, and networked categories. Some vendors have also utilized a hybrid design, implementing a decentralized architecture in closed rooms (e.g. private offices) and a centralized architecture in the open plan. However, the centralized architecture presents significant tuning challenges within the open plan—an area where occupants rely on masking for speech privacy and noise control.
Updating performance standards
Sound masking is a critical design element for which there is not much room for error. Without a set of performance standards, the client may not achieve the expected level of speech privacy, noise control, and occupant comfort. An ASTM Subcommittee E33.02 on Speech Privacy (part of ASTM Committee E33 on Building and Environmental Acoustics) is currently working on such a standard: WK47433, Performance Specification of Electronic Sound Masking When Used in Building Spaces. It is also in the process of updating:
In the meantime, a minimum performance guideline is to require the masking sound to be measured in each 90-m2 (1000-sf) open area and each closed room, at a height between 1.2 to 1.4 m (4 to 4.7 ft) from the floor (i.e. at ear height rather than directly below a loudspeaker). Some systems can adjust for smaller areas, but this is an acceptable baseline. Masking volume is typically set to between 40 and 48 dBA, and the results should be consistent within a range of ±0.5 dBA or less. The curve should be defined in third-octave bands and range from 100 to 5000 Hz (or as high as 10,000 Hz). It is a reasonable expectation to have ±2 dB variation in each frequency band.
The vendor should adjust the masking sound within the area as needs dictate, and provide a final report verifying the results. The report should indicate areas where the masking sound is outside tolerance and why (e.g. noise from mechanical equipment or HVAC).
Today’s design trends and increased occupant densities mean clients may be more reliant than ever on sound masking to improve speech privacy and control noise in the workplace. The measurements and adjustments an acoustician or vendor makes after a sound masking system is installed are an essential part of the commissioning process.
In all but extreme cases, it is impossible to know whether a sound is providing the expected level of masking. To make this determination, it must be ascertained whether the desired curve has been met throughout the facility. Clients can be assured of their sound masking system’s performance by requesting a detailed report of the tuning results. This document should demonstrate the desired curve is consistently provided throughout the space, ensuring the system’s benefits are enjoyed equally by all occupants.
Niklas Moeller is the vice president of K.R. Moeller Associates Ltd., manufacturer of the LogiSon Acoustic Network sound masking system (logison.com). He has more than 25 years of experience in the sound masking field and also writes an acoustics blog at soundmaskingblog.com. He can be reached at email@example.com.
Source URL: https://www.constructionspecifier.com/tuning-into-sound-masking-technology/
Copyright ©2020 Construction Specifier unless otherwise noted.