Sound masking is more annoying when it's done wrong than when it's done right
Most people install sound masking because their architects told them to, or because someone complained about hearing conversations three desks away. The Cambridge Sound Masking Design Guide is the document you follow to actually get it to work instead of making everything louder without making anything more private. I've spent years dealing with failed installations where the system was turned down so low nobody noticed it, and others where it was blasting at 52 phon and everyone blamed the facility manager for the "annoying hum." Here's how I actually use it on a job.
How to work through the Cambridge Sound Masking Design Guide
Start with the space audit. The guide walks you through measuring the existing ambient noise profile first, which most people skip because they think the masking will just cover whatever's there. It won't, not really, if the ambient is already above 50 phon. The goal is typically to hit a masked level between 42 and 47 phon in occupied zones, with a frequency response that follows the FRF target curve — that's the flat-ish slope from about 80 Hz up through 2 kHz where speech energy lives. Everything below 80 Hz gets rolled off to avoid the rumble that makes people's chests feel weird. Everything above 2 kHz gets attenuated because that's where the system starts sounding like a broken air conditioner. Speaker placement comes next. The guide recommends the 155-degree coverage pattern for standard 9-foot ceilings in open offices, with speakers spaced roughly 12 to 14 feet apart on center when using the standard grid layout. I always double-check this against the actual ceiling height. At 10 feet, you can tighten that spacing slightly. At 12 feet, you need to move to the longer-throw emitters or switch to a different mounting pattern or you'll get dead zones between the throws. The zone definition step is where most people mess up. You divide the space into acoustic zones based on function and noise source, not just square footage. A call center next to a break room needs a different masking level than a quiet focus area. The guide gives you the logic for this but it's easy to gloss over. I learned that the hard way on a project where we masked the entire floor as one zone and the break room end was essentially deafening while the far end was inaudible.
After you've got the layout drawn up, the guide moves into equipment selection. The emitters are rated by their output capability and frequency response. You pick based on the zone requirements and ceiling type. Plenum-mount speakers need a different approach than surface-mount ones because the ceiling tiles themselves absorb different frequencies. Standard 2x2 mineral fiber tiles take a bite out of the higher frequencies, which actually helps you meet the FRF target more easily. Acoustical ceiling tiles with higher NRC ratings can throw off your response curve in ways you didn't plan for.
Get the Full Details
A specific problem I ran into
Last year I was working on a renovation where the existing HVAC had been upgraded to a low-noise variable air volume system. The ambient was sitting at about 38 phon, which is quieter than what most open offices see. The Cambridge guide says you need at least a 3 dB headroom above ambient for the masking to be perceptible, so I set the target at 42 phon. The system was installed, measured, and sounded perfectly flat on the spectrometer. Nobody could hear it when they walked in. Two weeks later, the facility team turned the VAV boxes down further during a power optimization cycle, dropping ambient to 35 phon. The masking was now only 7 dB above a quieter baseline and the perceptibility dropped noticeably. People started complaining about "hearing things they shouldn't" again even though the measured SPL hadn't changed. The workaround was to adjust the masking output upward by about 3 dB across the board and to coordinate with facilities on minimum HVAC noise levels. The system needs to sit at a consistent delta above whatever the baseline is. If the baseline moves, the masking needs to move with it. An automated control system tied to ambient sensors solves this, but most installations don't have one. I now specify that as part of the design rather than treating it as an afterthought.
Counter-intuitive things the guide doesn't emphasize enough
First, more masking isn't better. There's a point of diminishing returns somewhere around 50 phon where the system stops improving privacy and starts making people fatigued. I've seen specs that call for 52 phon "to be safe" and then wondered why occupants complained about headaches. The sweet spot for most open offices is somewhere between 44 and 47 phon measured at ear level. Second, the frequency balance matters more than the overall level. A system that's 45 phon but skewed toward the low frequencies will feel boomy and intrusive. A system that's 43 phon with a clean FRF will feel neutral and do its job better. Spend time on the EQ adjustments during commissioning. The built-in parametric EQ on the processors lets you trim specific bands. Use it. I usually knock 2 to 3 dB off the 250 Hz region and check the 1 kHz to 2 kHz area for any peaks that might be causing tonal complaints.
Where this approach breaks down
Sound masking doesn't solve every noise problem. It reduces speech intelligibility, which is useful, but it does absolutely nothing for transient noises like phones ringing, printers jamming, or chairs scraping. If your main complaint is equipment noise, masking is the wrong tool. It also struggles in spaces with very high ceilings above 14 feet unless you use line array emitters or specialized fixtures designed for tall volumes. The guide covers this but the recommendations get less practical and more expensive. Another limitation: masking doesn't work well in partially occupied spaces where large areas are empty and others are dense. The sound field doesn't equalize fast enough. I've seen this in co-working environments where one corner is full and another is vacant. The masked level in the empty area ends up higher than needed and the occupied area doesn't get much benefit because the source distribution is uneven.
Getting the design guide
The Cambridge Sound Masking Design Guide is available through the Sonus Cambridge portal. It's a PDF, roughly 40 pages, with spreadsheets attached for the zone calculations and emitter spacing. The spreadsheet tools are actually useful if you plug in the correct room dimensions and ceiling heights. They're not magic — they still require you to understand what the numbers mean — but they save you from doing the math by hand and they flag issues like insufficient emitter count for a given zone area before you order equipment. Download it, print the FRF target curves, and keep them at the job site during commissioning. Measuring to a curve on paper is more reliable than trying to remember what "flat enough" sounds like after you've been in a loud space all day.
