It’s a hemi-anechoic chamber, to be precise. A silent, echo-free room for measuring computer gear that’s been an impossible dream for nearly seven years: We made it happen with the great support of the SPCR community, and it works… to the tune of 11 dBA! Settle down for a long and winding story…
Oct 23, 2008 by Mike Chin
The story of how SPCR’s far-flung but dedicated community
helped to fund the creation of an anechoic chamber and the acquisition
of new audio test gear suitable for testing low noise products. For details
of this amazing show of community support by silent computing enthusiasts,
please see the articles Make
SPCR Even Better, the accompanying
forum log, the article New
Audio Test Gear, SPCR 2008, and the forum log for the project, 2008
SPCR facility/equipment upgrades: hemi-anechoic chamber.
Silent PC Review began as a hobby site for a technical writer obsessed with
keeping the noise of his computer down. That was me, seven years ago. I was
confident about my ability to hear, listen, analyze and describe the noise of computer components.
I’d played around with piano, organ and guitar from an early age, and spent
years obsessing over high end stereo reproduction of music. The latter helped
hone my critical listening skills, which were surely good enough to discern
noises from computers.
Initially, the instruments to measure what I heard were crude or nonexistent. There was no budget for a serious sound level meter (SLM), spectrum analyzer or any number of other devices that would make the review work both more consistent and easier. Still, given the relatively high level of most "quiet" gear available then, SPCR provided the much-needed service of examining components and systems for acoustics in consistent ways. My review team and I developed practical testing methodologies to communicate and show what we actually heard in my lab. It was very DIY, but effective, well suited to the SPCR audience of those days.
Fast forward seven years: SPCR has now become a well-established "institution" on the web, a central meeting place for anyone interested in low noise computing. Its articles and reviews are often quoted and linked by other tech sites, who appreciate our rigorous testing if not our quiet obsession, and it is visited by numerous computer industry personnel. SPCR has become a kind of research reference library for PC acoustics, and we’ve helped to accelerate the development of quieter computing products. There are fewer hardcore DIYers, mostly because there’s much less need for serious modding to achieve low noise computing. The demand for reviews from both readers and makers of hardware is constant, while the supply of technically savvy PC users who understand and appreciate enough acoustics to write for SPCR continues to be in short supply. Meanwhile, the lab has taken over three rooms on the ground floor of my "Vancouver Special" house, and over the years, I’ve managed to buy, beg, or build enough measurement tools to fill up a closet or two.
SPCR operates out of the ground floor of a "classic" Vancouver Special:
An ugly design meant to maximize floor space on the lot within municipal building regulations.
There are other changes, two that directly affect the work at SPCR.
First, today’s best quiet PC gear is often at a level that only DIY custom modders could achieve when SPCR was launched. The number of such products is not large, but the "silent" sector’s noise floor has dropped significantly. This progress poses challenges in the task of differentiating between products. We can still hear the differences more easily than we can measure them; it’s difficult to measure differences when the competing products are below 20 dBA@1m. We’re pressed at the low level limits of our old Bruel & Kjaer 2203 sound level meter. A new, more sensitive SLM would solve the problem, but then we run into the issue of rising ambient or background noise levels. That’s the second significant change, completely beyond SPCR’s control or influence.
CAUSES OF INCREASED AMBIENT NOISE FOR SPCR
The city of Vancouver has grown, and the traffic of cars and aircraft within audible distance of the SPCR lab has increased significantly. In the past, external noise was light enough that we did not have to wait long to get the few minutes of low noise needed to listen or take measurements for a product. Now, both cars and planes intrude frequently, and there are many days, especially in summer, when it’s too frustrating to measure or record any product acoustics at all.
As in many cities, there are quiet pockets in Vancouver. This neighborhood, near the geographical center of Vancouver a bit south of Queen Elizabeth park, was pretty quiet and sleepy ten years ago. Now there are more drivers negotiating their automobiles through the neighborhood, sometimes in an effort to bypass heavier traffic on major routes. It’s blocked from downtown noise by a ridge roughly 100~125m elevation that runs east-west across the center of the city, with a peak at Queen Elizabeth Park (33rd Avenue), elevation 170m (550′) above sea level. The downtown peninsular 5km to the north is probably just 20′ to 100′ above sea level (similar to Manhattan), and Richmond, across Fraser River 2km to the south is flat river delta land just barely above sea level. As you go north beyond the ridge, noise from downtown (mostly vehicular traffic noise) increases steadily. South of the ridge, at about 90m elevation, this house is insulated from downtown noise, but more exposed to the air traffic to and from the Vancouver International Airport (YVR) some 6km to the southwest.
Pilot’s view of Vancouver from 7300m up, looking towards the west. Most of it is greenish, due to the large number of trees in the city. This contrasts against the browner/grayish patches downtown and vicinity, which is more conventionally urban — asphalt, brick and concrete. (Click on the photo for a larger 1024px wide image to appear in a new window or tab.)
While street traffic is annoying, the house is far enough away from the closest major thoroughfare, Main St, to be modest except during rush hours in the morning and late afternoon. The random local traffic is more intrusive but usually short-lived. We can still get SPL readings of just 18 dBA in our labs in the middle of the day, at least for brief stretches.
900 FLIGHTS A DAY
The noise of air traffic is more insidious. YVR on Lulu Island at the western end of Richmond is Canada’s aviation gateway to Asia and the Pacific, and it’s just 6km away from SPCR. A commercial jet plane taking off is audible from inside the lab almost as soon as it is airborne, if it is heading in an easterly or southerly direction. The planes coming from or going to eastern Canada or the US tend to fly east-west over Richmond, close to the Fraser River. They get closer to SPCR before they move farther away, and there are no hills or mountains to block the noise for a hundred km to the south. There’s a straight line-of-sight (and sound) between these planes and this house. The commercial flights that go north tend not to be as audible because they’re moving in a westerly direction as well, away from this house as soon as they are airborne. There are three main runways for large jet planes. Two run east-west while one runs NW-SE; it’s no surprise that the planes most approach and take off in the E-W path that keeps them in line-of-sight to this house.
Courtesy of YVR
There are numerous smaller (mostly propeller) planes, seaplanes and helicopters at much lower altitudes. These are not all landing at YVR; many land in pads in downtown Vancouver or onto the water at Burrard Inlet north of downtown. Many small planes appear to travel in a north / south direction, often over the immediate neighborhood, sometimes directly overhead. Some are running scheduled flights in standard routes; others are running tourist flights. The noise of these aircraft is of shorter duration, but just as intrusive as the big jets, as they fly at lower altitude.
Over the past year, there appears to have been a change in the air routes over the city for small planes. They were far less frequent in the past. The change could be a natural outgrowth of expanding traffic or business, but I cannot help suspect partisan group action by residents of the posher west side neighborhoods. The natural route between the airport and the Vancouver harbor north of downtown is directly over the west side. It’s not inconceivable that influential residents of the west side neighborhoods lobbied municipal and airport authorities to redirect traffic away from their neighborhoods.
The number and frequency of smaller planes over SPCR is difficult to ascertain, but they contribute to the noise as well.
(Copyright photo courtesy of conlawprof)
YVR states that the annual number of aircraft movements has increased only marginally over the last 15 years, from 290,297 in 1992 to 328,563 in 2007. However, the number of passengers passing through VYR during the same period has risen from 9.9 million to 17.5 million. Similarly, cargo movement has increased from 144,000 tons to 226,200 tons. This suggests that the increase in aircraft movements is due mostly to large commercial jets, which of course, cause the greatest noise. The average number of aircraft movements works out to be 900 daily. A study of daily flight schedules at YVR shows there are only a handful of flights between midnight and 6AM, so the vast majority of the 900 flights are in the 18 hours from 6AM to midnight. It works out to be about 50 flights per hour, on average.
Sunday Oct 5, 2008 between 8 and 9:30, I stationed myself with laptop and SLM on the south-facing sundeck and heard some 30 jet planes. Almost all of them were departures. My guess is that a plane is loudest when taking off because all the engines are going full blast while the plane is accelerating and climbing. In contrast, when landing, the engines are at much lower power and decelerating; only when the brakes are jammed on at touchdown does the noise actually increase, and by this time, it’s not audible from SPCR.
I could not see every plane, especially the landings. Each departing plane was audible for a minimum of about two minutes; most of the visible larger commercial planes headed east or SE could be heard for around five minutes. The recorded minimum SPL of these planes was about 50 dBA; the maximum was 75 dBA. When no planes were audible (a rare pause), the ambient dropped to about 40~45 dBA. This was a Sunday when there are usually somewhat fewer flights. YVR’s web site lists real-time info on passenger flight arrivals and departures. In the 24 hours following 9AM Oct 5, 332 departures and 304 arrivals were listed.
Commercial jet liner seen south of SPCR sundeck about one minute after takeoff.
Same jet, two minutes after takeoff. It stayed audible almost five minutes in total.
There was also a strange occurrence of a NAV Canada Bombardier CRJ200 twin-jet plane repeatedly flying loops over Vancouver. It flew almost directly overhead about a dozen times at probably no higher than 3,000′ altitude. This has been noted in the past; it appears to happen about twice a year.
Unbelievably annoying NAV Canada jet that flew a dozen passes at lower altitude over SPCR in about one hour.
(Cropped photo captured with Nikon D80
at 135mm – 7.5x zoom.)
Lest you jump to the conclusion that SPCR’s neighborhood must be incredibly noisy, I hasten to assure you that it is not. The average resident (not cursed with an inconsiderate noisy neighbor) would tell you that it is quiet and peaceful, even though planes and cars are audible from time to time. The lab is also on on the ground floor, which is considerably more insulated than the upstairs south-facing deck.
It’s the takeoff noise of larger jets that is most audible; these must average about 20 an hour. Trying to measure and records sounds under 30 dBA@1m — and often under 20 dBA — makes airplane noise difficult for SPCR. It’s the main problem that drove me to find the solution in a soundproof room… but that’s much easier to say than it was to build.
THE SOLUTION: AN ANECHOIC CHAMBER
An anechoic chamber is a room specially treated to eliminate internal sonic reflections, or echoes. The etymology is obvious: An, meaning no, and echo: No Echo. A corollary is that it should be resistant to noise interference from external sources; ie, soundproof. I’d fantasized about building one like an underground bomb shelter beneath the garage and back yard, but that could only remain a fantasy for all kinds of reasons — not least of which was a guaranteed divorce! Last winter, it occurred to me that a more realistic approach is to convert an existing room in the house. Was this feasible? The question deserved some exploration.
Since starting SPCR, I visited several anechoic chambers, two of which I’ve written about. The first is a full anechoic chamber at the University of British Columbia right here in Vancouver. This is one in which all internal surfaces of the room are treated to eliminate reverberation completely. In late 2002, I sought out the assistance of Professor Murray Hodgson at the Acoustics and Noise Research Group of the University of British Columbia; the man in charge of the UBC anechoic chamber. I was able to corroborate with Dr. Hodgson for a while, engaging some of his students in projects for SPCR. The close proximity made it seem possible for SPCR to establish a long term working relationship to access the anechoic chamber; alas, it is always busy with student projects, which take precedence over any non-UBC activities. Interestingly, it was not ideal for SPCR because the chamber is housed in a building that has a massive HVAC fan that’s going almost all the time. Only when that fan is turned off does the ambient noise level drop low enough to measure the quietest computing products accurately.
The small UBC anechoic chamber is local; alas, too busy for regular access by SPCR.
The second anechoic chamber I wrote about is some 30~50′ underground, at Gigabyte’s headquarters in Taipei. Gigabyte’s anechoic chamber is a pretty serious affair. It is a fairly large 5.5m x 4.8m x 3.1m hemi-anechoic room, which means the floor is deliberately left reflective, and the other five boundary planes are treated to be completely sound absorbent. This is the type of design most commonly stipulated in recent acoustics-related testing standards. The chamber was constructed at the end of 2002, and its sonic qualities are checked every two years by an independent consultant, Industrial Technology Research Institute. The residual ambient noise level in the chamber is reportedly 15 dB, unweighted, which is extremely low.
The residual noise in the Gigabyte chamber is an extremely low 15 dB, unweighted.
Like most anechoic chambers, both of the above have an outer skin of thick steel… just one of the umpteen technical features used in anechoic chamber construction. Discussions with various personnel in the acoustics industry revealed that typical anechoic chambers start at about $125,000. Over $200,000 is not unusual for a chamber of the size of Gigabyte’s.
PLANNING A DIY ANECHOIC CHAMBER
SPCR’s anechoic chamber could not hope to approach the above professional examples. There’s the obvious budgetary constraint. Then, the practical issues: Firstly, the available room, already being used for testing, is far smaller, just 12′ x 10′ with an 8′ ceiling. Secondly, it was unrealistic to consider any structural changes to the house. So what could be done? There are many web articles on soundproofing techniques for condo units and apartments, treating rooms for home theater use, etc. The information in many of these articles is relevant and highly educational, but there’s nothing specific on the web about building an anechoic chamber. No one seems to have done it successfully and posted the procedure in detail on any publicly accessible web site.
Once I decided to embark on this project, I listed basic objectives:
The first objective was obvious, and we’ll get to that soon enough. The second had other implications: Does SPCR’s audio measurement and recording gear have the capability to deal with quieter sound levels? The short answer: Probably not. The long answer is in the form of a separate accompanying article, Test Equipment Upgrades: SPCR 2008.
A lot of research was done to study how these objectives could be achieved with minimal work and cost. Many people provided valuable advice and information that helped me understand better the issues involved.
There are four basic methods used in soundproofing:
1. Mass barrier
2. Air barrier
4. Mechanical decoupling
Sorting through the mountains of information, the best practical options given the existing constraints appeared to be:
The room-in-a-room concept is not new, and it’s been used in many applications. Ambitious home theater rooms employ this design, as do recording studios that need to block sound transmission, whether coming in or going out. One of the more detailed web articles on this topic was posted by Revolution Audio, a Canadian store for home recording studio products and services. Building a Home Recording Studio is in two parts, one on the Magical Floating Floor and another on the Sound Proof Walls. These articles describe an approach very similar to that adopted for SPCR.
The concept is simple enough:
Russ Kinder, now a project manager at an architectural firm, has helped SPCR since its inception. Russ provided a cutaway drawing of construction suggestions based on a group discussion with his mates.
Floating room-in-a-room construction suggestions.
Regardless of the fine details, after the inner room is built, the internal space would be up to a foot smaller in each of the three dimensions, which would result in a working area of about 9′ x 11′ with a 7.5′ ceiling. AC outlets, a small closet in the room, lighting — all these would have to be dealt with. My brother in law Rob, a home builder with extensive experience building home theater rooms in high end homes, explained the procedures needed to build a room of this type. According to Rob, it is much easier to build from the inside out, but with a minimum crew of two, it is still doable. All the construction would have to take place step by carefully planned step inside the room.
Once this inner room is built, its walls and ceiling would be lined with sound absorbent materials to eliminate echoes. A 200 Hz cutoff frequency would require each sound absorbent wedge to be 1.6′ tall, which would reduce the room width to under 6′. This is too small. A 300 Hz cutoff frequency would require wedges of about 14", leaving a working space of roughly 7′ x 9′ with a 6′ ceiling. Tight, but workable. Foam wedges similar to the ones shown below would be ideal as they can be left uncovered, unlike fiberglass wedges.
Small anechoic chamber with foam wedges in a Vancouver corporate office. There are extra wedges strewn about.
The forum thread 2008 SPCR facility/equipment upgrades: hemi-anechoic chamber begun on May 27 provides a more detailed, day-by-day journal of what’s covered below in more summary form. Most of the work was done in July, but there was preparatory work before… and a long fallout after. It’s one thing to start with an empty room; it’s a whole nuther thing to start with a room chock full of stuff. All the furniture, gear, parts and components had to be moved out and squeezed into the other two main lab rooms which were already full of other stuff. This made June a total grind. August was mostly about cleaning up. The SPCR lab has never been one of those tidy, spotless hospital white places; after the dislocations of moving in June and July, it’s in greater shambles than ever, and we still stumble around looking for things.
* * *
first point of attack was the window in the room. It’s an aluminum frame double-pane sliding window that does not fit well. There’s no question that it’s the most direct route for external noise to get in.
The 4’x3′ window which faces the house next door is a sound leak.
The window was blocked up with a 1.2" thick piece of MDF board with heavy weather-stripping gaskets all around to make an acoustic seal. Long, heavy duty plastic straps were used to clamp the board to the bars on the window.
Window boarded up.
Laptop drive on the ledge for scale.
There’s a ledge shelf across the wall, at the level of the bottom of the window. This is actually the concrete perimeter foundation of the house. It’s up around 4 feet high. Sound transmission through lower portion of the wall is very low. Most sound transmission from the outside occurs through the window and the top half of that wall. All the other walls of this room are inside walls; there is at least one room plus an outer wall beyond them, so sound transmission from the outside via the other walls is low.
From the outside. The window was left unlocked to periodically ventilate the airspace so moisture does not build mold. It rains a lot in Vancouver, so this is a potential problem. The pebble rock stucco external wall provides much better sound attenuation than a typical internal wall.
The drop in noise level after this window fix was shocking. It became much quieter. At night past 10PM (when it’s quieter everywhere), the new acoustic measurement system reported around 12~13 dBA in the empty room. The old B&K 2203 SLM was unable to show any readings below about 15~16 dBA. This may be directly related to internal electronics noise; ie, it’s at its internal noise limit. When a fan was turned on at ~400 RPM a meter away from the mic, the SLM read 18 dBA, and happily, so did the B&K. This tends to confirm my conjecture about the B&K’s internal noise.
A THICKER WALL
A decision was made to increase the thickness of the upper portion of this wall to match the bottom portion. The logic was simple: If blocking the window with 1.2" thick MDF board made such a big difference, perhaps making the whole wall thicker would help even more. The photos below show how this was done:
|Here’s an acoustic interim report as a MP3 file in the room at this stage. It’s easy to hear the room acoustics in this recording, especially over good headphones. It’s a 2.7mb file. Caution: There are some loud handclaps, so don’t set the volume too high!|
The isolation from external noise was further improved, but no measurements were done to quantify the changes. Because the improvement was dramatic, and so much work was still needed for the room-in-a-room, it was too tempting to just install the echo-eliminating damping materials without any further soundproofing work.
Quotes from suppliers for acoustic foam damping wedges were discouraging. The lowest cost for wedges cut to fit the room perfectly for a 300 Hz cutoff was over $20,000, well over the funds earmarked for the project. Smaller wedge for a 400 Hz cutoff still cost $11,000, which was the entire budget. This was not a viable option; a cheaper alternative had to be found.
Damping materials are shaped into wedges for anehoic chambers because they maximize the surface area of absorption, and provide progressively increased density as sound waves approach the wall to ensure a good impedance transition. One of the objects is to create even absorption of sound at all frequencies. Discussions with several acoustics pros who had hands-on experience with anechoic chambers, including their construction, suggested that volume and mass are more important if linear frequency response is not a goal.
Chris N. Strahm, the principal engineer behind LinearX Systems, who created the industry standard loudspeaker development tool called LEAP and many other computer-based acoustics tools, pointed to photos and descriptions of his company’s anechoic chamber. This anechoic chamber does not use the traditional wedges. It is fully lined with 2′ cube modules made from staggered layers 3" thick fiberglass panels. A primary objective was to maximize the sheer volume and weight of damping in the chamber.
LinearX Systems. "A rectangular wedge design was chosen for ease of construction and yields virtually 100% material utilization. The modules were then covered with fireproof stage lining material in addition to the sprinkled room. Over 1,000 lbs of fiberglass cover the walls, ceiling, and floor of the chamber."
The results of this design are impressive and well documented on the linked pages, but the prospect of working with fiberglass was daunting. Anyone who’s worked with fiberglass can tell you that it’s nasty stuff. Microscopic glass fibers get everywhere, into your clothes, skin, eyes, nose, and lungs. It can be painfully irritating, and long term health effects could include cancer, though this is still unproven. I wasn’t about to play with any of the stuff, never mind a thousand pounds.
UltraTouch, mentioned earlier, is a replacement for fiberglass batting insulation made almost entirely of recycled natural denim and cotton fibers. Its acoustic properties are not that well documented, but it’s promoted as being better than fiberglass. Cotton was an acoustic insulation long before fiberglass came along, and it’s still used sometimes for that application. Many of the consulted acoustics pros use it regularly; one called it Blue Fill.
Some 700 pounds of UltraTouch was ordered from a local distributor in Abbotsford at a cost of about $2,000 and delivered via truck to the back lane. There were 15 bags, each containing 10 pieces of 48" x 16" x 8" batting. They filled one side of the two-car garage. It has a thermal insulation value of R30. Fiberglass batts of the same R-value are usually 9.5" thick.
700 lbs of
UltraTouch filled half the 2-car garage.
Initially, the plan was to just stack them up against the walls in the room, suspend a layer across the ceiling on a net of clothesline wire, then do a quick check of external noise transmission. The work on the ceiling was tedious, but successful in the end. It took a couple of days complete, what with the usual multiple trips to hardware stores for parts that were not anticipated or not purchased in large enough quantity. Some 200′ of clothesline wire was used, for example — much more than originally anticipated. The wires had to hold well over 100 pounds, so it was important to screw large hooks into studs, not just the drywall, which meant locating and drilling into the studs. Sounds easy, but with interior walls in a house that’s probably been slapped together at minimal cost, studs are harder to find than you’d think.
View of blue fill suspended a few inches below ceiling on mesh of clothesline wire. The air gap helps reduce sound transmission. Note batting on open door on right.
It wasn’t simple to line the batts against the walls, either. Stacking them just to try them, as originally intended, two problems immediately emerged: They compress, which means more than 20 were need to reach the ceiling, and there weren’t enough batts to cover the walls and the ceiling. Secondly, the batts are far too soft to be stable in such a stack. Wobble, wobble, tumble.
A compressed and wobbly stack of UltraTouch batts.
FRAMES FOR THE BATTS
It took some time and effort to devise an inexpensive and efficient way of installing the batts against the walls. My solution looks like a mad carpenter’s flimsy, minimalist bookcase about 6′ tall, with steel wires to hold the batts in place. Each frame was assembled outside the garage and 10 batts were secured to it. Then the ~55 lb assembly was moved into the room where it was secured to the wall, with about a 2-3" gap between the wall and the batting. Again, that air gap actually helps to reduce sound transmission. A total of 11 frames were employed, 3 for each wall except the one with the entry door. The whole process took many days as I was working on my own, with multiple trips to hardware and lumber stores for supplies. Having a small sedan instead of a pickup truck or van was a real handicap. Plus, I was still working on SPCR editorial content… and trying to enjoy Vancouver’s summer as well, in case it turned out to be short. Tennis as often as possible, hanging with friends, etc. There was also a week-long visit from Christoph Derndorfer, our EU correspondent who was travelling around the US and Canada on an extended holiday. Much beer was being consumed daily to help fuel all the yikyak that geeks (and other human beings) are wont to share. Exhaustion was a regular daily experience.
The batt frames look a bit like ridiculous bookcases.
|Working with UltraTouch: The material may be nontoxic and safe to handle, but it is dusty. It rubs off on clothes and carpeting, and some of it is fine enough to clog up your nostrils and make you sneeze. It also smells funny; not terrible but odd. It’s treated with an EPA registered borate-based chemical that acts as a fire retardant and provides protection against mold, mildew, fungus and pests. Bonded Logic, the manufacturer, makes extensive claims about its benign qualities and lack of toxic ones. After working with the stuff for weeks, I think breathing in all that dust is still probably not great for the lungs. The good thing is that if you’re not handling or moving it, the batts don’t release any dust. I found no anecdotal or documented evidence on the web of any toxicity related to UltraTouch.|
THE SECOND DOOR
After blocking the window, beefing up the external wall and installing all the blue fill damping, the weakest link was the door. It’s a cheap, hollow-core interior door that’s probably more cardboard than wood, as is often the case. It swings into the room from the outside. As per the original plan, a second door, a substantial solid-core door, was hung on hinges on the outside of the door frame. It swings outward. An air gap of nearly two centimeters between the floor and doors was filled using pieces of MDF as a door sill. This door sill has a groove cut into its bottom that acts as a in/egress path for the microphone cable. A small frame all around the inside of the door frame was fitted and screwed into place, with foam gasket strips applied to ensure a good seal when the door is closed. A non-locking knob was installed on the door, along with a latch for the door frame.
The above paragraph took just a couple minutes to write, but it was the first time that this amateur carpenter ever installed door — a heavy door not prehung or prepared with holes for the knob or mortised for the hinges — into a preexisting frame; it was no piece of cake.
After more than two solid weeks of grunt work, the window had been blocked, the outer wall built up, the surface of the walls and ceiling covered with 600 lbs of blue fill, and a second door installed to improve the soundproofing around the entry: In essence, a small hemi-anechoic chamber. Total cost of all the materials ran around $4,000. It was time to assess what had been achieved acoustically.
The assessment process took some time, working with subjective perceptions as well as the new recording / acoustic measurement system purchased before the anechoic chamber was built. This system is fully described in New Audio Test Gear, SPCR 2008. It is a PC-based audio spectrum analyzer with calibrated SLM functions as well as a high resolution audio recorder based around the following core components:
A forum thread I used as a daily blog during the planning and construction of the chamber contain notes upon which many of this assessment is based. As with all complex tools, which an anechoic chamber certainly is, it takes time to appreciate all the various qualities, both positive and negative. This is my assessment after using the chamber for about three months. It may change after a few more months.
1. General – The most striking aspect of the chamber is the absence of echoes. Obvious, right? It’s the primary quality of an anechoic chamber, so what’s the big deal? The big deal is that outside of an anechoic chamber, you will never experience this kind of acoustic space. It simply does not exist in nature or in normal buildings. A few visitors have become uncomfortable when asked to remain quiet for a minute in the chamber. Apparently, the loss of aural cues can make some people feel physically unbalanced. It can be a bit spooky. It’s also very quiet. It typically measures 10~11 dBA through most of the day, although both ground and air traffic noise breaks through and occasionally registers over 20 dBA. When you become acclimatized to the ambient background, it becomes easier to hear the rumbling sound (mostly of traffic 2.5 blocks away) which registers constantly on the spectrum analyzer.
2. Sound pressure levels and Frequency response – The image below shows the frequency spectrum of the chamber when there is no obvious external noise.
Note that the ambient SPL is 10.21 decibels, A-weighted. That’s an averaging of the curve, with the A-weighting filter applied. Note the position of 0 dB, which is defined as the threshold of human hearing sensitivity; above 200 Hz, most of the curve lies over 20 dB below 0.
3. Reverberation time – This is defined as the time for the sound to die away to a level 60 decibels below its original level. What’s ideal depends on the type of room. A general purpose auditorium for speech and music should have a reverb time of 1.5 to 2.5 seconds, depending on its size. A classroom should have a much shorter reverb time, less than a second.
SpectraPLUS has a reverb time measurement function. It requires a wide-range loudspeaker capable of producing pink noise at least 60 dB above the noise floor of the room. First the software activates the noise source long enough to saturate the environment (that is, an equilibrium is reached between direct and reverberant sound throughout the room). Then the sound is turned off and the decay of the signal captured and the reverberation response analyzed in detail. This test was performed in the anechoic chamber using one of the AudioEngine A2 speakers that we reviewed in the spring.
The anechoic chamber’s reverb time measured 50 to 100 ms through most of the frequency band. This confirms what can be heard by anyone; there is virtually no echo at all. It rose to a maximum of about 200 ms (1/5 of a second) in the range below 200 Hz. This increase indicates there are echoes in the low frequencies. It also explains the steep rise in level at lower frequencies; this rise is caused at least partly by "standing waves", not only breakthrough of external low frequency noise.
Is this level of low frequency echo acceptable? Yes, more or less. It would be nice to have less echo at lower frequencies, but the amount of damping material to make any significant improvement is vast. One series of calculations performed by an acoustics expert with anechoic chamber experience suggested that doubling the blue fill in the room might give us an extra 50 Hz before the bass boost in the curve above. The price is another $2,000 and a lot of grunt work. It would also mean losing another 16" in every room dimension. However, there’s a limitation in the size of the room; it’s really too small to push the low frequency resonances much lower. The cost / benefit ratio doesn’t appeal, at least not now.
4. Isolation from external noise – It is much better than the original room used to be. Before any of the sound proofing work was done, any sound outside the window was easily audible. The neighbors use the walkway along that side of the house to gain access to their back yard and side door; they were always audible before. Now, they’re never heard from inside the chamber. In fact, someone can be shouting loudly (as if in an emergency) just outside the window, and they might be very faintly audible from inside the chamber.
A test of sound isolation could be done with a loudspeaker, amplifier and signal source outside the window along with a SLM to measure the SPL, and a SLM to measure the level inside the chamber. This test has not been done, as the wiring, AC requirements, etc are tedious. It would not be surprising for the isolation to measure 70 dB or better through much of the frequency range.
The main noise sources that served as the impetus for this project are planes and cars. They haven’t gone away, so the most pertinent question is: Does the anechoic chamber allow the noise of planes and cars to be ignored? No, not exactly.
A scientific test has not been done, but in general, the noise of most commercial planes intrudes for a shorter duration than in the past. That is, if a passenger plane taking off forced us to stop doing audio recording or measuring SPL for five minutes, now, in the chamber, it stops us for only two minutes. Most of the noise that comes through is in the low frequencies, and these, unfortunately are exacerbated somewhat by the relative silence in the rest of the frequency spectrum, and by the low frequency standing waves or echoes. The same comments apply to automobiles.
Is this level of isolation from external noise good enough? This is a completely subjective question. The answer is all about how patient we are in the lab and how much work there is to do. The point is that we cannot ignore external noise, but as long as we’re willing to wait a couple of minutes, we can usually get the acoustic measurements and recordings done. There’s no question that the original floating-room-in-a-room design would give us better isolation through most of the frequency range. Whether it would help much with the low frequency noise that intrudes most is difficult to assess. The acoustic experts consulted on this question did not come to any clear consensus. A further complication is that STC, which guides acoustic building material choices, only applies down to 125 Hz. It would cost at least another $5,000 and several weeks of physical labor to find out. At this point, I’m not ready for that discovery.
ACOUSTIC RECORDING REPORT
For consistency, and also because it’s very revealing, I made an acoustic recording that begins in the converted kitchen lab just outside the anechoic chamber, then finishes in the anechoic chamber. It’s a downloadable MP3. There are some loud handclaps, so don’t set the volume too loud.
* The earlier acoustic interim report MP3 made in the bare room after the window was sealed up, before it saw any of the blue fill.
* Ambient noise in "kitchen" lab room vs. anechoic chamber – short MP3: 18 dBA in a live room vs. 11 dBA in the dead room. The much lower white noise (random noise that sounds like hiss) is due large to the absence of echoes in that frequency range.
The new anechoic chamber meets most of the initial objectives established for this project. It provides a much quieter space for accurate testing of ever quieter gear, and much higher freedom from external noise. There’s both breakthrough and echoes at low frequencies, but the cost and complexity of eliminating these is daunting, and it may just be impossible to achieve in a room of this size in a conventional house. There is room for improvement, however, so don’t rule out further changes in the future.
The chamber and the new gear acquired to take full advantage of it has been in use for three months. The improved low SPL accuracy and range has improved the resolution of acoustic analysis in our reviews, and the broad capabilities of the audio spectrum analyzer are only just beginning to be utilized. In time, as the number of reviews in the new chamber with the new equipment grows, the SPCR database of computer noise should become even more comprehensive, detailed and accurate than it is today. It’s probably safe to say that SPCR is unique among computer technology web sites in owning an anechoic chamber, the acoustic equipment to make use of it… and hopefully, enough know-how to pull it all together in informative, entertaining articles that do more than promote the latest passing computer gadget.
FUNDING THE PROJECT
In April 2008, I asked the SPCR community for help to fund this project. Corporate sponsors were also contacted. Both readers and corporate sponsors responded generously, with cash donations from readers, and product gifts, randomly awarded to the individual donators, from the corporate sponsors. It was a successful fundraiser; some $11,000 was raised for the project in about one month. The progress of the fundraiser was documented as it unfolded in this lengthy forum thread.
There are too many individual contributors to name them all (some 300), and most want to keep their privacy. To all of you: My heartfelt thanks! Without you, this project would not have gone forward.
The corporate sponsors also deserve recognition, thanks and praise for their part in offering giveaway prizes for individual contributors. Their support is much appreciated by everyone involved with SPCR!
Other SPCR articles of related interest:
New Audio Test Gear, SPCR 2008
Exclusive Tour: Gigabyte’s Anechoic & Thermal Test Chambers
Audio Recording Methods Revised
SPCR’s Test / Sound Lab: A Short Tour
University of BC Fan Noise & Airflow Research Project
* * *
* * *
Selected Web Sources on Soundproofing, Anechoic Chambers and Acoustics
- Brüel Acoustics – Web site of Italian acoustic consulting company with many useful articles about testing, chambers, etc.
- Canadian Building Digests – Canada National Research Council’s publication includes many informative articles about noise and buildings.
- Magical Floating Floor and Sound Proof Walls – Article on Building a Home Recording Studio by Revolution Audio, a Canadian store for home recording studio products and services.
- Sound and hearing – Highly informative section in HyperPhysics, an exploration environment for concepts in physics maintained by Carl R. (Rod) Nave, Department of Physics and Astronomy
at Georgia State University in Atlanta.
- Sound Transmission Class (in the Acoustic Info section) – Good summary of STC by State of the Art Acoustik, a Canadian architectural acoustic design company.
- Eckel Industries – Major designer / builder of anechoic chambers.
- Anechoic Chamber – description of Meyer Sound’s chamber.
- Orfield Laboratories – It’s anechoic chamber is officially the quietest place on earth… better described by Ode Magazine and Audio Junkies.
Web Links for Selected Audio Measurement / Sound Control Products (Many of these sites also provide extensive sound control information)
- Green Glue – A unique damping material applied between sheets of drywall that can dramatically decrease sound transmission. Lots of info and data on applications, along with suggestions for effective use.
- UltraTouch – Replacement for fiberglass batting insulation made almost entirely of recycled and recyclable natural denim and cotton fibers.
- Mason Industries – Maker of vibration and noise control products offers damping pucks made of neoprene or steel springs for resilient floors. The section on architectural engineering has many detailed applications for their products.
- LinearX Systems – Manufacturer of numerous professional computer-based acoustics design and analysis tools.
- Brüel & Kjær – Still the biggest name in acoustics instrumentation such as sound level meters and spectrum analyzers.
- DPA – Mic division spun off from Brüel & Kjær in 1992, now a major quality microphone manufacturer with a huge product range.
- Quiet Solution – Makers of QuietRock, one of the most popular drywall substitutes for improved acoustic isolation. Huge range of other soundproofing products. Plenty of interesting documentation.