A joint experiment between Quintessence Studio and Hyndman Contracting Ltd
Let's stick with perspective.
A professional world class recording studio should be built well enough to allow for metallica to be recorded at 120 dB and "not wake the baby" if the studio were built next to the nursery bedroom. To this end, they believe in concrete, solid rock, sand beds and mechanical vibrational systems (mega springs and hanging stuff), NASA quality laminar vacuum glass that withstands heat and pressure, and let's not forget space itself, loads of it, at the very least, the use of a sealed air gap. Expect to employ weird shapes, facets, angles and types of materials before budgeting in a swanky interior design.
This is not that. We are a residential construction experiment in progress. I've never built a recording studio, and neither has my builder-contractor Bill Hyndman. We are starting with what we know and we're building off what we've got already. We are hell bent on problem solving using creativity and intuition. And we're aiming for as high a sound reduction in our wall systems as we can create. Upper aimed value is 60dB drop in sound transmission at common wall to the house.
This is not youtube DIY at the opposite end of 'college try' either, it feels like a spiritual test. We are approaching this as experts in our respective fields, intending to demonstrate what can be done at a residential construction-renovation price point. It costs just as much to build a bad recording studio as it does a good one. So what does that mean? And how do you do it?
First thing: get solid on first principles and best practices in your field. I bring to the table the experience of being a musical artist and audio engineer combined with everything I know on the materials science of sound through time, matter and space, further physics of sound and hearing, and architecture for sound in space. Bill Hyndman brings the rest: knowledge of construction & design and code & regulations, beyond houses and their locations, the problems to be faced before they become problems, what materials to use, how to do the best build for your site situation, what to try and what we need to avoid because home inspectors will make us take it out. In three years of working with him, I've not seen him cut corners, not once, it's not in his blood. So we decided, "ok, let's not lose it on overkill, either."
Turns out we are an unusually effective team. I had Captain Bill Hyndman and Mike Vance help me hatch a very basic experiment to check for sound transmission loss - essentially, the progress of our build. Our job is to build a studio that contains sound and doesn't lose sound. We might not have a baby nursery next door, but it is attached to the main house "mothership". With this quick experiment, all we wanted to do was get an idea of how much sound reduction each layer of framing and insulation would provide mid stage during the build, then we'll check again after drywall is mounted and all wall structures are sealed. Keeping it simple, but controlled, we elected to use what was easy - an off the rack air horn and the dB meters on our phone apps. We made starting value measurements in silence to set the calibration differences between our sound meters. The greatest error in calibrations show up at low level measurements. The diagram above shows where we set our controls, incoming sound level from outside, where we're taking readings inside etc...these won't change in stage to stage measurements, to the best of our ability, and I allow a reasonable + or - 2dB error.
From a more formal perspective: testing transmission loss for determining STC values of wall systems in standard settings are made using a sweep of sound from 125Hz to 4000Hz, usually, with the sound source measured as close to surface of one side of the wall system, and another sound level meter as close to surface on the other side (the wall thickness does not count as "more distance" (ie <1m) and is not a concern as part of standard procedure within a 1m limit; the purpose is to assess the wall system's material density that impede sound transmission). Usually the thickness of the wall system is measured and stated for the experiment, and in our case, we expect this value to increase as we build the next layer of insulation. Converting transmission loss values to STC values (as ASTM rated) is pretty complex, measurement wise and mathematically, so we will stick with basic sound meter dB differential one side of a wall system to the other, at one frequency, as this will serve for our purpose. Any drop in dB more than would be expected in empty space (expect a drop of 12dB every 12 to 13 ft) counts as transmission loss.
Following our first 4" layer of closed cell spray foam insulation (and before new windows were installed), we used our air horn (A880Hz) and had Mikey sound it from 20ft away from studio walls outside (see red asterisk locations in diagram). I measured incoming sound at exterior wall (see green asterisks) - average of 76dB over multiple trials. Captain Bill remained indoors at the same position (see blue asterisk in studio) and took readings from all 3 directions incoming. Remember: without any wall systems present, one expects a 12dB drop in SPL just due to a 12 to 13 ft distance anyway. When doing the analysis, I've adjusted by calculation our measured values to situate below the 1m = 3ft distance, from one side of a wall system to the other.
Our casual experiment results showed the following after stage 1 (4" spray foam insulation R24) with measurements of distance apart stated, source 880Hz at 76dB:
North elevation (1 slit window): average 42dB at 12 ft away,
~ 34dB drop
West elevation (2 slit windows): average 45dB at 8 ft away,
~ 31dB drop
South elevation (plywood covering patio door and casement window): average 55dB at 12 ft away,
~21dB drop
Stage 1 wall thickness was estimated at max 6 inches, vinyl siding exterior and the interior has the spray foam. I calculated an 11.5 dB distance correction to the 42dB (north elevation) reading, and 8 dB distance correction to 45dB (west elevation) reading owing to Bill's position to adjust the sound level reading as though he had been standing immediately on the other side of the wall near the green asterisks where I was, 6 inches away. Using the two elevations that had old sealed windows (and not plywood), we report an approximate starting transmission loss value of 22.75 dB (+/- 2dB) for north and west walls at stage 1, specifically for 4 inches of closed cell rigid dry spray foam, allowing for windows making up only about 4% total surface area on average. We know from specs that the triple pane windows and patio door we've installed come in around STC29, meaning a 29 decibel blockage on average over a specified frequency band in sound levels (again, for healthy perspective, this is not NASA, we live in the country).
Stage 2 stud/framing thickness is 6" with a 1/4" gap between stage 1 framing. With R24 rockwool fully filling that width, there remains a substantial spatial air gap of varying volume between exterior spray foam wall and interior rockwool frame wall, where there won't be much transmission of mechanical vibration or sound/air movement. The only places where mechanical vibration can transmit is via the cleats that hold the inframed wall structure to the original exterior wall - to structurally reinforce the perimeter walls hinge point once the squat room floor was removed. We don't expect these contact points to be a problem.
We made stage 2 experimental sound readings, which now also include triple pane windows in place, following the same procedure as stage 1, the 76 dB source (A880Hz) just outside exterior walls at green star locations, Bill measured an average of 16dB on his meter where he stood inside (see top diagram). Added distance from wall edges contributes to an additional ~11.5dB reduction in sound volume. My readings, taken just inside across from green star locations, averaged about 41dB on my meter. Correcting for distances between us, again the same way as I did in stage 1 analysis, and averaging our readings between our 2 sound meters, the SPL at 76dB (outside) just inside the wall system came in at average 34dB at stage 2. This corresponds to an SPL reduction / transmission loss of minimum 41.25 dB (+/-2dB) at 880Hz, compared to 22.75dB (+/-2dB) at stage 1. This means our as yet unsealed wall system without one or two 5/8" drywall surfaces at this stage shows transmission loss of about 41 decibels inside the studio if a horn sounded outdoors at a level of 76dB measured at the house wall, an improvement of 18.5dB from adding 6 inches of R24 rockwool to stage 1 treatment of 4 inches of closed cell spray foam. Conversely, if we are creating live music inside the studio at 80dB, which would be typical of my experience with grand piano, amplified vocals and backing music played back on monitors, that means someone just outside the studio would hear the sound at 80-41 = 39dB, which is the typical sound level of a mostly quiet space, just under ambient level inside a library, where we would expect people walking, whispering and making a bit of hushed noise. Acousticians consider 30dB to be equivalent to a baby's room if no one in the house made any noise to wake her. Studio designers based what they build off of this "unofficial" working standard - if they intend to record metallica at 120dB next to the baby's room, then you better get designing.
So we might even go so far as to wager that 4" of R24 closed cell spray foam reduces sound by about 22dB, and 6 inches of rockwool held up by wood studs (no plywood or drywall) reduces sound by about 18dB. Standalone. Once the two frame wall system sandwich is fully sealed, that might up the game overall. We will have to see...by then, the drywall will add to it.
We chose to make a quick measurement from in house, from lower level on the other side of the triple stud common wall at stage 2. Mikey was in the studio where Bill stood, and sounded the air horn toward where we stood in house at the other blue star location, about 25 ft away. We know from our experiment already that from a 20 ft distance, the horn sound registers 76dB, that's what we're treating as source. For another 5ft away, we calculate it would drop to about 71dB (remember a 12 dB drop for every 12 to 13 ft). We measured an average 25dB interior house reading when the house was quiet already to begin with. That means we experienced a 46dB (+/- 2dB) reduction through unfinished triple stud wall, interior walls, the house steel door etc... In more real music studio terms, already, even if studio monitoring rose to standard 92 dB level (used for standardized sound testing as well), that might suggest that someone standing in house where we were would hear at 92-46 = 46dB, which is about the level of a busy active library if nobody engaged in audible conversation. Now I'm excited to install the heavy duty studio door, and get that double drywall on resilient channel mounted on the common-vestibule walls. Stay tuned for Stage 3 measurements! I might not make it to exactly baby's room specs, but the fact that we're working toward this goal says something, and it means even more that we're not lost in an impossibility either having gotten this far.
A.B. Arnold
Comments