Does Green Glue Actually Work? What 20 Lab Reports Show.
Does Green Glue Actually Work? What 20 Lab Reports Show
I never was a huge fan of Green Glue, I thought it wasn’t worth the money. This year, a real project made me stop and actually read the lab data instead of going back to my catchall “I don’t use Green Glue”. What I found changed how I view the product, and it explains something I had noticed for years without fully understanding why: my designs kept hitting their isolation targets without it even though many people still swear by it.
This is not a takedown. Green Glue works. The interesting question, the one nobody answers out loud, is exactly when it works and when it doesn’t.
The project that started this
We were designing a basement studio with a ceiling height problem. Every inch mattered, and every dollar had a job. The question on the table was simple: does Green Glue earn its material and labor cost in this assembly, or not?
The marketing answer is that Green Glue adds 8 to 10 dB. And that number is real. But it comes from one specific test condition that almost nobody names, and once you see it, you cannot unsee it.
What Green Glue actually does
Green Glue is a damping compound. Sandwiched between two layers of drywall, it converts the drywall's vibration into small amounts of heat instead of letting that vibration radiate through as sound. That is genuinely clever engineering, and it matters most under one condition: the drywall has to be driven hard by vibration arriving through a rigid connection to the framing.
Which raises the question that reframes everything. What happens when the drywall is not being driven hard, because you already decoupled the wall?
What the lab data shows
I went through the full Green Glue test database, 20 laboratory reports from Orfield Laboratories spanning 2005 to 2010, all run under ASTM E90. I pulled every comparison where the only variable was Green Glue itself, meaning the same stud material, the same insulation, the same layer count, the same spacing, with and without the compound. That filter matters. Several of the commonly quoted comparisons quietly change the insulation between the "before" and "after" tests, which inflates the apparent benefit.
Here is what the clean comparisons show.
Within each stud material, Green Glue's benefit shrinks as the wall becomes more decoupled. Wood and steel clusters are separate tests and should not be compared to each other.
On a rigid wood stud wall, Green Glue added 11 STC points, from 44 to 55. That is a massive gain, and it is where the marketing number comes from. Add resilient channel to a wood wall and the benefit drops to 8 points. On steel, a rigid wall gained 6 points, but the same steel wall built with isolation clips and hat channel gained only 2, from STC 62 to 64.
The pattern is consistent with the mechanism. Green Glue intercepts vibration that reaches the drywall through a rigid path. Clips and hat channel interrupt that path before the drywall ever becomes the bottleneck. A well-decoupled wall has already solved most of the problem Green Glue solves.
One honest caveat that I will repeat because it matters: each of these comparisons is a single lab test pair, not a statistical distribution. The right phrase is "consistent with," not "proven."
Where the remaining 2 points actually live
The cleanest single comparison in the entire database is a clip and hat channel steel wall, double 5/8" drywall on both sides, tested with no Green Glue and then with Green Glue at both interfaces. Identical assemblies otherwise. This is the true zero versus full comparison, and because both tests published complete frequency data, we can see exactly where the improvement lives.
In a decoupled wall, Green Glue gained 0 to 2 dB below 200 Hz. Nearly all of its +2 STC came from the 2,500 to 4,000 Hz coincidence dip.
Below 200 Hz, where drums and bass live, the gain was 0 to 2 dB. Nearly everything Green Glue contributed shows up between 2,500 and 4,000 Hz, patching a resonance phenomenon called the coincidence dip. In a decoupled wall, Green Glue is a treble fix, not a bass fix.
For a studio designer, that is the whole ballgame, because bass isolation is the hard part and the expensive part. Which brings up the question that decides whether those 1 to 2 low-frequency decibels are worth paying for.
What does 1 to 2 dB actually sound like?
Three numbers from the hearing science literature put this in perspective.
A perceived halving of loudness takes roughly 10 dB. Green Glue's low-frequency contribution in a decoupled wall sits at the edge of what humans can detect at all.
First, 3 dB is a doubling of physical sound energy, and it is roughly the threshold of what people reliably notice in real-world listening. Second, the just noticeable difference under ideal laboratory A/B switching is about 1 dB, and trained listeners at the easiest levels and frequencies can catch a quarter to half a decibel. Third, it takes roughly 10 dB, ten times the physical energy, before people judge a sound to be twice or half as loud. That last figure comes from S.S. Stevens' power law, published in 1957, and it has been the foundation of loudness science ever since.
So Green Glue's 1 to 2 dB of low-frequency benefit in a decoupled wall sits at or below the edge of human detectability, and nowhere near a perceived halving of sound. If you master records for a living, maybe you would catch it in a direct comparison. Your neighbor will not.
One nuance for the sharp readers: perception does vary somewhat across the frequency spectrum, and a decibel in the deep bass counts perceptually for a bit more than a decibel in the midrange. But the bands where that effect is strongest, below 80 Hz, are exactly the bands where the lab data shows Green Glue adding nothing at all. The nuance is real. It just does not change the conclusion.
While we are here: your STC rating is not what you think it is
This research surfaced something bigger than Green Glue, and if you take one thing away from this article, make it this.
An STC 64 wall is not a 64 dB wall. STC is a unitless rating produced by fitting a reference contour to measured data between 125 and 4,000 Hz. It contains no information about performance below 125 Hz at all.
The highest-rated wall in the database, STC 64, measures 71 dB at 3,150 Hz, 36 dB at 80 Hz, and 17 dB at 50 Hz. The rating only sees 125 to 4,000 Hz.
The best wall in this entire database, the STC 64 clip wall, delivers 71 dB of transmission loss at 3,150 Hz. At 80 Hz it delivers 36 dB. Down at 50 Hz, in the resonance valley every double-leaf wall has, it measures 17 dB. That is a 47 point gap between the number on the spec sheet and the measured performance at a frequency a bass guitar produces every time it plays a G.
This is exactly why we design low-frequency-sensitive rooms from transmission loss data and physics, never from an STC rating. It is also why the basement ceiling decision was easy once we looked at the data: that assembly was already well decoupled, and the money had better places to go.
So, does Green Glue work?
Yes. Genuinely, measurably, and in the walls most people are actually building. If you are treating a condo party wall, an apartment ceiling, or any assembly where full decoupling is not in the budget or the floor plan, Green Glue delivers some of the best dollar-for-dollar improvement available, up to 11 STC points in the lab data. I recommend it for those situations without hesitation.
But once you are building a properly decoupled assembly, the kind we engineer for serious studios, its measurable contribution drops to 1 to 2 dB in the bass and a few dB in the treble, an increment that sits at the edge of human perception. Using it there is not wrong. It is just not where the leverage is anymore.
That is why you will not see Green Glue in our designs. Not because it does not work. Because we now know exactly when it does, and exactly how much it is worth.
Designing a studio and want the data-driven version of this thinking applied to your room?
If you are planning a serious build, take our soundproof site assessment and discover whether where you want to build is going to be working for you or against you when it comes to sound isolation.
Sources: Orfield Laboratories test reports OL05 through OL10 series (ASTM E90/E413); Stevens, S.S. (1957), "On the psychophysical law," Psychological Review 64(3); ISO 226 equal-loudness contours; see also the peer-reviewed just-noticeable-difference literature summarized in our research notes. Lab report data referenced under fair use for analysis and commentary; charts are our own analysis.
