A brief summary
Through these audio experiments we used our own methodology to gather data that we used to calculate the sound attenuation values of SleepMuffs™ and other devices. This methodology was inspired by AS/NZS 1270:2002 Acoustics—Hearing protectors which explains how to calculate SLC80 ratings of different sound reducing devices.
We did this so our future users could get an understanding of what they could expect from SleepMuffs™ and especially a basis of comparison with other sound reducing devices.
Through our experiments we purposely tested, for the basis of comparison, the sound attenuation levels of:
- Blisstil’s SleepMuffs™.
- 3M’s MODEL E-A-Rsoft earplugs rated at an SLC80 of 21 (NRR 32 dB).
- Bose’s QuietComfort® 25 Active Noise Cancelling (ANC) headphones.
These achieved the following sound attenuation values with our method:
- Earplugs: 23 dB
- SleepMuffs™: 33dB
- ANC Headphones: 35dB
As you can see above there is only of 2 dB difference (9.5%) between the earplug manufacturer’s SLC80 rating and our measurements, meaning that our methodology is correct and our measurements have value.
Nonetheless these values are to be used as a guide as attenuation levels may vary from user to user.
Set-up and Methodology
The methodology presented in AS/NZS 1270:2002 Acoustics—Hearing protectors is used to calculate SLC80 ratings. To the best of our ability we tried to follow this methodology to calculate our own sound attenuation values. However, there were situations where we could not due to the complexity of the requirements and the limitations imposed by our surroundings and measuring instruments.
As a testing rig we used an Earphone Audio Response System (EARS) from miniDSP that mimics the human ear and calibrated as per manufacturer’s recommendations.
This differs from the AS/NZS standard which uses real people to measure sound attenuation.
The data we collected from this rig was then analysed through an audio measurement software called REW.
Although these instruments are not industry standard, they worked great for our purposes.
However, we did have to fasten our SleepMuffs™ onto the test rig since the fake ears of the audio response system are on a flat base and Blisstil’s SleepMuffs™ are anlged to better fit humans.
Nonetheless, for fairness we fastened the ANC headphones in the same way as the SleepMuffs™ and pushed the earplugs all the way into the fake ear canals, so they touched the microphones inside.
A sound bar with a built-in sub-woofer (black square) was used to generate Pink Noise. It was placed above the ground. The EARS (black triangles) were placed on a soft surface to minimise reverberation and vibrations.
The AS/NZS standard states that the room used to measure the attenuation needs to be silent and that the test signals consist of one-third octave bands of noise filtered from a pink noise source at specified centre frequencies.
In the case of the AS/NZS standard is seems logical to do this since they use humans, who are unable to discern sound pressure levels (dB) at different frequencies, to determine attenuation levels.
In our case since we are using an audio response system and an audio measurement software we assumed it was not necessary to use the different test signals as we could use a full spectrum signal (20 Hz to 20 kHz) and use the data from measurement software to determine the sound pressure levels at the specified centre frequencies.
Moreover, to counter the lack of absolute silence we decided to mask the surrounding noise with an 89 dB of full spectrum pink noise as seen seen in the image below.
Sound Attenuation Tests
We first started out by measuring our base line sound.
The audio measurement system recorded the following dB at different frequencies (smoothed at 1/3 of an octave):
We then ran 7 individual measurements with each sound attenuating device to reduce errors. The following audio responses are averages (only for viewing purposes not used in calculations) of these 7 measurements (smoothed at 1/3 of an octave):
Calculations & Results
In the hearing protectors standard each test signal is produced for 1/3 octave bands of noise filtered from pink noise which have centre frequencies at the following values
Since we did not have to play sounds at different frequencies because our audio response system is capable of discerning different frequencies from a single source, we averaged the decibels between the upper and lower band limits of the one third octave bands mentioned in the table above. This was done to understand the sound attenuation of the devices at the specified octave bands and mimic what a test subject (human) would hear. The lower and upper band limits can be seen in the following table (source).
Once we had collected the attenuation data from the 7 tests of each sound reducing device, we calculated the mean attenuation and standard deviation of these tests in a similar manner to that presented in the standard.
The following formulas were modified to suit our experiment i.e. in the standard it is based on the result from the different subjects (humans) whereas ours, on the repeated tests performed on our rig.
where X_i is the attenuation for the ith test; and
where d_i is the difference between the mean and the ith test’s attenuation.
This yielded the following results for the different sound reducing devices.
Using the data above it is possible to calculate the sound attenuation related to the SLC80 rating by using the following formula:
Where Al_i is the attenuated level at the ith specified centre frequency.This calculation yields the following sound attenuation values based on SLC80 methodology.
- Earplugs: 23 dB
- SleepMuffs™: 33 dB
- Active Noise Cancelling Headphones: 35 dB
As you can see earplugs achieved a sound attenuation value of 23 dB through our methodology and through a certified laboratory they achieved an SLC80 rating of 21 dB. This means there is approximately a 9.5% difference in these results, which we assume is acceptable. We can extrapolate to say that the methodology we followed is correct and the values presented have value for future users.