How to measure how much sound a product produces?
Consider the speakers in Picture 1. For the purposes of this article, we will assume that the speakers are producing a constant, steady state noise that does not vary over time.
To measure how loud the speaker sound is, one might believe it is as simple as placing a microphone close to the object and measuring the decibel level as shown in Picture 2.
But where to place the microphone? How far away? The further distance a microphone is away from a sound emitting object, the lower the decibel value will be (Picture 3). Distance certainly affects the sound readings.
In fact, in an acoustic free field, the sound pressure level drops by 6 dB as the distance from the sound emitting object is doubled. Going from 1 meter to 2 meters away would decrease by 6 dB. If a product had a requirement to be 50 dB, but the microphone distance was not specified, one could simply place the microphone far enough away to meet the requirement!
Even if a microphones were placed at the same distance away, the decibel reading could vary depending on the location relative to the object, as shown in Picture 4.
A microphone placed behind the speakers will not read the same decibel level as a microphone placed in front of the speakers.
How can one quantify how loud an object is, but independent of the distance or location of the microphone? The answer is Sound Power.
Sound power attempts to quantify the acoustic source strength of an object, independent of the distance and location of the measurement.
How is this done in practice?
Sound Power Measurements
There are different methods used to quantify the sound power of an object. A common method is to surround the object with sound pressure microphones (Picture 5).
For example, the microphones might be placed around the object in hemi-sphere, to capture all the sound emitted by the object in all directions. By taking an energy average of the microphones, one gets a measurement of sound that is independent of location. See the “Sound Pressure” portion of Equation 1.
To normalize the microphone readings over distance, the surface area of the hemisphere is calculated and then converted into decibels. See the “Surface Area” portion of Equation 1. By calculating the surface area of the hemi-sphere, one can also make measurements independent of the distance.
Equation 1 is the basic formula for Sound Power (Lw), where L is the sound pressure level and w stands for watts (the units in which sound power is reported). Sound power is typically reported in decibels referenced to 1 PicoWatt (1 pW).
The equations has two major parts:
The sound power of an object is always be the same no matter what size hemisphere is used to measure the sound power. The pressures and surface area work in conjunction with each other to make the total sound power always be the same (Picture 6).
As the surface area gets smaller, the microphones are at a closer distance to the test object:
So, the total sound power (Lw) remains the same!
Conversely, as the area increases, the microphones get farther from the test object. The farther the microphones are from the test object, the lower their sound pressure readings.
The sound power equation is setup so that any changes in sound pressures are offset by equivalent changes in the surface area, so the total sound power remains constant.
The final results of a sound power test would be a A-weighted octave spectrum (Picture 7). It has units of decibels referenced to Watts.
Sound power tests are run in a variety of facilities of differing quality and performance. Correction factors can be used to remove some of the variation found between test facilities.
In fact, the Sound Power equation that was presented in Equation 1 assumes that there is no other sound sources nearby. It also assumes that there are no reflective walls in close proximity to the test, other than the reflecting plane of the ground.
In a virtual CAE sound simulation, this could easily be the case (Picture 8). But in real-world practice, there can be reflections and other sound sources. The correction factors, K1 and K2, are used to remove the effects of reflections and other sources, within certain limits. These corrections are performed on an octave band basis.
K1 – Background Correction
When performing a sound power test, a measurement is made without the test object emitting noise. This is called the background noise measurement. This correction is done per octave band. Depending on the levels of the background noise compared to the actual test, a few corrective actions might be made:
K2 - Reflections
Some test environments are not perfectly anechoic. Sound reflects back from areas other than the reflecting plane, causing the sound power levels to be higher than they should be. The amount of reflected noise can be quantified and corrected.
To do this, a reference sound source is measured in the test environment. The reference sound source creates a repeatable, known sound power level. For a given octave band, if the reference sound source should be 90 dB, but 91 dB is measured due to extra reflections, the increase can be corrected.
These two correction factors are subtracted from the sound power value. See Equation 2.
With the addition of correction factors, the sound power equation is now complete.
Sound power attempts to quantify the acoustic source strength of an object, independent of the distance and location of the sound measurements.
There are four main factors taken into consideration when calculating sound power (Equation 2):
Sound power is often used in noise regulations and legal certifications because it is not location or distance dependent. ISO 3744 and other standards have in-depth details on how these measurements are to be performed.