Pure Tones Blog Post-Acentech
Blog > The Commonwealth’s Pure Tone Requirement May 29, 2019 |

The Commonwealth’s Pure Tone Requirement

In Massachusetts, the law stipulates that certain sources of sound cannot cause noise. What is defined as noise you might ask? Well, noise in the Commonwealth occurs for either of two conditions: sound from a source that causes an increase of 10 dB above the ambient/background sound, or if the sound produces what is referred to as a “Pure Tone.” We are going to dive into the second condition in this blog.

What is a Pure Tone? BBN founder, Leo Beranek refers to pure tones as “the simplest periodic sound”. His definition is “a Pure Tone sound is a pressure disturbance that fluctuates sinusoidally as a fixed frequency”. Another distinguished acoustics professor, Cyrill Harris gets right to the point and defines Pure Tone as “a sound composed of a single frequency”.

Let’s try a couple of examples of sounds that contain a Pure Tone. The example I use when speaking with clients is squealing brakes or the sound of fingernails on a chalkboard. My Noise & Vibration Group colleagues offered a few other real-world examples such as the sound from a tuning fork, a church bell, a boat or train whistle and sound from hydraulic systems. Likewise, if you have ever had a hearing test, they make you listen to sounds of certain frequency then change the amplitude and frequency. Those sounds are pure tones and are usually centered at certain frequencies, called octave bands.

In my opinion, when a person is bothered by sound, it is more likely the presence of a Pure Tone that is bothering them rather than just the sound level. At the wrong frequencies, a Pure Tone can be a highly annoying sound. The sirens on emergency vehicles use a modulated Pure Tone because they are easily heard over background sound. You may be asking yourself; how do I know if I have a Pure Tone? If the noise you are bothered by has a buzz, shrill or whine then it likely has a Pure Tone.

The acoustical character can be measured easily, but you need a sound level meter that has frequency-dependent measurement capability. This includes octave band, one-third octave band, FFT (Fast Fourier Transform) or spectrum analysis capability. Since the measurement of a Pure Tone is independent of its amplitude accuracy, a smart-phone application, such as Analyzer can be used to provide accurate Pure Tone analysis.

Pure Tones Blog Post-Facility Exterior-Acentech

I have been involved in many community noise investigations from both the community side and the industrial side. In 2009 a suburban community complained of bothersome noise from an industrial abutter over the period of a few years. Noise measurements were taken by a consultant to the company, but they did not register as a Pure Tone by the MassDEP standards. I was subsequently hired by the community group to conduct measurements on their behalf. What I found was very interesting. My measurements showed that using the Commonwealth methodology, there was, in fact, no Pure Tone. This involves measuring the sound in octave bands which is a coarse sound measurement. A Pure Tone is recognized when the sound pressure level in any octave band is more than 3 decibels above both adjacent bands.

Back in 2009, it was obvious to my ears that the facility in question was producing a Pure Tone. I arrived at the client and got out of my car and immediately heard the problem. Like an emergency vehicle siren, a Pure Tone in the mid frequency, say above 250 hertz is obvious. Because of what my ears were telling me, I acquired finer resolution data including both one-third octave band and FFT (narrowband) measurements. Data measured using finer 1/3 octave bands showed likely tones in two bands; 400 and 800 hertz. Measurements using even finer FFT/spectrum analysis produce the exact frequencies and, in this case, we had tones at 363 and 726 hertz. In this case, the 726 hertz Pure Tone is exactly two times the 363 hertz Pure Tone. The higher frequency tone is referred to as a Harmonic of the lower frequency tone.

Why did the octave band noise measurements not register as containing a Pure Tone, when finer measurements clearly found the opposite? This was the obvious question from the people involved in the case. Certain sounds that contain Harmonics, a sound that is a whole integer times the primary sound, can mask the octave band Pure Tone method used by the State. Using a medical analogy, it’s like using an X-Ray to detect a torn ligament which can only be seen by an MRI.

For example, a Pure Tone sound in the 250 hertz band with a single harmonic would produce a Pure Tone in the 500 hertz band. If either sound existed alone it would be defined by the MassDEP method as a Pure Tone, but when taken together it results in a situation where no Pure Tone would be classified using the MassDEP methods. As long as more than one tone exists in any two-octave bands, the MassDEP method would fail to identify a Pure Tone condition.

What to do? There are other methods to determine if a Pure Tone exists using finer measurement methods; one-third octave bands and narrowband. My recommendation back in 2009 is still valid today, the MassDEP needs to adopt a better method. Again, using the medical analogy, the doctor may at first have you get an X-Ray, but if that does not find anything, they will prescribe an MRI.

Author: Michael Bahtiarian

“I enjoy my work in this field because it requires you to know a little about nearly everything — including people.” Mike Bahtiarian works in both the Noise and Vibration group…

2 responses to “The Commonwealth’s Pure Tone Requirement”

  1. I have heard this type of approach referred to as band projection (sorry, I cannot cite a source for that). It is quite common in community noise codes to use band projection. Seattle for example, uses a 1/3 octave band projection method. The level of projection which qualifies as a pure tone also varies in three groupings of frequency bands. Despite being a few degrees “better” (in my opinion), it also suffers from the same basic flaws of mathematical tone masking due to harmonics and side bands, and also of false-positives due to insufficient resolution.

    AHRI Technical Committee on Sound has been working to redevelop AHRI 1140, which currently defines a tone metric that nobody uses. The core of the metric is a 1/3 octave band projection, but sprinkles in a dose of other factors. It suffers all of the same flaws listed above, with the added complication that the resulting number is not in dB. We have been evaluating various tone metrics in relation to how member companies test their equipment and have concluded that we are not ready to roll out a meaningful tone metric at this time. ASHRAE is currently funding research in this area and we hope the results of that research will give HVAC equipment manufacturers a clear way forward to provide meaningful tone information on our products. Though to be truthful, I’m not sure how useful such a metric would be for consultants. Even if I had the perfect narrow-band tone metric to describe the sound of HVAC equipment, you would lack sufficient information to predict the tone level as experienced in the listener’s environment because you lack narrow-band sound modeling information to predict everything in between the equipment and the listener.

    But your article isn’t really about predicting tonal annoyance, it is focused on assessing it. That is slightly easier (but still not easy). I understand the frustrations of consultants who must measure according to well-intended codes while your ears tell are telling you that the results are wrong. The point to my feedback is to indicate that in my experience, band projection methods to assess tonal annoyance are always going to have the flaws you have identified in your article. If you are evaluating tones, narrow band metrics hold more hope of success (which you obviously already know). Just be cognizant that predicting narrow band sound is going to be a long ways off in the future. If communities begin adopting narrow band metrics of tonal annoyance, consultants will be better suited to assess tonal annoyance in a way that agrees with their ears! However, they will have difficulty predicting tonal annoyance for clients hiring them to meet these codes prior to installation.

  2. Michael Bahtiarian says:


    Great comments and I agree with all your points. I can further point you to two narrowband methods I discussed in the 2009 paper, which are described in Annex A of ANSI S1.13.

    This “informative” Annex includes two methods: Tone-to-Noise Ratio (TNR) and the Prominence Ratio
    (PR). Both require measurement of sound pressure levels or power spectral density, using an equivalent Type 1 sound level meter, a digital Fast Fourier Transform (FFT) with a bandwidth less than 1% of the frequency of the tone.

    For example, for a 500 hertz tone you would need a bandwidth of no greater than 5 Hz. So if you measured to 1000 Hz you would need a frequency data set with 200 samples vs. 10 to 30 samples with octave band and 1/3 otave band, respectively. So these methods require a lot larger data files.

    Again, I appreciated the feedback.

    Mike B.

Leave a Reply

Filter By



33 Moulton Street
Cambridge, MA 02138

617 499 8000

Acentech is an employee-owned company, majority-owned by its employee stock ownership plan (ESOP).