Decibel meters, perception, and you

Why a noise might seem louder than its decibel reading

If you try to measure the decibel level of a noise and get unexpected results, you’re not alone. Your results might depend on your tools and settings.

What do decibel readers measure?

Typically, noise-measuring tools use A-weighted decibels, or dBA. An A-weighted measurement cuts off the extreme high and low frequencies, measuring the typical frequencies that humans can hear. In many applications, this is perfectly fine.

However, as Dr. Erica Walker explained in an interview with Soundproofist, it’s often the low frequencies that really bother people and cause stress. This includes rumbling that might be caused by a toddler thumping across uncarpeted hardwood floors, or by construction equipment, motorcycles, leaf blowers, jackhammers, and garbage trucks, to name a few.

These low frequencies might not get measured accurately if they focus on A-weighted decibels alone. The good news: A-weighted decibels are not your only option. You can also measure C-weighted decibels (dBC). When you use C-weighted decibels, they include the lower frequencies that get cut off by A-weighted measurements. Readily available apps such as the NIOSH Sound Level Meter, Decibel Meter, and SPL Meter give you the ability to choose between A and C-weighted decibels.

Comparing decibel readings

Whether you use a decibel app on your smartphone or a hand-held device, the settings can also produce different results. For example, the BAFX 3370 decibel reader in the images below measures A-weighted decibels (dBA) in a range from 30-130 dBA. It has a sampling rate of 2x per second. Price-wise, it’s on the low end of the spectrum (~ $20 USD). This model doesn’t offer a lot of settings or features and it doesn’t offer C-weighted measurements. But it’s an easy, entry-level way to get a quick dBA reading at the press of a button.

NIOSH app (dBA)

In this first example, I placed the handheld BASX meter next to a smartphone running the free NIOSH Sound Level Meter. This is a complex app with a number of settings that enable you to calibrate it manually, set a threshold level (in this case, the threshold was set to “80”), use a frequency weighting of A, C, or Z, and choose a slow or fast time weighting. In this example, my NIOSH app was set to “slow.”

Curiously, I think the hand-held meter in this photo might be more accurate than the NIOSH app on the smartphone. Normal room noise in my home is typically around 35 decibels, so the 23.5 dBA reading in the NIOSH app seems rather low. A few tweaks to the default calibration or threshold settings might change the results.

NIOSH app (dBC)

In the second example, I’ve switched the NIOSH app to use C-weighted decibels with a “fast” response time. The reading varied a lot, and this app still showed a lower decibel level than the BASX meter, which registered at 38.2 dBA.

The NIOSH app also lets you run an analysis over time and generate a report, so I clicked the “run” icon on the lower left to gather more data. Here you can see that the instantaneous level was 31.1 dBC, the LAeq was 27.7 dB, the Lmax was 54.3 dB, and the LCpeak was 78.2 dB. So what does this mean? Let’s check this handy acoustic glossary from Gracey & Associates in the U.K.

LAeq is the A-weighted sound level (in this case, 27.2 dB)

Lmax indicates the maximum A-weighted sound level, which was 54.3 dB.

LCpeak is the peak of the C-weighted sound level, which in this case was 78.2 dB — higher than the Lmax.

SPL Meter (dBA)

In this third example, I’ve used the SPL Meter app to compare it to the handheld device. In this case, I measured some passing traffic outside the window.

The SPL Meter app also has a range setting (from 40 to 120), and a toggle switch for slow or fast response time. Like the NIOSH app, SPL Meter also has the option of using A or C-weighted decibels. In this first test, I set it to “A.” Then I set the response time to “slow” and set the range lower (about “40”).

As you can see, the average A-weighted decibel level shown on the SPL Meter (41.2 dB) is closer to the reading on the handheld device (40.6 dB). The SPL app also displays a maximum decibel level of 57.8 dB. This makes sense because the noise level varies when cars pass through the area.

SPL Meter (dBC)

In this fourth example, I changed some settings on the SPL Meter app to measure C-weighted decibels with a “fast” response time. Again, I measure some passing traffic.

Notice the gap starts to widen between the reading on the BASX meter and the dBC level on the SPL Meter app. The decibel level shows 43.7 dBC on the app compared to 40.3 dBA on the BASX, with a maximum reading of 66.0 dBC on the SPL Meter app. This is where you start to see more of a difference between using a C-weighted reading and an A-weighted reading for low-frequency noise.

Example: motorcycle noise

Let’s use the real-life example of a loud motorcycle and put the C-weighted decibel reader to work. In this example, a neighbor works on a very old and very loud motorcycle outside my building, and he tries to start it multiple times per day. The low-frequency noise penetrates through our closed, double-paned windows. I tried to record what it sounds like in this audio file. But in reality, it sounds louder and definitely more annoying.

So how loud was it? According to the NIOSH app (C-weighted), it was about 70 dbC with an LCpeak of 86.6 dB. This seems plausible to me, and accurately reflects the loudness that I perceive through closed windows. This is where C-weighted measurements come in very handy because A-weighted readings come in significantly lower (in the 50-60 dB range).

You might wonder why I didn’t include a comparison photo of the two devices side-by-side again. This was simply because I never knew when the motorcycle engine might start, and I didn’t have enough time to set up a photo.

In conclusion

Download some budget-friendly decibel apps to practice measuring noise. Your best options? Choose an app that offers at least two modes of measurement (A-weighted and C-weighted). Some apps also include the ability to save and log your data, too.

Use the A-weighted settings for airborne noise, conversation, distant noise, and environmental noise. Use the C-weighted settings to measure deep, low-frequency noises.

You can buy a physical noise meter online, and there are plenty to choose from. The lower-priced ones are easy to use, but they’re less adjustable and might offer fewer options. Getting into the mid-high range (over $200), you’ll find devices with multiple settings, the ability to record noise over time and log data to a smart card or to the internet. Getting into the high end, you’ll find instruments like the Brüel & Kjaer 2245, which has a very long battery life and can be connected via Bluetooth to your smart phone for efficient control.

Get familiar with your measuring tool of choice and log your results. With practice, you’ll start collecting reliable data.