Car audio is complex, but measuring it doesn’t have to be! SoundCheck measures a range of automotive infotainment parameters from simple frequency response to complex communications measurements. These include:
Frequency response. Frequency response is the clearest indicator of overall audio system performance. We recommend evaluating it according to TC-AA recommendations using 80dBA monophonic, correlated pink noise on both left and right input channels on the head unit. SoundCheck also supports the use of uncorrelated pink noise, which may yield additional useful information.
Seat-to-Seat Spectral Uniformity. Ideally, each passenger in the car experiences the same high-quality sound performance from the infotainment system. However, it is extremely difficult to independently equalize the loudspeaker playback system for each seat in the car. Trade-offs are made in the infotainment system tuning, typically favoring the driver’s seat. This results in the frequency response in the other seats being less smooth or ideal. SoundCheck measures the frequency response in each seat, then compares it to the driver’s seat to identify the seats with the greatest differences.
Total Harmonic Distortion (THD). THD is a valuable measurement which indicates the overall non-linear performance of the device; a measurement outside the limits may indicate a poorly centered voice coil, or mechanical defects. SoundCheck measures THD using both conventional and frequency-normalized methods. Frequency-normalized distortion offers greater insight to what is causing distortion for in-car measurements where there are many reflections.
Intermodulation Distortion. Intermodulation distortion is useful for evaluating multi-way speaker systems with a crossover. Intermodulation distortion can be observed when a small driver is trying to play back low and high frequencies simultaneously. Measurements are made using a two-tone stimulus with a fixed tone set to 50Hz and a sweeping tone from 20kHz to 150Hz, and responses are power-averaged prior to display.
Maximum SPL. Maximum SPL makes it easy to compare the maximum volume of an infotainment system in a car. It is the maximum sound pressure level that a car’s infotainment system can reproduce inside the cabin with its windows, sunroof, and convertible top closed, and it combines all the components of the system – speakers, amplifiers, power supply, etc. SoundCheck measures both the overall Maximum SPL and also the Max SPL spectrum, displaying them simultaneously.
Max SPL versus Frequency for a given percentage distortion. While Maximum SPL is the commonly specified measurement, it’s important that the audio system sounds good when it’s played loudly – this makes Maximum SPL at the onset of audible distortion an important metric. While time-consuming and manual with some test systems, SoundCheck’s conditional looping makes it easy to configure and automate. For each measured frequency, the tone-burst stimulus is played multiple times, increasing the level in 3dB steps until the THD exceeds, for example, 10%. The level is then reduced by 3dB and increased in smaller increments of 0.5 dB to more precisely find the 10% THD threshold. The Max SPL is recorded at that frequency and the test sequence continues to the next frequency. Although time consuming, the entire process is completely automated via the test sequence to minimize human interaction.
Impulsive Distortion. Buzz, Squeak, and Rattle (BSR) caused by vibrating door panels, loose wire harnesses, fasteners, etc. degrades the listener experience. Impulsive distortion can be used to measure this, and also serves as a useful measurement to make before performing system characterization measurements to ensure the infotainment system is correctly installed in the vehicle under test. SoundCheck offers three different impulsive distortion metrics: Crest Factor, enhanced Perceptual Rub & Buzz (ePRB), and enhanced Loose Particle detection (eLP). While Crest Factor is suggested by the AES Technical Committee on Automotive Audio (TC-AA), ePRB results correlate better to human audibility and provide increased reliability in noisy environments. SoundCheck’s eLP algorithm records individual transient distortion events versus time which reveals any loose particles or rattling wires in the car with greater precision.
https://www.listeninc.com/wp/media/2024/07/AES_automotive_in_car-scaled.jpg25601920Julius Wijonohttps://www.listeninc.com/wp/media/logo.pngJulius Wijono2026-02-18 17:16:562026-02-19 11:09:367 Key Automotive Audio Measurements You Can Make with SoundCheck
Recent transducer development, particularly piezo-MEMS loudspeakers, have brought renewed attention to Normalized Distortion (also known as Frequency Normalized Distortion), because it is essential to accurately characterize their behavior.
Normalized Distortion is a measurement algorithm first proposed in a 1993 AES paper by Listen president Steve Temme during his time at Brüel & Kjær. Although it has been available in SoundCheck since 1996 and used by some engineers for nearly 30 years, interest in the technique has surged recently. This renewed focus is driven both by the industry’s increasing emphasis on correlating measurements with audibility – an area Listen has long prioritized – and by the emergence of new microspeaker technologies that cannot be fully evaluated using conventional distortion measurements.
In this seminar, we’ll first explain distortion mechanisms and talk about correlating measured results to what we hear using masking curves. We’ll then examine how normalized distortion differs from the conventional approach and show why it’s essential for characterizing modern piezo-MEMS loudspeakers.
Learn how to average measurement data the right way in SoundCheck. In this video, we’ll walk through curve averaging, averaging multiple curves, and waveform RMS calculations to get meaningful results from your data.
Additional Resources for SoundCheck
Detailed explanations of how to use SoundCheck can be found in the SoundCheck manual.
Video Transcript: How to Average Data in SoundCheck
There are several ways we can average data in SoundCheck. The sequence step that you use depends on your input data and what you’re actually trying to get out of it. Let’s look at some examples.
First, let’s find the average of a single curve. We can do this using a post-processing curve average step. Here, my input is curve one. My operation is average. And when I output this value, it’ll give me the average value of all of the magnitude values from my curve as the Y axis, and also the average phase data as the Z axis on the value.
Now let’s find the average of a number of curves. We use this all the time, for example finding the power average of multiple microphones in an array. Here I have eight different curves and I want to find the average not of one but of all eight. Instead of a curve average step here, I’ll use an average curve step. First, I need to put all these curves into a group. I’ll just call it curves. And then in this post-processing step, I’ll choose the group that I’m interested in, which could be either curves or waveforms, and the averaging type. I have the option of complex, which will include phase data, or power, which won’t. I’ll use power for now. Rather than outputting a single value like we had before, this generates a curve that’s the average of all of my curves. And I can even put this on my display. And you’ll see that this curve here, the average curve, is now on my graph alongside the other curves.
We can also average waveform data. Here I have a one volt RMS sine tone. I might want to know what the average of this is here. If we use a curve average step, because this is a waveform and there’s positive and negative data, the average is going to be about zero, which isn’t very interesting to me. Instead, we could use the waveform RMS step. I can input my waveform. And this gives me an output of one volt RMS, which is as expected since that’s what I have set for my waveform here.
So that’s 3 easy ways we can find averages in SoundCheck. Check the manual for more details.
https://www.listeninc.com/wp/media/2026/01/Average-Data-Blog-TN.png10801080Julius Wijonohttps://www.listeninc.com/wp/media/logo.pngJulius Wijono2026-01-06 05:00:152026-08-27 13:59:38How to Average Data in SoundCheck
Understand when to use SoundCheck’s FFT Analyzer and RTA live versus in a sequence. In this video, we’ll show how live multi-instrument FFT/RTA measurements are ideal for real-time adjustments and troubleshooting, while sequence-based analysis steps enable automated, repeatable measurements, averaging over time, and advanced calculations like frequency response.
Additional Resources for SoundCheck
Detailed explanations of how to use SoundCheck can be found in the SoundCheck manual.
Video Transcript: SoundCheck’s FFT Analyzer and RTA: Live or in a Sequence
In SoundCheck, you can run an FFT or an RTA either live via the multi-instrument or through the analysis step in a sequence.
Let’s outline the differences.
Taking an FFT or an RTA via the multi-instrument is always an input-only measurement and it displays the live curves and changes over here in the graph. So, as I’m talking, you’ll see that these are displaying the current measurement on my reference microphone. This is very helpful for live adjustments – for example if I’m changing the DSP on a device and I want to see how it behaves, or checking the seal on some headphones. When you’re looking at it live, you can change settings on the fly, so using these controls I can see what happens if I switch from average to max hold, min, or change the windowing or weighting directly in the instrument. Most of this can be done in the analysis step as well, but you might have to rerun the acquisition or open an editor and click apply before you’re able to see the results, whereas here you can see the changes instantly.
The instrument is also very helpful for troubleshooting. If something is wrong with a measurement,you can just open up an FFT or an RTA to troubleshoot what’s happening and then solve it.
If we instead look at the FFT and RTA in a sequence, I can use an acquisition step set to record only. I can then run that reference microphone recording through an FFT spectrum analysis step and an RTA spectrum analysis step to get the FFT and RTA spectrum response for my reference mic. Instead of seeing live changes, I can define the full time length I’m recording and then average across that full recording. This is useful if you want to test something over a set number of seconds and just see the total average there without any additional user interaction. It’s much easier to automate.
There are some calculations we can do in the live multi-instrument. For example, I can look at the power average of multiple inputs and do some basic operations like subtraction and addition. One of the options we have here that’s exclusive to the live multi-instrument is a masking curve. I can play a signal here, add a masking curve, and I can then quickly see what parts of my spectrum are being masked by a dominant tone. That’s only available in the live multi-instrument.
Virtual instruments can be added to sequences in an acquisition step. When used this way, both the instrument and analysis steps output the curves to the memory list of the sequence once the step is complete. Post processing, limits, or statistics can be used for further calculations on the measurements.
There are two common use cases where an analysis step would be required rather than using the virtual instrument: Using previously recorded waveforms and measuring a frequency response.
For example, if I have wave file recordings from a past measurement, or an open-loop test, I can recall those and add them directly into an analysis step.
We also use an RTA analysis step to measure the frequency response of a device under test. The virtual instrument only calculates an RTA Spectrum of the input signal path, but the analysis is able to calculate spectra for both the Stimulus and Analysis waveforms. This then allows a frequency response calculation. The FFT analysis step will only measure a spectrum for the response waveform, but using the Transfer Function analysis step you can perform complex FFT on the stimulus and response waveform to output a frequency response.
https://www.listeninc.com/wp/media/2025/12/FFT-Analyzer-and-RTA-Blog-TN.png10801080Julius Wijonohttps://www.listeninc.com/wp/media/logo.pngJulius Wijono2025-12-23 05:00:172026-08-27 13:49:53SoundCheck’s FFT Analyzer and RTA: Live or in a Sequence
Normalized Distortion, or Frequency Normalized Distortion, is a distortion measurement algorithm first proposed by Listen’s president, Steve Temme (then at Bruel and Kjaer) in a 1993 AES paper. Although it’s been available in SoundCheck since 1996, it’s only recently gained traction. This is likely due to a renewed interest in correlating measurements to audibility – something that we’ve always considered important at Listen – and also because it is the only way of accurately measuring some new types of microspeakers. Let’s take a look at how it differs from conventional distortion measurement and why it is valuable for identifying the cause of distortion.
A typical speaker response, shown in Fig. 1, isn’t flat. There are typically amplitude irregularities due to reflections and resonances, and high and low-frequency roll-offs due to the passband nature of the transducer. This introduces some interesting artifacts when we look at harmonic distortion.
Fig. 1. Typical loudspeaker response
Fig. 2 shows the fundamental, second and third harmonic of a speaker. If we’re measuring harmonic distortion by sweeping from 20 hertz to 20 kHz, to measure the second harmonic, our filter tracks at twice the stimulus frequency and goes out to 40 kHz. Let’s assume that the second harmonic is 20 dB below the fundamental, i.e. 10% distortion. The third harmonic starts at three times the fundamental frequency and goes to 60 kHz, and let’s assume that this harmonic is 40 dB down below the fundamental, or 1% distortion. You can see the bumps and dips in the response align with the fundamental as the harmonics are filtered by the same linear frequency response.
Fig. 2. Loudspeaker response with second and third harmonics at their measured frequencies
The problem with conventional THD is that we always plot the harmonics at the stimulus frequency and then we compare the harmonic level to the stimulus level at the stimulus frequency as shown in Fig. 3.
Fig. 3. Loudspeaker response with second and third harmonics shifted to stimulus frequency
You can see that the responses are shifted along the x-axis to start at 20 Hz, like the stimulus frequency. This change in alignment between the harmonics and the stimulus causes the irregularities to be offset, and the high and low-pass roll-offs look entirely different.
Fig. 4. The effects of aligning harmonics with the stimulus before calculating THD
When we now calculate THD, (Fig. 4) we get some interesting results. Even though we showed earlier that the second harmonic was uniformly 20 dB down, when we calculate it this way, the distortion appears much higher at low frequencies because the fundamental rolls off sooner. Instead of being -20 dB (or 10%) at 20 Hz, it’s actually only -3 dB, which is equal to 71% distortion.
We see the opposite effect at high frequencies; the distortion rolls off prematurely compared to the stimulus frequency. This means it’s calculated to be -30 dB, or 3.1%, rather than -20 dB (10%).
In the pass-band region, the distortion could be under, or over-estimated, particularly around the resonances and reflections, where the calculation differences are more pronounced. This isn’t limited to the second harmonic; it’s the same for the 3rd and any subsequent harmonics so all these differences stack up.
If you look at a typical THD measurement of a loudspeaker, you’ll notice it’s always high at the lowest frequencies and rolls off at high frequencies. This is because conventional distortion measurement methods tend to over-estimate distortion at low frequencies, and under-estimate it at high frequencies.
So why does this matter? If you want to truly understand what’s causing your distortion, you need to separate the linear from the nonlinear. The good news is that this isn’t difficult – it’s just a small change in the order of operation when the THD is calculated.
What we proposed back in 1993, and I am still a strong advocate for today, is calculating it as follows: Instead of plotting the harmonics at the stimulus frequency and comparing it to the fundamental at the stimulus frequency, reverse the order. Plot the harmonics at the actual measured frequencies, compare it to the fundamental at the harmonic measured frequency, calculate the percent distortion, and then plot it at the stimulus frequency (Fig. 5). This is what we call Frequency Normalized Distortion.
Fig. 5. THD measurement when THD is calculated before aligning harmonics with the stimulus
If you’re wondering why this looks familiar, it’s because the graphs and distortion calculations exactly match the actual distortions shown in Fig 2.
As well as removing the influence of the linear frequency response on the calculated distortion, this method also minimizes the influence of room reflections. This enables you to measure the free field harmonic distortion in a regular untreated room with reflections – a particularly useful technique if you’re measuring somewhere like a car, which is full of reflections.
Now we’ve covered the theory, let’s examine some real-world examples: a conventional speaker and a piezo-MEMS speaker.
Fig. 6 shows the measured harmonics for a dynamic speaker driver. The fundamental starts at 50 Hz and the 2nd and 3rd harmonics are plotted at their measured frequencies, starting at 100 Hz and 150 Hz respectively. The roll-offs are clearly similar at the low frequencies, as expected. At frequencies above around 8 kHz, the shape of the third harmonic is particularly similar to the fundamental, clearly demonstrating its linear filtering effect. Even the second harmonic shows some similar characteristics.
Fig. 6. Measured harmonics for a dynamic speaker driver.
Fig. 7. Conventional and Frequency Normalized THD
Fig. 7 shows the THD plotted both the conventional way (purple line) and the normalized way (green line). The conventional plot indicates around 10% distortion in the lower frequencies, and a bump around 200 Hz – which could potentially be a problem. The reality, however, is that it’s just the linear response boosting the low frequencies. The conventional plot also has a little bump around 1 kHz, but this is caused by the slight bump in the fundamental curve at higher frequencies. If we now look at the frequency normalized curve, it’s clear that we don’t have a low-frequency issue, although we do have a little bump at resonance around 300 Hz – and that is indeed a nonlinearity. This is much more valuable information if you need to know precisely where the distortion is coming from so that you can design around it.
The differences between conventional and frequency normalized distortion are even more significant with piezo-MEMS transducers. Piezo-MEMS transducers are an interesting new class of miniature transducers that use semiconductor fabrication processes to create actuators with piezoelectric materials that drive a membrane, rather than traditional mechanical components. They behave very linearly in the lower frequencies – you typically don’t see any increase in distortion at all – but often have a little high-frequency resonance. Due to these characteristics, a major vendor of these devices advises their customers that Frequency Normalized Distortion measurements are a critical part of accurately assessing distortion.
Fig 8. shows the measured fundamental and harmonics of a piezo-MEMS speaker plotted at their measured frequency, and Fig. 9 shows the same harmonics plotted at the stimulus frequency.
Fig. 8. Harmonics plotted at measured frequency for a Piezo-MEMS speaker
Fig 9. Harmonics plotted at stimulus frequency for a Piezo-MEMS speaker
In Fig. 8, the shape of the third harmonic again clearly mirrors the fundamental, and the second harmonic has something going on – maybe a resonance – close to 10 kHz. In Fig. 9 where the harmonics are plotted at the stimulus frequency, the low-frequency distortion, as before, appears a lot more significant than in Fig. 8 where it is plotted at the measured frequency.
Fig. 10. Conventional and Frequency Normalized THD for a piezo-MEMS speaker
This is clear in Fig. 10, where we compare the conventional and Frequency Normalized THD. Plotted the conventional way, the distortion is calculated at around 20% at 100 Hz, which is quite high. However, when calculated the frequency normalized way, it’s down at 6-7% . It’s also quite flat throughout the 100 to 10 kHz range, which is typical of these drivers. They’re simpler – there are fewer components to cause distortion, and they’re not acting in a nonlinear magnetic field. As with the traditional driver, relying only on conventional THD measurements could lead to a lot of wasted time trying to design around problems that don’t exist.
Whenever I explain this distortion measurement method, most engineers immediately recognize its benefits, and wonder how the conventional method became the norm. I suspect the reason traces back to early measurement practices: in the 1980s, Bruel & Kjaer pioneered the tracking-filter method, and plotting harmonics at their actual measured frequencies simply wasn’t possible with the chart recorders that were in widespread use. That constraint shaped the convention we still use today, even though the technological limitations that created it no longer exist. Perhaps the demands of modern transducer technologies may finally drive the industry to adopt it.
Summary
Conventional and Frequency Normalized distortion measurement methods are compared in Fig. 11. Conventional THD isn’t wrong. If you listen to sine waves it is what you hear. However, we don’t generally listen to sine waves, and Frequency Normalized Distortion correlates better with listener perception of broadband signals such as speech and music.
More significantly, to correctly identify distortion mechanisms, and design products that address these, it’s essential to separate the linear response from the nonlinear distortion, and this requires Frequency Normalized Distortion.
Fig. 11. Comparison of conventional and Frequency Normalized distortion
Check out the recording below to learn more on Frequency Normalized Distortion: Why THD Isn’t Enough:
https://www.listeninc.com/wp/media/2025/12/Fig-2_LR.png437630Julius Wijonohttps://www.listeninc.com/wp/media/logo.pngJulius Wijono2025-12-12 15:16:362026-08-12 12:51:59The Value of Normalized Distortion
Measure time delay in microphone arrays with SoundCheck! In this video, we’ll show you how to calculate both absolute and relative record delays across multiple microphones for more accurate, time-aligned measurements.
Additional Resources for SoundCheck
Detailed explanations of how to use SoundCheck can be found in the SoundCheck manual.
Video Transcript: Measuring Time Delay in a Mic Array
When we make an audio measurement, there is always a time delay between the stimulus playback and the recording. We call this “record delay”. In a mic array, each microphone has its own record delay and sometimes you need to know the absolute delay between each mic and the stimulus, or the relative delay between each mic using one of them as a reference.
Here is a simple sequence to calculate the absolute delay.
I play a sweep through my speaker. There’s a play and record acquisition step which records simultaneously on three calibrated microphones in my array, mic one, mic two, and mic three, so I get three recorded time waveforms from this step.
Let’s run the sequence up to this break-point here…
Now you can see we have our three three recorded time waveforms in the memory list. I’ve grouped these and labelled them RTWs.
Next we use an analysis step to calculate the delay or the record delay between these microphones. Any of these analysis algorithms can be used as they all output the delay value for the waveforms.
To find the absolute delay for these microphones, we’ll use the stimulus itself as the stimulus waveform and the group of three RTWs as the response. When I click apply, I get record delay values for the three microphones.
Now I’ll remove this break point and run the entire sequence one more time. We hear the stimulus; it’s recorded simultaneously on the three microphones, and the display shows the fundamental response for each microphone along with the record delay values. These record delay values can be represented as time in seconds, or in inches or samples.
Now let’s calculate relative delay. We’ll use mic 1 as the reference mic, and calculate the relative delay of mic 2 and mic 3 compared to mic 1.
To do this, let’s use the recorded time waveform for mic 1, since it’s my reference, as the stimulus in the analysis step. And we’ll use other recorded time waveforms as the response. Let’s run the sequence.
Now you see that I get the record delay values compared to the first mic. Mic 1 is showing a delay of zero as that’s the new reference point. Mic two has a time delay of 90.7 microseconds, compared to mic one, and mic three has a record delay of 68 microseconds.
This concept of using a non-stimulus waveform as the reference in the analysis step is a valuable technique with many other applications in audio test, for example, dual channel impedance measurements.
https://www.listeninc.com/wp/media/2025/12/Measuring-Time-Delay-Blog-TN.png10801080Julius Wijonohttps://www.listeninc.com/wp/media/logo.pngJulius Wijono2025-12-09 12:18:182026-08-27 13:51:15Measuring Time Delay in a Mic Array
Easily calibrate your measurement chain by importing correction curves directly into SoundCheck. In this video, we’ll show you how to load external data, apply it to a signal path, invert response curves when needed, and ensure your corrections are properly applied during analysis for accurate test results.
Additional Resources for SoundCheck
Detailed explanations of how to use SoundCheck can be found in the SoundCheck manual.
Video Transcript: Importing Correction Curves for Calibration
It’s simple to calibrate signal paths in SoundCheck using either imported data or measurements you’ve made. This is useful, for example, if you need a frequency and phase correction curve for an input transducer, or a DRP to ERP correction.
To apply a calibration curve, the calibration data first needs to be in the memory list. If your data is in an external file, for example if you have a microphone frequency response curve from a calibration lab, you just select “open data” from the data menu of the memory list… and you can see now it shows up in the list. SoundCheck supports a wide range of formats from our own .dat format to txt and csv files.
Alternatively, you can use a recall step to bring your measured data into the memory list.
Now we have our correction curve loaded into the Memory List, we can use this curve to overwrite the calibrated device. Let’s open up the calibration window, and we’ll select the ‘copy from memory list’ button and choose our source curve. You can either apply the curve as it is, or invert it using this “invert curve” checkbox. You would use this, for example, if you want to remove the frequency response curve of the microphone from your measurement.
So, let’s check that box, and we’ll hit “apply” to save it, and OK.
Now we’ve added the appropriate correction curve to the signal path, the final step is to apply this new correction to our input response by enabling the checkbox “Apply Correction In” in the analysis step. It’s important to make sure this is selected otherwise, the correction will not be applied.
https://www.listeninc.com/wp/media/2025/11/Correction-Curves-Blog-TN.png10801080Julius Wijonohttps://www.listeninc.com/wp/media/logo.pngJulius Wijono2025-11-25 05:00:562026-08-27 14:03:06Importing Correction Curves for Calibration
Find the start and stop times of any recorded signal with SoundCheck’s intersection post-processing step. In this video, we’ll show you how to define a threshold, locate where your waveform crosses it, and calculate timing for applications like wake-word latency or response detection.
Additional Resources for SoundCheck
Detailed explanations of how to use SoundCheck can be found in the SoundCheck manual.
Video Transcript: Calculating Intersection Times with SoundCheck
In this video, I’ll show you how to find the start time of a recording.
Here I have a sequence which recalls a wave file. I’m using the ITU real speech wave file which is provided with SoundCheck as an example.
I’ll apply the recall step, and now we have the waveform in the memory list.
Let’s say I want to find the start time of this waveform, which is around 0.5 seconds.
We can do this with an intersection post-processing step, but first I need to define a signal level or a threshold.
We need to select a value that is above the noise floor but within the signal of the recording you’re using. Most of the signal in this waveform lies above 30 m FS, so I’ll use 30 mFS. This means SoundCheck will find the point where the waveform crosses the 30 mFs value.
Once this value is defined, I use an intersection post-processing step. I select the wave file or the recording as my operand A and the signal level that I have defined as operand B.
We’ll use the “search up” option, which starts at the beginning of operand A, which is the recording, and finds the point where it intersects with the signal level.
“Search down” would do it the other way, and “return all intersection points” will identify all points at which the waveform corresponds to the signal level.
Now let’s run the sequence, and we get the start time as a value, which is 492 milliseconds, approximately 0.5 seconds as we saw in the waveform graph.
This is a simple example of how to calculate the start time. This could be used, for example, for measuring the latency of a response to a wake word. If you wanted to find the stop time, you could use the “search down” function in the intersection step, and you could even calculate waveform duration by subtracting the start time from the stop time.
https://www.listeninc.com/wp/media/2025/11/Intersection-Times-Blog-TN.png10801080Julius Wijonohttps://www.listeninc.com/wp/media/logo.pngJulius Wijono2025-11-11 05:00:252026-08-27 14:04:49Calculating Intersection Times with SoundCheck
How do you measure ANC headphones? In this video, we show how to measure active noise cancellation (ANC) performance using SoundCheck. You’ll learn how to test unoccluded, passive, and active responses with a head and torso simulator (HATS), an AmpConnect 621 audio interface, a SoundConnect 2 power supply, and pink noise source to calculate total noise attenuation. This simple setup can also be expanded for real-time ANC analysis or used to measure THD and distortion caused by ANC circuits.
Today I’m going to show you how to measure active noise cancellation performance, or ANC, of headphones.
The set-up is simple. I’m using SoundCheck to generate the test signals and analyze the results. I have a head and torso simulator to capture the ear’s response. I’m using an AmpConnect 621 audio interface which also provides amplification for the speaker… and I have a SoundConnect 2 microphone power supply to power the mics in the head and torso simulator. If I was using a newer HATS with IEPE mics, I wouldn’t need this – I could power it directly from the AmpConnect. I’ve got a basic speaker here to generate the pink noise that we’ll use to measure the noise cancellation, and I have a nice pair of noise-canceling headphones.
We’ll play pink noise from this speaker. First we’ll measure the unoccluded ear response, then the passive response. That’s with the headphones on the ear but not turned on. Then we’ll turn on the ANC and measure the total attenuation – the full effect of the active and passive isolation.
Let’s run our test sequence. First we’ll take the unoccluded measurement without headphones. You’ll hear the pink noise playing, and we’ll see our measurement on the graph.
There we go…
Now the sequence pauses while we put the headphones on.
Next we’ll play the pink noise. I’m going to adjust the headphones to establish a good seal. Typically you would make this measurement 5 times and average it to account for variation in fit, but I’m just going to do it once in this demo.
Now let’s continue. [pink noise plays]
So now we have the occluded measurement – that’s telling us how much sound the headphones are passively attenuating just from their materials and geometry.
Now we’re going to turn on the ANC and let the sequence continue….
Here we have the measurement with ANC turned on.
And you can see on the top graph here, we show the sound pressure level for the unoccluded measurement, the occluded measurement, and the measurement with ANC turned on. And this bottom graph here uses a simple calculation to show the passive, active, and total noise attenuation of the headphone.
That’s the test in its simplest form. You can get more fancy with your speaker setup. For example, instead of using just one source speaker, you could use multiple equalized source speakers to create a more realistic spatial background noise environment. You could also measure both ears at the same time. If you do this, you can use SoundChecks multichannel RTA to check for fit – you just measure the two channels simultaneously and make sure the left and right are showing similar responses. This saves time as you don’t need to make multiple measurements.
You could also use the multichannel RTA to look at how noise attenuation changes in real time. You’d need an external reference measurement microphone positioned here, outside the headphone, and you could play complex signals, like real recorded background noise from an airplane and watch in real time how the noise attenuation changes. This gives a more realistic representation of how the ANC circuit responds to dynamic signals.
This set up is good for much more than just measuring active noise cancellations. We’ve done a lot of research on audio distortion caused by active noise cancellation. With the same setup, but different test sequences, you can measure THD, intermodulation distortion, time-variant distortion and more. Check out the technical papers on our website to learn more.
Detect loudspeaker air leaks automatically with SoundCheck’s advanced analysis tools! In this video, we’ll show you how to use THD, Rub & Buzz, ePRB, eLP, and Crest Factor measurements to identify air leaks and port noise – no human listening required.
Additional Resources for SoundCheck
Detailed explanations of how to use SoundCheck can be found in the SoundCheck manual.
Video Transcript: Detecting Air Leaks with SoundCheck
Traditionally, loudspeaker enclosure air leaks are detected by human listeners during the final test process. SoundCheck’s algorithms can detect air leaks and characterize port noise automatically. Let’s take a look.
I’ve made recordings of a closed box speaker system with and without an audible air leak using a 60 Hz sinewave. Let’s listen to them. First without the leak <play wav file> and then with the leak <play wav file>. Looking at these two recordings in real time on a spectrum analyzer, we can see the spectrum of the speaker with the air leak has significantly higher harmonic content than the one without, particularly the odd order harmonics.
Now, let’s take a look at the SoundCheck algorithms that can detect the air leak. For these demos, I’ll recall previously saved waveforms for analysis.
First, we’ll establish a baseline using a speaker without a leak. We’re going to look at THD, Rub & Buzz, enhanced Perceptual Rub & Buzz or ePRB, Enhanced Loose Particle detection or eLP and Crest Factor.
This display shows us these results, displayed on different graphs. As you would expect for a good speaker, THD, Rub & Buzz and Crest Factor are low, ePRB is below audible levels and there are no eLP events detected.
Now let’s analyze a recording where I’ve introduced an air leak by loosening the screws that secure the front baffle to the rear enclosure. THD and Rub & Buzz are significantly higher, ePRB is well above the audibility threshold of 12 phons and hundreds of eLP events have been detected.
These measurements were made with a microphone pointing at the side of the enclosure, the actual location of the airleak. If we next analyze the results of the same sweep captured by an on-axis microphone, we’ll see that there is a much smaller delta between the leak and no leak data.
Configuring pass/fail limits for any of these metrics is much easier using the side mic data so you can see that microphone location is critical for leak detection. If leaks are commonly coming from the front baffle, then a single, on-axis mic should detect them but if the leaks are coming from the back or sides of the enclosure, two or more microphones should be deployed. The good news is if you are doing this on a production line, although you need an extra microphone, the measurement can be made simultaneously with your other QC measurements.
It should be noted that this method may also pick up defects from other causes, but if you’re using it to reject faulty speakers, then that doesn’t really matter.
These techniques can also be used in R&D for the characterization of port noise, allowing you to compare the results using ports of different shapes and sizes.
Some of these methods may work better than others on your specific design so I’d recommend that you evaluate all of them with your own products, and focus on those that yield the most reliable results.
https://www.listeninc.com/wp/media/2025/10/Air-Leak-Blog-TN-1.png10801080Julius Wijonohttps://www.listeninc.com/wp/media/logo.pngJulius Wijono2025-10-28 05:00:552026-08-27 13:47:51Detecting Air Leaks with SoundCheck