Car Audio Measurement – Part Two

In this second part of our car audio measurement demo, we’ll be covering how to measure maximum SPL at the onset of audible distortion and detect impulsive distortion using advanced SoundCheck sequences. You’ll see how to identify buzz, squeak, and rattle with perceptual Rub & Buzz, transient distortion, and crest factor analysis.

Watch this Video Demo of Car Audio Measurement Techniques

Learn More about How to Perform Car Audio Measurements

Here are some additional resources:

• Standardized Automotive Audio Testing Seminar: https://www.listeninc.com/standardized-automotive-audio-testing-seminar/
• Automotive Audio Measurements Seminar (Frequency Response, Max SPL, Buzz, Squeak & Rattle): https://www.listeninc.com/automotive-audio-testing-seminar-2/
• Standardized Automotive Audio Measurements (presented at ISEAT conference): https://www.listeninc.com/standardized-automotive-audio-measurements/

Video Transcript: How to Measure Car Audio – Part Two

Today, I’m going to continue on from our last automotive video where we measured frequency response, spectral uniformity, THD and IM distortion. This time we’ll  demonstrate Max SPL at the onset of audible distortion  and impulsive distortion. Let’s make some measurements.

Again, I’m using the test configuration recommended in the AES TC-AA white paper on automotive measurements. We’re using a precisely positioned 6-mic array in the driver’s seat to represent our listener. This is connected to  an AmpConnect 621 Audio Interface. Our laptop is also connected to a measurement-grade Bluetooth interface  which we’re using to send the stimulus signal into our car’s head unit. If you want to know more about the test configuration, you can download the AES white paper, and we also have full details on our website. The inputs and outputs of the test system are already calibrated, as is the  audio interface, so we’re ready to measure.

Maximum SPL is one of the measurements outlined in the TC-AA proposal, and it’s important because it lets you compare maximum sound levels of infotainment systems. It’s defined as the maximum sound pressure level that the infotainment system can reproduce, and it combines all the components of the system – the speakers, amplifiers, power supply and so on.

The simplest way of evaluating this is to measure the overall Max SPL and Max SPL spectrum regardless of distortion. In these tests we set the car’s volume control to maximum and measure overall Max SPL and SPL Spectrum simultaneously using a 30 second broadband 20 to 20 KHz monophonic pink noise stimulus at minus 12db FS . This is quite a basic measurement and I’ve demonstrated it in other videos, so we’ll just briefly discuss the results here, then I’ll demo a more advanced measurement. 

Here we see the individual microphone measurements and the average, the black line. The overall level of 103dBC, regardless of distortion, isn’t too loud but maybe it might be limited by design to prevent severe distortion at low frequencies from clipping the amplifier. 

What I’ll show you now is a Maximum SPL measurement at the onset of audible distortion. More volume is only good as long as the output sounds clean, so it makes sense to define the Maximum SPL based on when the sound quality starts to degrade.

This is a more complex test sequence and takes longer to run because you need to repeatedly play the stimulus until you reach distortion across a range of frequencies – similar to how we measure Max SPL in other applications. The good news is that this process can be highly automated in SoundCheck.

Our test sequence plays the stimulus multiple times for each test frequency, increasing the level in 3dB steps until the THD exceeds a predefined level in percent. It then reduces the level by 3dB and increases it in smaller increments of 0.5 dB to more precisely find the THD threshold. The Max SPL is recorded at that frequency and the test sequence continues to the next frequency. The entire process is completely automated via the test sequence, making this measurement fast with minimal intervention.

This sequence actually takes about 20 minutes to run in total, so I’ve edited this recording to skip the repetitive bits. We’ll set our range from 50 Hz up to 500 Hz, and we’ll set the percentage distortion. I’m going to set that at 3%. That means that once the distortion level goes above 3% it will record it and move onto the next frequency.

We’ll start at minus 12dB. The first thing it does is an auto range and it measures the time delay in the system… Now it’s playing 50 Hertz…It’s auto ranging. In a moment, it’s going to show the first result.

All right, here we go.

So, we can see the 50 Hertz sine wave of the 6 microphones capturing it, and we can see that we’re around 3.4% distortion.

So now it’s going to lower by 3 dB.

And it’s going to start going up again.

So now we’re 114 dB at 2.4%.

Now it’s going in half DB steps to very precisely find the sound pressure level at exactly 3%.

I’m just going to skip quickly through the next few measurements until we start graphing results

OK. We got 50, 63, and 80 Hz. You can see on the bottom graph is the input level and on the top graph we have the maximum sound pressure level versus frequency, the blue curve, and the orange curve is the THD level for that sound pressure level.

Again, I’ll fast-forward through the rest until we get our end result – it’s just more of the same at a bunch of different frequencies

OK, here’s our final result. The top graph shows the Maximum SPL and corresponding THD, and the lower graph shows the maximum stimulus level plotted when the THD reaches 3% with the head units’ volume control set to max.

At some frequencies, especially higher frequencies, for example 400 Hz and 500 Hz, the maximum input level is reached before the system distortion reaches 3% THD. The stimulus level is typically limited by the lower frequencies where there is more distortion.

So that’s it for system performance measurements. Now let’s look at system integrity.

Impulsive Distortion is a useful system integrity metric. It detects any audible buzz, squeak and rattle that may be caused by the speakers vibrating the door panels, loose wire harnesses, etc. These noises degrade the listener experience so it’s important to assess the vehicle thoroughly. It’s actually a good idea to make these measurements before you make the system characterization measurements, as this ensures that the infotainment system is properly installed in the vehicle under test

You can also use this test as an end-of-line check to verify vehicle integrity.

There are several methods of measuring impulsive distortion. The AES White Paper suggests a crest factor measurement with a logarithmic sine sweep and a tracking high pass filter, to evaluate impulsive distortion at 80 and 90 dBA. Sine waves have a crest factor of 3dB, and transients, such as rattling wires, loose fasteners, and noise, for example background noise, typically have a crest factor greater than 8dB. Generally speaking, crest factors over 3 dB, may indicate buzz, squeak or rattle. However, you have to be a bit careful as background noise, especially in a noisy factory, can result in similar Crest Factor results. You have to be very careful to minimize background noise and make sure that the results correlate with what you can hear.

I prefer to use perceptual Rub & Buzz and Transient distortion versus time instead, as they are less sensitive to background noise, and only highlight distortion sounds that are audible to humans. Perceptual Rub & Buzz uses masking curves to simulate the performance of the human ear to measure only audible distortion.  Transient distortion versus time measurement tracks individual events in the time domain to reveal any loose particles or rattling wires. It offers greater precision than frequency domain measurements since it tracks individual transient events.

I’ll measure both these parameters simultaneously, using a stepped sine sweep stimulus.

There’s a couple of frequencies there where I could hear some buzzing or rattling sound, I’m not sure if you could hear it. Let’s see if we pick that up in the measurement.

OK,  so here’s our recorded time waveform of all 6 mics.

And here… we can see the perceptual rub and buzz and indeed there’s a little bit of a peak going on at around 224 Hz and probably at a higher frequency here to around 340 Hz. These are the little buzzing sounds that we heard during the sweep.

Now, if we go to the bottom waveform graph, we can actually see if there’s any transients. Those are typically caused by things rattling around.

We don’t have much here – at low frequencies we don’t have a lot of output, so I suspect this is just background noise. So this car isn’t bad – just a little bit of Rub & Buzz and no transient artifacts.

I actually tried cranking the SPL up to 90dB to try to get some distortion at higher volumes, and here’s what I got…

Let’s just take a listen to the recorded response…. At this volume you can hear the distortion a lot more clearly.

Here we have the results of crest factor, Perceptual Rub & buzz and transient distortion.

The distortion is so obvious at this volume that it even shows up on the Crest Factor response – you can see the Crest Factor is above 8dB between 200 – 300 Hz.

It’s also much more obvious on the PRB curve – you can see a peak in the same point showing it correlates well. This is a little more sensitive than crest factor though –  we can also see a smaller, sharper bump at 600 Hz which is not perceptible using crest factor measurement, even though I could hear it during the sweep. This measurement technique is also considerably more reliable in a noisy environment.

This 3rd graph shows the time transients of the impulsive distortion measurement. You can see the constant background noise as a steady noise floor and impulsive distortion is highlighted as well-defined transient spikes. You can see transient spikes from 2 of the microphones around 11-12 seconds into the sweep. These are closer to the driver side door where a door panel is rattling at low frequencies, around 200-300 Hz, in the sine sweep. A neat trick with this measurement method is that you can actually listen to the recording without the fundamental, let me just play that back now… and that really highlights the audible distortion.

So that pretty much concludes our suite of measurements. There are of course many variations of these measurements as well as other things you could measure with the same physical hardware setup, but that’s all we have time for today. There are some more in-depth automotive audio testing videos on our website and YouTube channel, so if you want to learn more, check them out.