Using Masking Curves for Measurement Insights

Discover how masking curves in SoundCheck 23 bring psychoacoustics into your measurements with real-time tonal masking and threshold of hearing visualization – an industry first. In this video, we’ll show you how to use this new feature for production failure analysis and R&D insights by correlating what you see with what you hear.

Additional Resources for SoundCheck

Detailed explanations of how to use SoundCheck can be found in the SoundCheck manual.

Video Transcript: Using Masking Curves for Measurement Insights

A great new feature in SoundCheck 23 is the introduction of real-time tonal masking and threshold of hearing curves in the spectrum analyzer—an industry first. This lets you visualize how a dominant tone audibly masks other frequencies for a visual correlation between what we measure and what we hear. Let’s take a look at how to use it.

In production, it can give a deeper insight into failure analysis. Here, I have a small micro speaker. These designs are notorious for having rub and buzz problems around resonance due to their single-point suspensions. Let’s run our test sequence and examine the results.

The upper XY graph shows the fundamental and the enhanced perceptual Rub & Buzz (EPRB) curve. There’s a strong EPRB component focused right around 600 Hz. In the lower XY graph, we’re plotting the EPRB curve along with the electrical impedance curve, and we can see a strong correlation between the defect and the resonant frequency of the loudspeaker. On the waveform graph, zoomed in around 600 Hz, we can clearly see how distorted the waveform is.

Now, let’s use the masking curve tool to investigate this defect further. I’ll open up a saved virtual instrument configuration that contains a signal generator and the FFT analyzer – the two tools you need to use masking curves.

The test level in our sequence was 200 milliVolts, but I’ll start the signal generator at 70 milliVolts and 600 Hz. At this level, a couple of harmonics are visible. Now let’s add our masking curve. We’ll add a calculation and select “Masking Curve” from the dropdown list. Our calculation is now a masking curve. Its name is calculation one and our operand is the FFT1 live curve. Now on our graph we can see our fundamental tone, our masking curve, and we can see that harmonics 2 and 3 are well below the masking curve, suggesting they should be inaudible.

But now, watch what happens as we increase the test level. Around 150 mV is when perceptual Rub & Buzz really takes off. Now nearly all harmonics above the fifth up to the 20th or so are above the masking curve. Now when we bring this up to our final test level of 200 mV, nearly every harmonic up to the 30th is touching or above the masking curve, so we’d expect this to sound pretty bad. Let’s take a quick listen… and as we would expect, we can hear a lot of distortion.  

Finally, I’m going to remove the cursor and increase the frequency up to the point where the measured EPRB in our sequence should drop down. It’s doing some averaging, lets wait for it to settle. Now we can see at 800 hertz things have calmed down quite a bit. There’s still some visible harmonics but they’re pretty much at or below the masking curve. Let’s take it up to a kilohertz, now all our harmonics are below the masking curve and we should have a nice pure tone when we listen to it. 

So there’s a look at how one might use this new feature for failure analysis on a production line. It also has R&D applications, such as monitoring the effects of component size and quality during product development—for example, the contributions of the voice coil and magnet system to harmonic distortion. To learn more, check out the SoundCheck manual.