Automotive Inc.

There are many different aspects of automotive testing ranging from the measurements themselves, how you connect with the infotainment system, and more complex testing needs such as voice assistance, readiness, and pre-qualification. Firstly, I'm going to talk about the measurements themselves, things And then we'll make some in-car measurements so that you can see how easy it is. Next we'll look at various connection methods that infotainment engineers have to contend with And finally we'll take a look at some of the other audio measurements that you can make using soundcheck. For example, active noise cancellation and voice assistance readiness. Let'start with the measurements. I've been part of a working group on in-car acoustic measurements working under the AES technical committee on automotive audio. Over the past couple of years we've been working on trying to develop recommendations for some of the important audio measurements that need to be made in cars. Specifically, frequency response, max SPL, and impulsive distortion, w I'll give you a quick explanation of what those measurements are, then show you how we can make those measurements in soundcheck. Frequency response is a common audio industry measurement. We use it all the time for speakers, headphones, any device that plays sound. It's a quantitative measure of the sound pressure level as a function of frequency, and it gives a clear visual indication of the spectral balance in the car. One of the t We want to make sure that there are no sudden peaks or dips in the curve that would indicate that certain frequencies are being emphasized or attenuated with respect to the rest of the spectr Max SPL is important because cons The maxim closed. There are two methods that are recommended, overall max SPL and max SPL spectr It is very difficult to characterize distortion in a modern-day infotainment system because all the signal processing, w Impulsive distortion measurements give us a way of quantifying unwanted distortion artifacts, such as buzzing panels, rattling cables, loose particles, and air leak noise in a car. These are important to identify as they are easily discernible by the passengers, giving the perception of poor quality. These may be either parts defects, for example, the speaker itself, the enclosure, or the subassembly. Or it may be introduced by the mounting, for example, plastic clips, or even the panel themselves might be excited by So next, I'm going to demonstrate how we make these measurements in a real car, using soundcheck, of course. We'll make all three of the measurements I mentioned earlier. Fortunately for us, they all use the same test setup, w So let me talk you through the setup for these tests. Now, the physical setup that I'm using is based on the recommendations from the AES Automotive Audio Technical Committee that I mentioned earlier. Obviously, if you want to compare measurements between cars, it's critical to use the same measurement conditions. So there is value in the whole industry standardizing. If you're just doing it for your own in-house evaluation, it's also important that you are consistent, but you can, of course, determine your own setup. All right, let's make some measurements. We use a 6-microphone array to represent the driver's head in the driver'seat. T Let's get t First, I'm going to put the mics into the array. I'm using our own SEM measurement mics, the short version, because they are nice and compact, but you could use any quarter inch measurement microphone. Now I'm going to attach t Now I'll connect all the cables. Now let's look at the positioning in the car. I am positioning the tripod legs as far back as I can get them on the seat right against the backrest. The height and angle of the seat are set to the minim Now the AES is working on defining some very specific dimensions between the pedal and backrest and the height of the stand. 105 centimeters from the brake pedal to the backrest in a straight line and a microphone array height of 72 centimeters from the seat to the second and fifth microphones. That's the two middle ones. It also specifies the incline of the backrest so it is 10 centimeters from the t For the purpose of the front passenger seat should be in about the same position as the driver'seat. There we go. That should do it. Now let's connect our audio interface. I'm going to use our amp connect 6 to 1 w You can use any audio interface that works with your microphones and offers the resolution and other characteristics needed for reliable and repeatable audio measurements. T It's also Teds compatible so if you happen to have Teds microphones, you can save yourself some setup and calibration time. Now let's plug the microphones in and I'm going to connect t Now as we'll demonstrate later, there are many ways that your audio measurement system can be connected to the car infotainment system. Soundcheck can connect to the car using a USB drive, the car's Bluetooth or even with the A2B bus. Here I'm going to use Bluetooth because it is fast, convenient, I don't have to get in and out of the car and most importantly I don't have to worry about damaging my hearing when measuring max SPL. I've got a BTC 4149 Bluetooth interface. T I have t I'm going to add it here under my Bluetooth devices, add device and start search. You can see it shows up as a Bluetooth device so I can add it and we're all set to go. Now the last part of our setup is checking that the head unit's audio mixer settings are in the default position for standard and repeatable configuration. So we go to the settings panel. First we're going to make sure we have flat tone controls so make sure all of these are set to the middle position. We want to make sure the balance is in the middle w That's the default setting. We also want to be able to use different sources for example Bluetooth, USB memory stick and so forth. Although I'll be doing t Now the important particularly for the max SPL test we need to crank the vol I didn't realize until I came to demonstrate it that the sound on t So now we've done that we'ready to go. Now let's get testing. For the actual measurements I'm going to use our soundcheck test system. Over the past couple of years we've introduced several new features that are really useful for automotive testing. For example the multi-channel RTA. Anyway let's j Frequency response, maxim We'll start with frequency response. We're going to measure then I'm going to run max SPL that's what the vol I'm going to show this as a spectr Lastly I will show impulsive distortion where we want to look for buzz, squeak and rattle. Okay let'start with the frequency response. The Bluetooth interface is paired to the head unit and the pink noise is played out of the loudspeakers. The reference excitation signal is a monophonic full-scale pink noise signal with an RMS level of minus 12 dB FS as defined in the AES-17 standard. I'm just going to You'll see the RTA come up and you can hear pink noise being played. You can see the output of the 6-micro-a with each individual microphone plus the black line w You can see on the multimeter that it's at 80 dba. T And here you can see the frequency response curve. It is pretty smooth with a significant amount of bass and some treble boost. Now let's measure max SPL. We'll use the same broadband monophonic pink noise that we used for the frequency response but this time we'll crank the vol Yep that's loud. If you do t It's 5 dB and we're going to play it for 30 seconds and take the linear average for LECU. Here you can see the reading on the six microphones and the black line representing the spatial power average. And now you can see that it calculates the overall max SPL as 111 dBc weighted. Also we have the max SPL spectr Lastly I'm going to make an impulsive distortion measurement. Here we're looking for buzz, squeak and rattle due to interior vibrations caused by the loud speakers mounted in different locations in the car. We're using a Okay here's the results. What we're looking for is some spikes in the loose particle envelope curve on the bottom right and some b You can see no b Now I'm going to introduce a piece of plastic in the left door pocket to represent a loose component that will vibrate. Hopefully we'll get to hear and see the difference when t Let's make that measurement again but at a You can see there is some rattling going on in the loose particle envelope and some Rubin-Buzz going on just below 200 Hertz near resonance. We have a reading of about 45 fonts on the perceptual Rubin-Buzz measurement. Any PRB above 20 fonts is quite audible. So you can see it's one simple setup and you can make all these measurements very quickly. T So the next t Cars are complex systems and most offer several ways of connecting audio to the head unit to play test signals out of the infotainment system. I just did all my demos using Bluetooth because t You can also put test signals on a memory stick and play them via the head unit or even access the car's infotainment system via the A2B bus. One of the neat t We're now going to review these three common connectivity methods Bluetooth, USB drive and A2B. We'll demonstrate how they work and discuss the pros and cons of each. Firstly, Bluetooth. Let me first address one of the questions I am often asked is Bluetooth accurate enough? The answer to that is yes. Absolutely. Provided that you are using a T Now I'm going to hand over to my colleague Steve Taterunis who will explain more about testing via Bluetooth and demonstrate some measurements. Thanks Steve. Most commonly cons Bluetooth connectivity gives us a standardized method to connect to the CARS audio system to both send and receive audio test signals simplifying what might otherwise be a complex hardware and wiring setup. Earlier you saw Steve Temme perform a variety of in ve The Bluetooth signals were transmitted using the BTC 4149 Bluetooth interface from our measurement partner Portland Tool and Die. Un 0 compliant RF interface. It supports a comprehensive suite of codecs protocols and other significant settings and can be configured through touch screen controls, a command line interface or message steps in a soundcheck sequence. These options provide unparalleled control over protocol selection, pairing options and transmission power. I'm often asked if using Bluetooth is accurate enough for automotive testing. There's one t You should not use an unqualified Bluetooth dongle. Only a laboratory grade instr Configuring the BTC 4149 is simple. When soundcheck detects the presence of a BTC, it will automatically configure soundcheck's hardware and calibration with the appropriate digital signal channels and signal paths. Once t Let's take a look at how we can use message steps in a soundcheck sequence to configure our BTC 4149 to establish a Bluetooth connection and send a test signal to the audio system. Since we're in the studio today and don't have a car available, I'll be connecting to t Okay, so now it's time for a quick demo and soundcheck. And what I hope to show you here today is how we're going to be using message steps to configure the BTC and discover available devices and open up the audio channel between the BTC and the device so we can continue on with our acquisition and analysis in soundcheck. So as you can see here in the sequence editor, I have three message steps called BTC Setup, BTC Get Address, and BTC Connect Audio. And I'll be using those to send commands to the BTC w So let's quickly open these up and take a look at the functions of these steps. So BTC Setup is for the codec selection and it allows us to select the A2DP codec. In t I'm not going to be using HFP in the next step, BTC Get Address, is where the BTC will go out and look for discoverable devices. So t And there are connection options here where we can connect by name, connect by address, or disconnect. And the other user configurable parameter here is the inquiry time or the discovery time. How many seconds do you want to spend searc Pretty good value here is 10 seconds. And then lastly, we have the actual connection step where we open up the audio stream between the BTC and the connected device. What is most important here is the audio channel selection, whether or not we're going to connect over A2DP or HFP. And in t So otherwise, the sequence is pretty straightforward soundcheck sequence. We're going to be using virtual instr We'll be using signal generators and RTAs. And then a display step and a couple of operator dialogues to start and finish the sequence that give the operator some options about connectivity. So let's run the sequence. So the first step in the sequence is a message step. And it asks the operator if they're currently paired with the device under test, w So we're going to say no, and we'll go through the pairing routine. We're searc And there's a couple that have showed up here as available. And I know from experience that I actually want t So once I click OK, we'll proceed to the next message step where we're going to connect to the audio and the sequence will progress from there. And so here's our final display. And there's one last message step asking me if I want to disconnect from the dot. And I have the option of keeping the connection open in case I want to make another measurement here. Or if I'm done, I can disconnect. So I'm going to say yes. And we're left with our final display. And what we're actually looking at here is the response of the car kit sending the signal over to an FM tuner. And then we're sending the tuner's response back into soundcheck. So that's why we don't have a very flat response here. We'really seeing the characteristics of the tuner here in the final response. So there's a quick demo of how to use message steps in a soundcheck sequence to connect to your Bluetooth device, send audio and disconnect. Using soundcheck in a device Back to you, Steve. Next, I'm going to cover USB drives. T However, it's not without its drawbacks. You need to record the stimulus with a long period of silence at the beginning to enable you to plug it in, It's not ideal, but if you have no other way or only need to make measurements infrequently, it works. The reason that t If you t In soundcheck, we offer several different techniques for triggering a recording and automatically detecting the time delay in the measurement. And if you want to learn more, there's plenty of information about open loop measurements on our website. The other situation we encounter with automotive testing is the A2B bus. T Most modern cars contain microphone arrays for everyt I mentioned earlier that soundcheck is great at getting signals into and out of any device. And t Soundcheck can communicate directly with the A2B bus to enable you to make all the same measurements that you would with a conventional wired setup. I'm now going to hand over to Mark Lachaw, who will demonstrate how we can measure MEMS microphones via an A2B interface. Hey, my name is Mark, and I'll be demonstrating an A2B application for automotive testing. A2B is a2B is capable of transporting data, control information, clock, and power using a single two-wire uns In an automotive application, there are some major benefits of using A2B. Multiple devices can be daisy-chained together. A2B also supplies power to all devices in the chain, so local power supplies aren't needed. Since only a single twisted pair of cable is required, the total weight and complexity of the system can be much lower. In contrast, traditional wiring would require multiple connections, power supplies, with lighter ve In t I'm then connected to two boards over here, each of them with two PDM MEMS microphones. So in total, I have four MEMS microphone inputs, two line inputs, and two line outputs. We're also connected over here via USB to my computer. Using analog device's program Sigma Studio, I can set up the configuration devices. I'm also using an audio connect and an SC amp connected to a speaker over here so I can play out pink noise to test these microphones. Now we'll take a look at Sigma Studio by analog devices. So in t So I have my sound card configured, it's connected to the AD2425, and then I can select w So again, on t If your hardware is different, you'll need to set it in Sigma Studio to match what you're using. Now that it's configured, we can set it to the device, make sure that it matches the settings, and what I've done in t So we'll take a look now in soundcheck. If we go into setup hardware, we'll be able to see that I now have six ASIO inputs into ASIO outputs, w I also have my two inputs and outputs for my audio connect. In setup calibration, we'll open up the table, and you can see for the input side, I've created four input signal paths that are all tied to the microphone. So inputs one through four will correlate to the MEMS mics one through four on these boards over here. And on the output side, my source speaker is configured to play out of my audio connect out of the speaker here. So we'll save t Great. So as you can see there, we had an RTA with four discrete input channels. Were reading the four microphones on my setup over there, and were able to put these into memory, set limits on them, and test them whatever we want further in a soundtrack sequence. Once the data is captured via A2B, we can test it the same as if it was from any other source. For example, if these microphones are going to be used for noise cancellation, we could then write a sequence for that specific use case. So as we come to the end of the seminar, I just want to mention some of the other t We focused here on basic in-car measurements, but soundcheck is extremely powerful and flexible, and can actually be used for virtually any audio measurement inside or outside the car. For example, wit It can also carry out a wide range of tests on in-car voice assistance, such as Google Art and Apple CarPlay, as well as proprietary voice assistance and hands-free communication systems. Outside the car, it can be used for laboratory measurements of virtually any component, speakers, microphones, communications, etc. , at the design stage. In fact, we even have many specific options for communication measurements, such as Polka, DoubleTalk, and other industry standards. Thank you.