Why your measurements don't match. Mic Array vs Single Mic RTA testing.

0:06 Hey everyone, this is Jeffrey Hall from Total Tuning. Before we get started, I want to be very clear about the purpose of this video. This is not about proving that one measurement method is superior and another is wrong. This is about understanding why different measurement techniques produce different results and how those differences affect consistency, repeatability, and decision making during a tune. If you've ever wondered why two measurements taken in the same car can look completely different, or why a tune that looks perfect on screen doesn't always translate the way you expect in the seat, this video is for you. What I'm going to show you is real data

0:51 taken in real time using multiple common measurement approaches. The goal isn't perfection. The goal is consistency, accuracy, and getting to our desired result quickly. Now, let's get into it. So, on screen is the live recording of my screen as I was doing all of the testing. So, I just want to commentate and highlight any key points there as we go through everything. Just pressing play. This is a uh speedun through a quick tune on my car using a generic curve using a generic process. This is not the normal total tuning process or curve. This is just to show that the

1:38 speakers are following the targets on screen. Uh time aligned and phase aligned accordingly. um just so that I can get the most accurate testing as I go through each individual microphone. So next I'm going to uh connect the multipplexing again and I'm going to show the summed response of the left mid and the left tweeter in this particular instance. Going to go ahead and take that capture. And what I'm trying to point out here is that the left mid-range and left tweeter are both following a target. They're both uh in time and they're both in polarity. So, they're going to sum very nicely when I uh play them. So, when I capture that tweeter and I capture that mid and I capture the sum between them, you can

2:24 see that there's a really nice bridge in between the two. Um just showing that, you know, the data is working according to our process. Next, I'm going to break apart the five microphones from the multipplexing. I'm going to bring them all on screen and show you that, you know, each microphone is contributing a very different RTA response. So, if you're unfamiliar with the Jail Audio Max microphones, there's five of them. And in this particular instance, they're all different colors on screen. Uh what I'm doing is I'm capturing each individual trace uh on the tweeters. And I'll be doing the same thing on the mid-range here in just a moment. The point of all of this is to show how we had that really nice summed

3:10 response where everything was nice and smooth, but now each individual microphone is showing a very different response. And something to keep in mind is that these microphones are, you know, a little bit wider than your normal headspace. So, you wouldn't expect these crazy deviations from, you know, being that close together, but there's a lot going on in a car. And that's what I want to point out here. So in this particular instance, I'm saving all five traces on the mid-range next. And what I want to highlight here is that there's some pretty big differences between uh two of those responses in particular. Uh in this particular case, I'm going to go ahead and get rid of all of this data

3:56 just to try and clean it up for you. And I'm going to show the uh multipplexed version of the mid. So the red trace on the screen uh is going to be your average response between all five microphones multipplexed on JL Audio Max. The purple one on screen there is actually your far left top microphone uh slightly above your ears. And now I'm using the set A set B function on JL Max to show that there's a pretty big difference right here at uh 944 hertz. There's over a 3dB difference between the average and what one microphone happened to contribute. So it's a pretty big deal, especially if you're, you

4:42 know, making half dB or 1 dB increments on your EQ adjustments. I mean 3dB is over twice the power that you're applying to the speaker and you know that can result in a very different sounding system. Um this is another super impactful measurement in this particular case. This is the uh left top microphone microphone one. The red is the average and the brown is actually the center microphone where people would normally put a static microphone. We can see that's over a 9db difference at 1.2k. So again, this is that brown trace is where you would typically put a microphone. If you were just going to shove it in the middle of the headrest, start tuning away, you would see a pretty crazy response. Something else to

5:30 point out is that, you know, this car has done very well on the competition scene. These are highly optimized uh speaker locations with reflection control and really trying to get the best possible scenario. and I'm getting these crazy of uh deviations between all five microphones. So, just goes to show that in a less optimized install such as a factory locations, you know, speaker replacement kind of scenario, what most people would fall into, you're going to see even crazier responses than this. So, I'm going to go ahead and uh scoot through to my next sections here. Um, and in this particular instance, what I'm doing is I'm taking that center microphone. I'm trying my best to tune

6:16 it to a center trace as if I was putting the center microphone with no body in the seat. Uh, right now I'm saving that on screen. This is the uh, single red center microphone. And next, what I'm going to do is I'm actually going to turn back on the multipplexing. So this is all five microphones against what the center microphone showed us using the center microphone's EQing. So again, in my opinion, the the uh multipplexing array shows us more accurately what we might hear. And we can see that the mult the multipplexed array we have a a boost between that 12 and 1400 range. We have another boost in the uh upper 2K range and it just doesn't quite match what the red microphone showed us. We can uh see

7:04 that the red microphone showed us some dips uh down there in that lower mid-range and actually some boosts in the you know 7 to 900 Hz range. There's a couple uh boosts right there that the microphone array when multipplexed showed us that that actually wasn't you know as accurate to what we might hear because there's other data happening in other parts of that listening environment. Now what I want to show is what each microphone is contributing towards that multiplex. So again, you know, there's five microphones in this particular situation. This happens to be um my labeling of the first microphone and I'm just going to capture all five of those. And uh as we get to see here, we'll see that the the red one matches

7:50 perfectly because that is the microphone that we tuned this particular setup on uh during these particular sets of traces. Now, we can see there's a couple uh commonalities here. The yellow and the gray on screen have that peak uh and they just change to brown and red because they're save traces. They have that peak in that mid 2K range. Uh microphone one is that uh kind of lighter blue. It's got that pretty large peak uh right there in the 1000 to700 range. And what I did was I just reduced the banding so it was a little easier to see. The yellow highlighted trace on screen there is the multiplex again. So just kind of showing what all five microphones average out to be. And then um this is me highlighting

8:37 the um the individual microphones and showing how big of a difference between any particular set of microphones. So we can see in this particular set over 8 1/2 dB difference at 1.2k. Uh again using that banding. So, um, if I were to make it a, uh, less aggressive banding such as like 148th, uh, which is what I'm doing right there, we can see that that that dip gets even worse. The the deviations get pretty, uh, pretty wild there. Um, the next thing that I'm doing, uh, which is actually on screen, the yellow trace is that again, that center microphone with no body in the seat. the red microphone trace, which

9:23 I'm saving, and it'll change color here in a moment. That was me sitting in the seat holding the microphone in front of me. And what I want to point out is that now it's gray. So again, the red is a live trace. So let's ignore that for now. We can see that the there's a dip at 1K or 1.2K on the nobody trace. We can see that gray trace looks like my torso, my body moved the cancellation down to like 7 or 800 hertz. So, pretty significant there. You know, understanding that the target changed or the target would need to change based on if you were sitting in the seat or not sitting in the seat and understanding what the measurements are showing you on that that screen there. And next, what I'm taking is I'm actually taking the

10:10 average or a moving mic average of a single microphone. So again, body in the seat. You can see that slight dip that I've kind of averaged out by moving the microphone around my head. And that dip doesn't look quite as pronounced now, right around that 7 to 800 Hz range. So it shows you that the moving mic method is a a valid way to kind of clean up some of that data and make it easier to digest, make it easier to make good decisions. But we can still see that there's going to be some some differences here. Once I uh get done saving this trace, we can take a look more in detail. The green trace is what I just saved, and that happens to be microphone facing up as I'm moving it. And we can see there's there's a a boost right around that uh 4 to 500 htz range.

10:56 You know, it's not exactly accurate to what the uh multiplex is, but it's a lot closer. It's a it's a it's really a lot closer than the previous method of just putting the microphone in the seat. Now, in this particular instance, so we can see that facing forward, I've got more of that dip in the lower mid-range, and I've got more of a boost in that upper mid-range, that 1 to 2k region. So, again, I'm just, you know, what I did was I just reset the data, and I'm taking a new capture. So, I'm moving it back and forth. And we can see that just changing the microphone position, orientation, I should say, changed the end result. And you know, that's where calibration files come in on your microphones and understanding how they're going to read in one scenario or

11:43 another. Uh, knowing what might work better for you, what your target curve compensates for, things like that. Just capturing some of that data. The the last thing that I'm going to do here is I'm actually going to get out of the seat and repeat those same steps. So again, microphone facing up going back and forth. Microphone facing forward going back and forth. And then uh as a bonus thing, I'm going to do microphone in a a figure eight. What I'm doing right now, I'll overlay it on screen. I'm actually in the back seat reaching to the front seat and moving the microphone uh back and forth. And then I'll change it to the uh facing up position here. And then uh again, the last one here is going to be my figure

12:29 eight. And this is the most accurate to the array that I saw from the single mic measurements. And just making that figure eight gesture, getting, you know, double or even triple the data that you do from in front of your face to the sides of your head. I think that that greatly contributed to a much nicer response. So, right now I'm highlighting the um figure 8 versus the uh multiplex average. And I'm going to pause it right there because that's that's really the most important part of this whole uh commentary is that your multiplex average is that lighter turquoise kind of color and then your figure 8 is uh that lighter gray uh kind of color. on

13:15 screen. We can see that they they meet target or they meet each other pretty well, pretty accurately on the two crossover slopes, but the multiplex of me being out of the seat. So, this is five microphones, no body in the seat, and then they're being averaged in in in a multiplex kind of one at a time way, uh the way that J Audio Max does it. And then the gray data set is me from the back seat doing a figure eight with a single microphone taking an average. And it was our best results, but it's still not the same. There's still a a pretty big difference around 1K. Um, and then again, uh, you

14:03 know, a a dip around that 1500 Hz range. Both very critical regions in my opinion. uh you're going to get a lot of um you know upper vocals, guitars, uh snare drums, a lot of important things around that region. So, I just wanted to bring this up in regards to knowing what your measurements are going to do for you and knowing how to interpret those measurements so that you can make the best tuning decisions possible. Each microphone shows a dramatically different response. This is not a measurement error. This is physics. Think of it like watching leaves on a tree to understand the wind. If you look closely, you can see direction, beaming, and interference patterns in real time.

14:51 What you're seeing here are localized effects. Comb filters from reflections, off-axis behavior, and very small positional changes that radically alter frequency response. A single microphone is not wrong. A properly functioning microphone does not lie. But microphones don't interpret data. People do. And that's why understanding your measurements is a huge part of tuning. The most common mistake is assuming that a single static microphone represents what we hear. In reality, every static position will show dips caused by comb filtering at that exact location. If you aggressively EQ these dips, you're not fixing the

15:38 system. You're fixing the microphone's response on screen in that exact spot. Every sound you've ever heard, every venue, every concert, every car has been heard with your body in place. Your body is always a part of the acoustic system from your point of view. That doesn't mean we must tune with a body present. It means we must understand what changes when we tune with a body in the listening environment. If you move the microphone to create an average, results improve. But now consistency becomes the challenge because the nature of who's taking the measurement. How fast are you moving? How wide is the

16:25 motion? Is your body in the way? Two people waving a microphone will not produce the same data. Even the same person won't produce it perfectly twice. And no two people have the same size, body, shape, or size, which means the data changes again, and the target curve now has to be adjusted. If you only have one microphone and want better repeatability, measuring from the back seat while tuning the front can greatly reduce the body's interference. But this is still a compromise. Even after compensating for all of this correctly, you'll notice that this technique takes more time than an array. A microphone array solves all three

17:11 problems. It averages space, helping to show more accurate representations of what we hear. and removes motion variables along with body shape variables. Microphone arrays can be arranged in many ways. Because the microphones are fixed, the measurements are repeatable. And because no one is in the car during the measurement, the system is optimized consistently regardless of who is eventually going to sit in the seat. Most target curves were designed around either a single microphone or an array, but not around the human body occupying that space. If you tune with a body present, the body becomes baked into the data, and

17:56 your curve may not be designed for this. If you tune without someone in the seat, the body becomes a predictable filter applied later. Neither approach is wrong, but only one is easily repeatable across cars, locations, and tuning sessions. This is ultimately about simplification. Every added variable, body position, microphone position, inconsistent averaging, it all adds uncertainty. My method is not laboratory perfect, but it is extremely consistent. More variables exist such as temperature, humidity, and the exact microphone position and calibration, but those are outside the scope of this

18:43 explanation. However, laboratory level testing and my tuning can agree on this that consistency is what allows meaningful comparison, repeatability, and reliable results. There's also a practical side to using an array outside of speed and accurate results. Extended exposure to pink noise can affect your listening perception later. With a microphone array doing most of the heavy lifting, we can greatly reduce the exposure to loud sounds and use our ears only for the final refinement. Different measurement techniques will always produce different traces. That's why it's nearly impossible to match data from another tune in another location with a different measurement system.

19:29 Honestly, the impressive part would be if someone could take all of those variables and end up with the same measurement. The goal is to choose a method that is consistent, accurate, and repeatable. In my opinion, that's how a great tune should be approached every time. [clears throat] If there's one takeaway from all of this, it's this. Measurements don't exist in a vacuum. Every method has assumptions. Every technique has tradeoffs and every added variable introduces uncertainty. There is no single correct way to measure or tune a system. But there's ways to measure that are more consistent, more repeatable, and easier

20:14 to reuse over time. That's why I focus so heavily on controlling variables, simplifying the measurement process, and using methods that allow results to be compared meaningfully from one session to the next. When you really understand what your data is telling you, and just as importantly, what it isn't telling you, you stop chasing traces and start making better tuning decisions. That's how reliable results are built. That's how good systems become great ones. And that's how professional tuning should be approached. Thanks everyone for watching. Be sure to like and subscribe to my YouTube channel and also let me know if you'd like more content like this in the future. To book your professional DSP tuning, go to totaluningusa.com.

21:02 See you on the next one.