Educar has been involved with training 12 volt professionals on process, theory, and system design for many years. These articles address the process of selling, installing, tuning, and how to present new audio system technologies effectively.

Why do so may OEMs use the 6x9/2” two-way front stage?

Recently on the Educar Integration FB group, someone asked about this from an upgrade point of view – “What’s the best way to upgrade a Toyota with this 6×9/2” speaker setup? Should we just go straight to a 6-and-tweeter two-way?”

I’ve recently been working on a few of these vehicles, and I think I understand better than before why some OEMs use these sizes – and what we can do successfully in these cases.

And, as often happens, it all circles back to what I call The Three Problems of Car Stereo:

  1. One Side is Louder.
  2. One Side Sounds Different.
  3. One Side Gets To You First.

All of these are problems that foul up stereo in a car, but can be solved in the home simply by sitting in between the speakers and aiming them properly.

For the purposes of this exercise, let’s assume we’re talking about an amplified premium system with active channels. Yes, I understand that we are often upgrading base audio packages – but the speaker provisions are often the same, and those speaker provisions were designed for the premium systems. This assumption helps explain the logic behind the design decisions which put that size of speakers in those locations.

1. One Side is Louder

So, we know that OEMs prefer 2-seat stereo to 1-seat stereo. The Three Problems are harder to solve for two locations than for one location.

If you put 2”-class speakers in the forward corners of your dash, the distance to the near dash speaker and the distance to the far dash speaker are closer to the same. They aren’t identical, but they are closer than if those speakers were in the top of the door, for instance (think 2G Tacoma, for example – the same-sized speaker was in the top of the door rather than the dash, and this principle did not help out in that vehicle).

So the difference in loudness related to proximity – to attenuation over distance – is reduced with those dash locations. It’s not perfect, but it’s improved.

2. One Side Sounds Different

We know that speakers have a different frequency response in their upper ranges when speakers are pointed AT you, as opposed to when they’re pointing AWAY from you. This is the single biggest reason that speakers sound so different in cars than they do in display boards.

So, in this hypothetical fully-active system with a door woofer and a dash wideband speaker, you can restrict the range the door speaker plays, and prevent it from playing up into that range of frequencies where it becomes directional. The fact that we don’t let it play up into the midrange means that we can keep the frequency response of the left speaker and the right door speaker essentially the same. As far as this door speaker is concerned, we have addressed the Second Problem of Car Stereo pretty well.

What about the upper frequencies, the ones played by the wideband speaker in the dash? Well, it turns out that neither the left-side dash speaker nor the right-side dash speaker is close to on-axis to the listener. BOTH speakers are aimed straight up, almost 90 degrees off axis.

So we have also largely solved the Second Problem of Car Stereo for the dash speakers – since neither speaker is on axis, they sound very similar. (That also means they both have rolloff in the upper treble, so we will probably have to solve that with EQ).

3. One Side Gets To You First

OK, that’s two out of three addressed. So what about the Third Problem? The Third Problem is essentially phase cancellations caused by arrival-time mismatches, which are caused by path-length differences. Sound travels relatively slowly, so speakers which play the same sounds, but are different distances away, interfere with each other like this:

Well, one way to address this problem is delay, and we use delay a good deal in the aftermarket. Of course, delay only works for one unique location, and OEMs prefer two-seat solutions (or solutions for even more than two seats!)

A way that many OEMs are partially solving the Third Problem is phase equalization – changing the phase at specific frequencies to prevent cancellations at the listening positions.

Simple Phase Equalization in cars to me means only addressing the worst problem – which is always around 200 to 250 Hertz when door speakers are used. That frequency is a function of the arrival-time difference, which is a function of the path-length difference. This varies in cars due to the width of the car, but it’s usually around 2 milliseconds, and that translates to 250 Hz.

Let’s create an imaginary car with door woofers and dash speakers, and a 250 Hz cancellation in either front seat.

So “phase equalization” uses all-pass filters to change the phase. All-pass filters are filters which invert phase in specific frequency ranges. Let’s use one all-pass filter at 250:

Using one AP filter, at 250 Hz in this imaginary vehicle, will get rid of that 250 Hz cancellation in either front seat. The imaginary vehicle will now have a 250 Hz cancellation if measured in between the two seats – where the path lengths are equal and the arrival times are simultaneous – but we don’t care, because we don’t sit there.

So right there we have addressed the Third Problem to some degree, for the notes played by the door speaker. What about the dash speakers?

Well, some base OEM systems don’t do much of anything up there. When we are upgrading these systems, we may decide to bias the system for driver-seat performance, and so we may decide to use delay on these channels to solve this problem for the driver.

I’ve done this in a 2018 Subaru with base audio, and a 2018 Toyota Tacoma with base audio. Both of those cars had a 250 HZ all-pass filter.

In the Subaru, we installed a full-active 6/3/tweeter 3-way front stage, plus a subwoofer in the back. This was intended to be a VERY high-performance driver’s-seat system.

When I tuned it, I matched the level of the left and right door woofers, I matched the response of the left and right door woofers with EQ, but I entered the distance to the passenger woofer for both the driver and passenger woofer channels when setting delay. That means that the signal to the left and the right woofers were not delayed relative to each other. With a normal signal, there would have been a cancellation at 250 in either front seat – but Subaru’s audio vendor had already fixed that problem for us by using Phase EQ at that frequency, and their solution worked fine for us. If we had tried to use delay in the normal manner, then we would have had problems.

When I tuned the mids and the tweeters, I did it the normal way. I matched the levels and the responses left and right at the driver’s seat, and then I entered in the actual distances. That solved all Three Problems of Car Stereo (for the driver, of course).

Another way to look at it:

This does mean that there was a slight phase deviation in the crossover transition band from the top end of the door woofer to the bottom end of the dash midrange on the driver’s side. We aren’t matching phase perfectly, but we aren’t 180 degrees out either. The difference is a dB or 2, but not the 30dB theoretical maximum that results from 180 degrees of phase cancellation. It wasn’t noticeable to me, and the client told us this system was the best-sounding car stereo system he’s ever heard, so I guess he didn’t have a problem with it either.

(Technically, there also is a very slight one at the bottom end of the door woofer where it integrates with the subwoofer signal, but the wavelengths are so long at this frequency that being slightly off really isn’t an issue at all).

In the Tacoma, we used a 6×9 and 2” two-way speaker system in an active setup up front, and a sub behind the second seat. Same arrangement:

Other OEMs will use more phase equalization on the dash channels (coughBOSEcough), to address more of these cancellations. In these situations, we don’t want to use delay without first putting those channels back onto phase with each other. (That requires either some all-pass filters and testing, or an external preamp that eliminates the OEM processing). In a future article, we will discuss options for these vehicles.

So, what if you aren’t using an OEM source? Is any of this beneficial?

It can be. In last month’s magazine, we profiled a Toyota FJ Cruiser project which used an aftermarket Sony head unit. In that project, we used phase equalization ourselves for a two-seat presentation. We have had questions about how I did this, so here are the specifics:

We kept the L and R level settings matched for each speaker. If we matched the actual levels for one side, that would foul up the levels for the other. The output levels are of course lower for the dash speakers than for the door woofers, but Left and Right dash speakers are set to the same output level, and so are Left and Right door woofers.

We used no delay anywhere in the system, except for 12mS of delay on the D-pillar L-R effects speakers (not relevant to this discussion).

We tied L and R EQ together for the dash speakers, and also tied L and R EQ together for the door speakers. Again, making any adjustments for one seat would just compromise the sound in the other seat. This way, any changes we made affected both seats.

We restricted the door woofer to frequencies below 500 Hz, which meant the door speakers only played the big 250Hz cancellation. The one at 750 wasn’t a concern. One all-pass filter at 250 Hz with a Q of 5 solved that biggest cancellation.

Then we high-passed the dash speakers at 500, and used an AP filter at 1000 with a Q of 6. We tried a few different approaches, but this was the one which sounded best to me out of what we tried. Remember, the dash speakers are different distances than the door speakers, and much closer together, so the cancellations start at a higher frequency.

Now, we also did a one-seat tune – you may remember we installed a button to switch between the two. You won’t confuse them. The one-seat tune has a pinpoint image, and the two-seat tune definitely feels wider and more spacious. Well, some research has shown that normal people often prefer a more spacious presentation, and guess what? My sister and brother-in-law, who own that FJ, are two of those normal people. I think they run it in 2-Seat mode all the time, even if there’s no passenger.

Myself, I prefer the one-seat setting from the driver’s seat – but by a little, not a lot. Because of what I know about sound, I notice some of the out-of-phase cancellations which are not corrected – but I’m not “normal”, and I’m not the customer.

This system addressed – to some degree – all Three Problems of Car Stereo. The left and right levels are closer to matching than they would be with most 6-and-tweeter set. The left and right responses are definitely closer to matching than they would be with a 6-and-tweeter set. Finally, I think these improvements combine with our addressing the biggest phase cancellations caused by The Third Problem of Car Stereo, and the result is the worst of those cancellations solved, and an enjoyable two-seat presentation. That’s a result we couldn’t have achieved with a 6-and-tweeter 2-way system.

Back to the original question: Should I use the OEM sizes, or a 6-and-tweeter?

If you’re using an OEM signal, you NEED to know if two-seat Phase EQ is in use. If it is, then your answer depends on your goals. Summing is a bad idea when there’s a lot of all-pass filters in use. Better to use the OEM crossover point, and NOT sum – and to use the OEM crossover point, you need to use similar sizes.

If there’s phase EQ on a full-range signal – think Toyota or Subaru above – there’s no question of summing. In that case, what are YOU doing? If you’re doing 2-seat stereo, you probably want to go with the OEM sizes, so you can use the technique explained above. If you’re going for one-seat, and you go to a 6-and-tweeter 2-way system, then you need to undo that phase EQ somehow – with a DSP that has the capability to do it. If you can’t, re-consider that technique above.

If there’s no phase EQ in use – which is true if you’re using an aftermarket source or an external preamp (which bypasses the OEM processing), then the next question is “Are you aiming for stereo for one seat, or two?”

If the answer is one-seat, then it doesn’t matter which sizes you use. A 6×9 is probably a bit louder than a 6.5, and a 2” is probably a bit louder than a 1” tweeter. But it’s not critical. If you’re using a 6×9-and-tweeter 2-way component set, I recommend using one with a tweeter which plays as low as possible. If you can get another octave of low-end extension out of the tweeter, that’s an octave that we aren’t trying to get out of the top end of a 6×9.

If the answer is two-seat, then consider using the same 6×9/2” approach as the OEMs – IF you have equipment with all-pass filters available to you to do your own two-seat phase equalization. Very few DSPs have all-pass filters available.

6×9 woofers and 2”-class midranges are available from Audiofrog, Audison, Morel, Kenwood, and some others.

Audison Prima AP690 and AP2

Audiofrog GS690 and GS25