Low Stage Part 02 - A case of Lobing

At the very least, matching tweeters mounted on the "A" pillars with mids installed in the kick panels is a terribly risky idea. 
Why, then, is it that so many cars use the technique so successfully? Well, it all depends with what you call successful.
In Low Stage Part 01 - Two Speakers, I tackle some of the myths about fixing perceived stage height. In this second part I will further comment on the challenges resulting from the system's lobing patterns.
Sound judges today, seem to have forgotten that live music is large. Instead, they award higher scores to systems projecting caricature-like images. The more violins seem minuscule in size, the sharper the image these judges think it to be.

Miniature Instruments

These judges equate small with sharply outlined. But image sharpness comes from transparency rather than size. It is only after removing many layers of veil that we get to see the difference between the suspended instrument and the empty space around it. Closing our eyes while listening to great quality unamplified music demonstrates this. But if asked whether these sound judges routinely listen to real unamplified music, the response is that of an empty stare. Just think about it, even the Boston Pops amplify most of their performances. It is getting harder and harder to listen to the magic that comes from real good musicians playing real good instruments without ElectroVoice ruining the presentation.

I remember going to listen to Sarah Brightman. The beautiful lady captivated me from the moment I first listened to her music. So, when I had a chance, I paid for the tickets. Armed with excitement, I listened attentively. So can you imagine my surprise when her sound was flat and brittle? It was terrible. Whatever the sound engineer was doing, I hoped would end immediately. Surely his acid trip and the loss of high frequency sensitivity made it impossible for him to understand what good sound was. But then again, maybe I was being hypercritical. Then came intermission. As the lights came on, dozens of people around me begun to complain about the terribly bright sound. It seems I was not alone in my conclusion that the sound sucked. How sad indeed.
The bottom line is that when asked about how to earn critical listening skills, I tell people to just listen to real music. It is that good when compared to the garbage we are now exposed to.

When I audition a system, I pay lots of attention to the size of the instruments. Properly aligning speakers will inevitably result in larger, deeper images. It is this quest that makes me conclude that most systems with speakers on the "A" pillars fail the time domain test. While placing  speakers like tweeters up high helps with the perceived image height, it does nothing to create realistically sized instruments. Have you ever listened to a grand piano. The sound is as large as a room. But most car audio systems, even the better ones, shrink instrument dimensions in the pursuit of so called image definition as they define it.
On the other hand, a system using properly time corrected speakers will do both, large instrument size and solid image height. It is interesting to note that the time correction I am referring to is independent of speaker positioning. In other words, I am not against speakers on the "A" pillars. I am against using them as a way to patch fundamental problems without properly addressing such issues directly. Most system designers path their height issues and completely miss the fact that inadequate time response has killed any hope at realistic image dimensions. Soon a self fulfilling prophesy is created where judges begin to reward poor performance with higher scores because of herd behavior.
At a recent Consumer Electronics Show, I spent time with my friends Chris and Melissa Owen at their high end room. Orca had lent them speakers sporting their latest design. To be clear, the speakers looked fantastically finished. Moreover, these deployed really good drivers from Orca's arsenal. But as soon as Chris fired up the system for me, I told him that the crossover had problems that were creating comb-filtering. It seems as if the designer, like many others within the industry, used 12dB/oct parallel crossovers between drivers mounted on a surface that stood perpendicular to the floor. Anyone who pays attention to this stuff would know that this was trouble waiting to happen. But for the many who think that amplitude linearity is all that matters, the speakers were fine, just as their LEAP modeling software had predicted. So, as we move onto the car, consider that the problems I will address are also ubiquitous in home speakers. Again, it's a bit of a herd problem.
I recently showed a dear friend what happens when a tweeter on the "A" pillars is matched with midrange speakers in the kick panels. I used modeling software that is freely distributed. Oh, how I wish I had these tools when I started in the industry. Back then I had to interpret all the multidimensional variables while trying to make sense of what I was listening. I used no equipment when designing passive crossovers; just my trusted ears.

Polar response for tweeters in "A" pillars and mids in kick panels

 Take a look at the illustration above. It models what the polar response would be for a system with a tweeter mounted on the "A"pillars while the mids are in the kick panels. The crossover uses a 12dB/oct sloped at a frequency of 3.5KHz. Note that both of these settings are quite common in the industry. As it should be immediately evident, destructive interference would make it very difficult to find a listening position where spectral performance could be maximized. Just move your head an inch or so and all dimensional sense will evaporate. This kind of lobing error seems to correlate with poor vertical spread or instrument size; no matter whether in a car or a home system. In fact, one is often surprised of how many so called high end speakers display poor polar response. For many of them, tilting the speaker backwards would solve the issues. This explains why there are so many speaker platform designers that facilitate just such tilt without risking that the speakers fall backwards. It is an expensive way to correct what could otherwise be handled at the design stage. Unfortunately, the problems depicted here are much more gruesome. No amount of tilt will fix them. Either the distance between the drivers needs to be reduced or the crossover point should come down in frequency by quite a bit.

Polar response for tweeter and mid in kick panels

Now look at the second illustration for what happens when both drivers are located side by side to each other in the kick panel. Crossover frequency and slope remain the same. The only aspect worth noting is that to fix the problems associated with the 12dB/Oct design the tweeter has to be physically located on a plane behind that of the mid. This helps til the lobe upwards to ensure that maximum spacial performance occurs anywhere a common listener would place her head.
And if in need of further clarification, the second set up would also project images at eye level. The main difference is that images from the latter setup would be much larger, just as God intended.

Polar Response at 10KHz between
tweeters in the "A" pillars and kick panels

Again, I am not saying that height is wrong when using tweeters on the "A" pillars. I am simply clarifying that by themselves, they are no solution. The same attention to detail needed for properly setting up speakers in the kick panels will be needed with speakers mounted above the dash. Furthermore, to prevent the negative effects from comb-filtering, a less common slope should be matched with a lower crossover point aside from proper time adjustments.
And while we are on the subject, what do you think happens wen people use both drivers in the kick panels and then add a second set of tweeters on the "A" pillars to "lift" the sound. Often, well intended attempts use a higher frequency high pass filter for the tweeters mounted above the dash. Well, chaos is what happens. Look at the illustration to the left. It represents what happens at 10KHz with a 6dB/Oct highpass slope for a tweeter mounted on the "A" pillar and matched to a tweeter in the kick panel. It begs the question of how in the hell would one expect reliable performance; forget about good sound, just getting the same sound all the time would be a bonus.
Can anyone hear the difference? Yes. Can anyone know what's wrong with the music? No, In simple terms, the change is visceral. We just know it's wrong. These types of problems tend to fatigue us quickly.
Are the issues discussed correctable? Yes, but only when the constraints are understood. Unfortunately, many industry designers miss the science side of sound and attempt to solve everything as if it was all an art. While music is certainly an art, we should not waste valuable creative time on things better left to our tools. I hope you found this post as helpful.