12volt Tech: Localization Cues and How to Manipulate Imaging

- Image: Sound localization dummy head

When it comes to achieving proper "staging" in a vehicle, there are many considerations that are highly critical. There is a lot of voodoo surrounding this concept and it is interesting to hear how much weight is put into only one aspect by even very seasoned sq competitors. Hopefully when I help identify what gives you your stage, you'll see what needs to be done to make improvements. Keep in mind that even with the greatest stage ever, with no beaming or lack of bin-aural separation, will still be very disappointing if you don't have the tonality and detail to go with it. You only sound as good as the weakest link in the chain.

So, three very critical concepts that you need to understand before going forward.

1. Inter-Aural Level Differences - ILD (or sometimes referred to as IID where the second I stands for Intensity) refers to the differences in SPL between your left and right ear.
2. Inter-Aural Time Differences - ITD refers to the differences in arrival time of a sound between your left and right ear. Most often, this is referred to as PLD, or Path Length Differences. It's a bit of a misnomer, I suppose.
3. Head Related Transfer Function - HRTF refers to how sound is altered by the diffraction and reflection of your head, pinna, and torso.

The first and second concepts refer to the localization of sound in the horizontal, or lateral, plane. HRTF relates to the localization of sound in the vertical plane. In simple terms, your left, left-centre, centre, right-centre, and right locations of the stage are determined by ILD and ITD, whereas your stage height is related to HRTF. ILD and ITD are considered binaural localization cues that rely on a comparison of the input from your left and right ear. HRTF is considered a monaural localization cue as it relies on the filtering of the source material by external structures.

Now how are these relevant? Let's start with ILD. The goal here is to control the SPL (or frequency response) at the right and left ears. If SPL is higher at a given frequency at the right ear than the left ear, you will interpret the sound as originating from the right. The larger the difference, the more bias you'll interpret. If we assume that SPL is identical at both the left and right ear, then this will be sonically interpreted as being at the center of the stage. This effect is most predominant in higher frequencies covered by your upper midrange and treble portion of the spectrum. In rather general terms, it occurs around 1500Hz, or around the frequency at which the wavelength is shorter than the distance between your left and right ear. There are many factors that ILD is dependent upon, and they are highly related to where you put your speakers and how you manipulate its surroundings. We know that SPL drops the further we are from the sound source, so if the right channel is closer than the left channel, your stage will be smeared to the right (more on this concept in a moment). If the right channel has considerable reflection off the window and creates constructive interference at a given frequency, you will again hear the stage "pull" to the right, as SPL will be higher at the right ear than the left. The opposite can occur as well: if the right channel encounters some form of diffraction or reflection that attenuates response at a specific frequency, any content at that frequency will be pulled to the left, as SPL will be higher at the left ear. There is an easy way to measure this: a microphone carefully placed where your ears will be. Obviously, this becomes substantially more challenging to manage when you're wanting to create a realistic stage for both you and any other passengers in the vehicle, but I'll touch on how we implement this knowledge in a moment.

When it comes to ITD, it's a substantially easier concept to understand and manage. The easiest way to avoid ITD related issues is to make the distance from your left channel to your left ear and the distance from your right channel to your right ear identical. ITD is most dominant in your midbass and lower midrange, so that's where we need to focus our attention. This, again, is a little complicated to compensate for electronically with multiple passengers, so it works best when physically installing your speakers in an area where there is very little ITD.

I'll save HRTF for a moment as it's a very complicated concept. How do we work ILD and ITD into our installation and product choice?

As I mentioned above, ILD and ITD are hard to compensate for electronically for multiple passengers. With ILD, if we boost response on the right channel at a given frequency to improve ILD for the driver, all passengers on the right side will see the whole stage shift to the right. Obviously, that's not a good thing, though I know it's hard to find dependable passengers as dedicated to the cause as we are. With ITD, if we use time alignment at the head unit to delay the left channel and improve ITD for the driver, the passenger will find the left channel even more delayed than it already is, as the left channel is already farther away from the passenger than the right channel. Given these examples, it's quite clear that we simply cannot compensate for ITD and ILD electronically and achieve good results for both the driver and all passengers. The only way to do this is to physically make this work with your installation methods.

Regarding ITD, this involves installations as far forward in the vehicle as you can make happen, which usually means installing in the kickpanel. There are some rather noticeable downsides related to kickpanel installations. Diffraction caused by your feet and legs can be quite an issue. Usually you can estimate that the frequencies most affected by diffraction will have a 1/4 wavelength equal to the width of the object. For example, if you assume a 6" leg is in the way, then you can estimate that diffraction will be a concern for all content above 550 Hz. Your midrange portion, where ITD occurs, has considerable material above 550 Hz (note that the frequency this occurs at gets lower as your leg gets bigger). So the kickpanel resolves some of our ITD issues, but diffraction is still a considerable problem (and some avoid kickpanels because of the muffled sound). The kickpanel is also one of the most obtrusive and unattractive areas to install a speaker. Still, the benefit of achieving the best ITD possible in a vehicle is a hard one to turn down. Maybe just saw off your legs. ;)

Regarding ILD, this involves careful planning and, if all else fails, some clever EQ work. There is really no way around this. If you have your tweeters down low in the kickpanel, you will deal with diffraction and early reflections from the dash, floor, and door. If you have your tweeters installed up high, either on the dash or in the a-pillars/sail panels, you will have to deal with considerable refraction, which can be a little harder to tame. An equalizer with as many bands as possible above 1000 Hz is a really good idea. Additionally, you'll want to treat the surrounding areas. If installed up high, consider a dash mat at the absolute least. I know that they are rarely appealing, but it's a necessary evil to help reduce the effects of comb filtering. If you're doing something in the kickpanels, consider using soft floor mats, keeping each kickpanel identically shaped/sized, and perhaps treating the areas around with some sort of material that will dampen sound, similar to using a dash mat. You will also want to consider the center console: if it's tall, it's going to be a cause of considerable diffraction when material from the channel on the opposite side of the console attempts to reach your ear. All in all, you want to keep as many objects out of the way as possible with the least interaction on dispersion of the sound wave. Barring that, your only choice is to compensate for this with an equalizer and boost or cut at various frequencies on each channel until you have minimized the effects of ILD.

That's pretty much the gist of managing your localization cues in the horizontal cue, ie. the left and right portions of the stage. But even more challenging is the managing of the stage height, and that's entirely related to Head Related Transfer Function.

Unlike ILD and ITD, HRTF is not something that is universal. With minimal ILD and ITD, you can swap people in and out of your car all day and the width of the stage will be about the same for everyone. This is not the case with HRTF, as it relies heavily on the physical attributes of the listener. The diffraction and reflection properties of the head, pinna, and torso are of considerable importance when determining HRTF. In mathematical terms, HRTF can be described as a ratio between the output signal and the input signal as a function of frequency. Earlier we determined that the kickpanel is a good location to keep ITD to a minimum. However, this has a tendency to pull the sound stage lower in the range of frequencies where HRTF is critical. I don't believe I've mentioned this yet, but HRTF begins around 1-2 kHz. This is something very worth noting: if our localization cues along the vertical plane (given by HRTF) don't begin until 1-2 kHz, is it as critical as everyone says to make sure the tweeter is mounted right beside the mid? If we can't localize the mid in the vertical plane, we should be able to put it where ever we want and still not feel that it's pulling our stage down. Try it and I'm quite sure you'll agree that's the case.

I fear I'm rambling a bit, but to get to the point, HRTF is listener dependent and begins around 1-2 kHz. Your HRTF can be directly calculated in the frequency domain using sweeping sine waves, but it can prove fairly challenging. It is perhaps more prudent to do this on a trial and error basis. The beauty of HRTF is that it CAN be compensated for electronically. Though it is easy to move your speakers up high to give you the stage height you desire, corrections using an EQ or a notch filter can raise or lower the soundstage as you see fit. Doing so, without having calculated your HRTF, while require some trial and error though.

When we take all those parts and put them together, where is the best place to put your speakers in a vehicle? Obviously this is very, very product dependent, but here are a few generalizations I'm relatively comfortable making.

1. 2 way (1) - If you're willing to compensate for HRTF electronically, you can put a large, high output midbass in the kicks with a tweeter of your choice. The pros of this route is that a) ITD is not an issue, b) ILD is less of a concern here than if mounted near the windshield, c) HRTF can be compensated for electronically and you can still get that stage height you want. The con is obvious: you must compensate for HRTF electronically! Also, it can be challenging to use a high output midbass and still get good upper frequency extension from the same product: output comes from displacing a volume of air, the more the better. This can be achieved by using a large diameter mid, but this is tough to install in the kicks. You could also use a mid with good linear excursion, but this generally requires a longer voice coil, which has higher inductance, which rolls off your upper frequency response. Still, this isn't a bad way to go.
2. 2 way (2) - We'll go a little different route here. First, we'll still use our high output mid (everyone wants it; it makes crossing over with your subwoofers very easy). This will be installed in a kickpanel, once again. Second, we'll use a tweeter with great low frequency extension. By great, I mean that it has an Fs of at least 750 Hz, preferably lower. The tweeter will go up in the a-pillar. Ideally, the tweeter will be crossed over as low as possible. It's usually preferable to crossover no lower than an octave above Fs, but you can go lower if you increase the slope. This is a good alternative because you can use a high output mid without worrying about it's top end extension too much, all vertical plane localization will be coming from the tweeter which is mounted up high, and ITD is not an issue as the mids are in the kickpanels. You still need to try and correct for ILD by using an EQ to even left/right response. The real issue here is finding a tweeter that is extension below 2kHz and is still very detailed. For the reasons outlined so far, horns are very popular in 2 ways: they have the extension necessary and do not lack much in the detail department. Using standard tweeters, this can be hard to achieve, but the results are very pleasing. The Seas Lotus Performance series comes to mind when I think of successful implementations of this nature coming from a manufacturer. You must be extremely careful with tweeters up high, though, because it can be very hard to control the comb filtering effects the dash introduces.
3. 3 way - A 3 way design is great because it lets you use a large midbass for the bottom end without worrying about upper frequency extension, a small tweeter for the top end to give you the detail you want, and a widerange mid that can handle most localization cues. I think the tweeter in the a-pillar passed above at least 5kHz, the midrange in the kickpanels handling 300Hz-5kHz, and a strong midbass in the doors lowpassed from there (obviously transitioning to the subs at whatever frequency works best in your implementation).

As you can see, there is a compromise behind everything. What you work with depends on the product you use, the effort you're willing to put in, and the electronic capabilities you've got available to you. If you're planning for your first SQ competition, you should understand the concepts above as a bare minimum. If you're just an enthusiast, try to manage your localization cues as outlined above and hear how great sound can really be.