There is a mountain of stuff to talk about here.
Getting a perfect sound stage for all listeners isn't feasible without a center channel and a lot of processing. getting a nice sound stage for both front passengers is doable without a center channel' date=' but the "center" for each listener will usually be different (closer to them).
For over 10 years I've ran head units or processors that give me T/A presets. I have a preset for myself only. A preset for the passenger only. And a preset for both seats. The first two produce an excellent sound stage with the center being truly center of the sound stage boundaries. The preset for both seats still sounds good and has correct height, but the center appears more in front of the listener which means it's not in the center of the sound stage boundaries. This is fine for daily listening but obviously not going to cut it for world finals.
With a passenger or a demo, I don't mind giving them the preferred sound stage.
Achieving the desired sound stage requires you consider IID, ITD, PLD, and HRTF. You need to know what effect each speaker will have on localization. This is determined by frequency response, aiming, location, and airspace.
Proper time alignment should be done in steps and requires considering the virtual source you are creating.
Step 1 is to align all of the front left drivers so they appear as one virtual source.
Step 2 is to align all of the front right drivers so they also appear as one virtual source.
Step 3 is to then compare left vs right and add more delay to all of the left speakers to center yourself inside the virtual source.
Step 4 is to bring in rear fill (if any). I haven't found a good algorithm for this other than to say good rear fill should meet the following criteria
....a. it should be band limited between 400Hz-4,000Hz
....b. it should be delayed to create the sense of a larger listening space (not just delayed based on path length)
....c. the level should be adjusted so it is not noticeable when it's on, but turning it off is an obvious change.
....d. it should be an L-R or R-L mix to remove center channel information
....e. the goal should be to simulate rear reflections in a larger listening room that has proper acoustical absorption/diffusion.
Step 5 is to bring the subwoofer system into the mix. Again, I don't have an algorithm yet, but in general, you don't want it localized and it should not take away from bass in the soundstage (T/A, crossover, and level play a role)
the concept of a "virtual source" is important to grasp. ideally, we would place/aim all drivers on each side such that their acoustic center axis would converge at some point outside of the car. simply imagine (or draw) a line through the center of the speaker. placement and aiming can work together to allow these lines to converge. we use this concept in pro audio speaker arrays.
You hear people say "swap phase on something" to fix an imaging issue because it is out of acoustic phase with other drivers. . This can be achieved in two ways:
1. reverse polarity on the driver (physically or electronically)
2. change phase by adjusting crossover slope.
The biggest culprit in car audio are crossovers. Passive or electronic - crossovers change acoustic phase. In addition, passive crossovers are influenced by impedance - and impedance is not constant. You need to pay close attention to what your crossover settings are doing to the phase of the speakers. 12dB or 18dB slopes will reverse phase. with modern processing equipment you can choose slopes so that you retain the same acoustic phase for all drivers, which negates the need to change speaker polarity.
phase is relative.
polarity is absolute.
note that running a simple 2-way component set with passive crossovers using 12dB slopes (common) means that both the woofer and tweeter phase are reversed. no problem since the tweeter and woofer are still in phase relative to each other. what if the component system is bi-amp or if you run active? you still have the 12dB tweeter high pass and woofer low pass - now if you put a 12dB slope high pass crossover on the woofer it is now out of phase relative to the tweeter. This may cause harshness and other issues that people attribute to bad speakers/tweeters. reversing polarity on the woofer will return it to be in phase with the tweeter. Now the subwoofer phase relative to the woofer comes into play in a similar manner. if the crossover slope is 12dB then everything is in-phase now. if the sub crossover slope is 24dB/oct then the woofer and sub are out of phase. a simple phase reversal of the sub system will correct this, and the reason most equipment include the phase switch (0/180 deg).
i typically combat this by using 24dB/oct slopes for the tweeter and woofer channels and 18dB/oct slopes for my midrange and midbass channels (since midrange and midbass have bandpass crossovers - phase is reversed twice).
you really need to pay careful attention to phase relative to each driver type and crossover slope. until that is resolved, nothing else can be successful.
the only real mystery is what electronic crossovers are doing to phase. what they should do and what they actually do can vary.
another item i'll touch on is level. level, or volume, is important to consider for each driver. localization of high frequencies is mostly influenced by level (IID). i believe this is the real reason passive crossovers include level adjustment - to allow you to turn down the tweeter closest to you. doing this can help greatly with imaging even without T/A. tweeter aiming also plays a strong role since tweeter response rolls off very quickly as you move off-axis. this is why my tweeters are aimed at the opposite passenger's head rest.
lastly (because it's late), reflections are detrimental. aiming and placement should consider what effect reflections will have on phase interference (when i use the term phase interference, i mean it to include both constructive and destructive). reflections will create additional virtual sources which will greatly confuse the sound stage. sadly, we have a huge reflecting plane next to our ear. that allows information meant for the right ear to be easily reflected into the left ear, a short time later. this plays a role in localization because it has an effect on the HRTF and intended shadowing.[/quote']
