This tutorial was done in my home, but the same concepts can apply to a car.
My home stereo is in an asymmetric room and I have problems with the left and right speakers sounding very different. This disrupts my stereo imaging, is very annoying, and cannot be fixed with DSP. So I examined the room and experimented with what I could do to improve it.
There is a ceiling over my right-side tower but there is no ceiling over the left-side tower. This creates extra reflections and more energy on the right side of the room. I did not see an easy way to add similar reflections to the left side of the room, so I tried removing the extra reflections on the right side of the room.
To find the reflections, there are a few methods. A common one is to use a mirror on the wall and then you can "see" the reflections. If you sit at the listening position, and someone holds a mirror on the wall, they can move the mirror around until you can see the reflection of the speaker in the mirror. The reflection occurs at the point on the wall where the mirror is at. I don't have an extra someone to hold a mirror for me at 2am when I'm tinkering, so I tried something different.
I put a flashlight on the speaker and aimed it at the wall near where I think the reflection is. Then I walked over and held a small mirror on the wall, and moved it around until the reflection of the flashlight landed at my listening position. I used a microphone boom to hover where my head usually is when I'm sitting on the couch, and I worked until the mirror shined the light right at the end of the mic boom. I noted where the mirror was on the wall, and marked the spot on the wall with a piece of masking tape.
[]
I did this a couple times on the surfaces where I knew there would be a reflection like the bounce off the ceiling, and the bounce from the rear of the speaker to the front wall (I have dipole speakers so I have rear-firing tweeters and midrange). I marked these spots with masking tape. These are all pretty good guesses at the moment.
Then I did some measurements with a microphone at the listening position. You want to find the impulse response, and you can do this with REW or SMAART or OpenSoundMeter, or any program really. You also want to do this with the high frequency drivers, one at a time if you can. I measured the impulse response of my left and right tweeters individually, below.
I saw some extra energy on the right side tweeter's impulse response. There are two extra squiggles that are not present in the left tweeter's impulse response. So, there are two reflections I need to find and treat. While I'm at it, I wanted to try and identify each of the reflections to learn how to find and identify any arbitrary reflection, like practice I guess.
The first problem is the reflections in the impulse response can be difficult to see when viewed with a "linear" scale like the pictures above. Most programs allow you to view a "log" or logarithmic scale too which can make the reflections stand out much more clearly. Below is the same measurement but plotted in log scale to show the difference.
REW can certainly make these measurements, but if you are using an XLR microphone (not a USB one) and have a loop-back cable on your sound card, then you can do these measurements in real time. I can take a pillow and hold it on the wall, move it around, and see in real-time the effects of absorbing the reflection. This makes it super easy to pinpoint where the reflection is, if the absorber I'm using is working well, and saves an incredible amount of time.
So that's what I did. I held a pillow on the wall in the two places that I marked with masking tape earlier when I used the flashlight/mirror method. I watched the impulse response change in real-time and I could see one of the peaks dissapear when I had the pillow in exactly the right spot. Turns out, my masking tape from the flashlight/mirror method was spot on.
With the pillow on the ceiling, I could look at the impulse response and clearly see the peak at 3.25ms decrease:
Further experimenting showed the first few reflections were:
1.00ms - floor or trunk or couch itself
3.25ms - bounce of ceiling
3.75ms - rear firing tweeter bounce off front wall
6.00ms - rear firing tweeter bounce off front wall, then bounce off ceiling (oblique reflection)
The bounce off the front wall at 3.75ms is present in the left tweeter also, so I do not have to treat it on the right tower's tweeter. However, the stereo image was still pulling to the right so I treated it to remove a little more energy from the right side of the room.
Below you can see the three early reflection paths, and the couch cushions that reduce/absorb the reflections. I will get some different materials for long-term, but these cushions were nearby and convenient.
Measuring again shows I've removed the asymmetry caused by the reflections from the ceiling.
And here is one more picture showing the imbalance of reflections between the left and right speakers. Notice the left speaker has no ceiling reflections at all. The sound from the left tower travels up to the 2nd-story landing without obstruction.
Result: a better match between the left and right speakers from the couch. However, still more work to be done. There are other influences in the room I have not discovered yet. I also need to find a long-term solution for the materials so I can keep using my couch cushions, and stop tying junk to the ceiling with rope lol.