Just recently tested my car with the RTA and some pink noise, and there were some pretty bad dips in the overall frequency response. Wish I had a pic of the display, but I don't. Basically had a huge >10dB dip at ~600Hz 1/3 OB and a similar but lesser sized dip at the ~1.2kHz 1/3 OB, but the rest of the FR was fine. Stuck the mic at the speakers, and they turned out quite flat, so I guess it's just a problem with the car. Any tips on getting it to be a little better other than boosting and cutting out sections on the EQ? I've got a 31band 1/3 OB eq that I'm using. Too bad I don't have control over the left and right channels separately. I haven't yet tried checking each channel separately and seeing if it's a problem with that. BTW, i'm using the Koda component set, crossed over at ~60Hz to the sub which is an e8a. Koda mids are mounted in the stock location in the doors of the car (Basically a civic), and the tweets are mounted in the kicks, aimed <10 degreesish towards the centre of the car (aimed almost straight towards the back of the car). I guess it's not a problem with the crossovers, since they're using the standard passive set which crosses the tweets and mids at 3.5kHz. Any suggestions would be helpful. B)
I assume there is no center console right? A dip like that is likely cancellation by the 2 mids firing at each other. If it were a boost in frequency I would suspect more of a refraction issue creating unwanted reinforcement. Optimal correction would be to aim them slightly towards your head... basically making the crossing of the 2 wavelengths not quite so direct, hopefully eliminating the cancellation, but more likely just reducing it and hopefully tremendously. Also, you may be able to build a soft "console" or loading panel in the center of the car. This will effectively eliminate the waves colliding. I suggest soft, because by using a carpet or a high pile felt you will be avoiding refraction/reinforcement. The cool thing about this is that you can simplify testing using a hank of cardboard and throwing some high pile felt over it loosely to find the optimal hieght and location. When you have it perfected, build it out of glass or wood , mount it, and cover it.
Aiming of the mids is not possible due to the fact that they're mounted in the doors and there is practically no leeway in moving them. Also, i'm not sure how aiming them would help - considering that changing their directions basically change the level differences between the two, the wavefronts still reach me at the same time - moving the distance of the speakers, say to the kicks, would be better for the pathlength - which I would think makes more sense to stop cancellation. I guess if it's possible to reduce the level difference through aiming, it will probably reduce the cancellation level, but it'll still be there. But the level is fine for the other frequencies - and it might just move the dip to a different frequency. Quick calculation: freq vs wavelength ~600Hz ~= 0.6m ~1kHz ~= 0.35m Hmmm..... There IS a centre console and it doesn't rise high up, but it still does block direct sound from the mids to each other. I'm afraid that if I make an absorber panel, it'll look pretty ugly. Especially using high pile carpet.
I'll try to ask about that on tuesday and see if Dan has any ideas. BTW, I did notice those dips were exactly an octave apart. That should tell us something, and might make it easier to fix. BTW Seth, He's got IIRC a newer Civic. Steven Kephart Adire Audio
Wait a minute.... I was using uncorrelated pink noise for the tests... there should be no cancellation from this type of output - and even if there was, it wouldn't be to this magnitude.
Alrighty - a bit more detailed listing of my findings. I didn't have much time to work on it, nor actually to listen to some real music... I'll put it up in spreadsheet format - easier to visually see it. These are 3s Leq measurements of the FR on a 1/3Octave band analyser. I did several measurements in the general vicinity of where my head would be and it didn't change too much. As you can see, there's huge differences in the left and right channels. Especially at 100, 400 and 800Hz. Slightly lesser differences also at 160 and 250Hz.
I've got a 95 Civic, and I've always had very bad dips right around my midbass region.... not an issue with the speakers... and definitely lower than where yours are - but possibly for the same acoustical reasons (these cars tend to evolve over time, not become radical all-encompassing redesigns). I'm still trying to solve mine, honestly (I never have time to work on my own damn car)... I've messed around with flipping the phasing on individual drivers, EQ, everything short of time alignment - which I DO have available to me (Alpine HU). Maybe you'll solve yours before I have a chance to really sit and work on mine again, and you can tell me what you ended up doing?
Darn... time to get a new car? lol I guess it's time for me to get an EQ that allows separate left and right channels.... preferably parametric... I've been toying with the idea of hooking up the mls measurement system that I can borrow to check out in more detail the actual FR of the system.... It'd be interesting at least. If only I could get off my butt to do it... I'm actually quite busy now...
