PDA

View Full Version : OmniMic in the car



ErinH
02-21-2012, 08:45 PM
Been using OM to help me with some car mode problems. Originally posted this on DIYMA (my stomping ground) but thought some of you home folks might like seeing OM used for car audio application:



figured it's time to update and see if I can get some convo going again. not for the sake of me seeking help, but to share some insight and maybe make some folks go "hmmm"...


I long ago said that my midbasses make the system perform much better when they have no HPF enabled. People are apalled to learn I rock 7" drivers all the way to 20hz. Been doing it for about 1.5 years now and so far, so good.

I did, however, recently try a stint at using a HPF on them at 50hz and using the sub around the same point. Sub with a 36dB slope and midbasses with 24dB slopes. Got it?
Okay...

The problem arose whenever I'd get a midbass guitar or kickdrum; you'd feel it in the back. Grrrr.
I played around with varying the sub to midbass transition in a multitude of ways (changing the crossover points, changing the slope, flipping phase, etc). Nothing really worked as well as when they midbasses were played with no low end filter. I had an idea why but I wanted to see if it held up. I popped in a track that makes everyone's car do the nasty: Spanish Harlem. Yes, I know you hate it, but it does a great job at highlighting deficiencies in a stereo system.

I used my omnimic with TrueRTA and set TrueRTA's "peak hold" to on. I played the first few seconds of Spanish Harlem at a pretty loud level and recorded the sound with the midbass crossover slope enabled and at 24dB/octave.
I then played the track again, this time with the HPF slope inactive (0dB/octave).
The goal here was to capture the bass resonance of the stand up bass behind Rebecca that often "blooms" in many audio systems and causes the entire audio response to be compromised. Capturing this would allow me to see exactly what was going on with the varying crossover/slopes.

What you see below is a picture of the response of the first 10 seconds of the track, with Peak Hold set to on with both 'versions' of the system response. 24dB slope is in Green. Slope off is in Purple.
http://i18.photobucket.com/albums/b138/bikinpunk/resonancemidbasscomarison.png

You'll note the large gap in response at 40hz. This means something is out of phase. Surely this sounds bad, right? Well, as it turns out, it seems to do a good job of creating an anti-node. What do I mean? Simple: the car has a modal response driven by geometry at about 45hz. I know this because I did some testing with the omnimic and the OM software using their Bass Decay feature.*
Here's the response of the car's bass decay. This essentially is a means of determining resonance in the 'room' (ie: room modes).
In the picture below, you can see the scale on the right side that gives you color vs. level of resonance. This level is absolute; it is not given relative to the frequency response. You'll see the large increase in resonance at about 43hz.
http://i18.photobucket.com/albums/b138/bikinpunk/resonancemidbassallpass.png
(side note: See that one at 160-200hz? That's a serious issue in my car, too)

As it turns out, while the sub to midbass transition is out of phase here when looking at the FR measured (not pictured here), it actually sounds very good. The midbass bloom goes away to a large degree and the upright bass is in front of you. The majority of any tactile feedback in the seatback is removed and the bass shifts up front by a perceptually large amount.

My hunch is that the two drivers being out of phase at/near the crossover is helping to absolve some of the room mode issues and is lending itself to more 'upfront bass' response, rather than 'hit me in the back of the seat'. It seems that I've created an anti-node right where it was needed. Only a theory...
And, this actually does help prove to some degree my OP even further: don't always solely depend on a graph. You'll recall (before this thread went to hell and back) my goal was discuss room modes, using an RTA, using your ear, and coming up with a helpful discussion on the low end response; the Achilles' heel of car audio (and home audio, too).

I can't say,... won't say..., if this always works for everyone. It is simply a means to further display how tricky this hobby is. Hopefully this will give some folks motivation to try something outside of the norm and possibly have good results.



*Note:
I love the omnimic. Yes, it's expensive compared to other products but it's incredibly simple to use and is packed FULL of great features.


Edit: This post is NOT intended to start a "run your midbasses to the ground with no/low crossover" discussion. Just an observation regarding phase and mode relationship and how they can (possibly) be played against each other with good results.



- Erin

ErinH
02-21-2012, 08:50 PM
a bit more from the discussion on diyma...



Mark, take a look at about 180hz. See how the 15-20dB resonance lasts for about 75ms?
That specific area in my car is an issue. I think it's a corner loading effect due to the midbasses mounted in the kicks. I've had problems there since day one and have had to either build the crossover around it or EQ the crap out of 160 & 200hz to bring it down.

The problem is that a modal issue can't be EQ'd, really. It can to some degree but always an underlying issue. This is where a bass trap would help, but you can't fit them in a normal sized car. The next best thing is an equalizer with adjustable Q and broad levels of attenuation. This is why the Arc PS8, Helix C-DSP, Mosconi, and Zapco DSPs are so nice. You can vary Q. However, the Arc is the only one I know of that allows such a narrow Q of 25 (iirc).

Just posting this to highlight another issue and how useful these kind of measurements can be. I never knew until 15 minutes ago that the issue at 180hz was so obvious. It stands out more than the other issues, to me, because it's such a high resonance with a long decay.

musiciansdelite
02-22-2012, 07:54 PM
What is your system and car type?

ErinH
02-22-2012, 09:44 PM
2006 Honda civic sedan.
Scanspeak illuminator 18wu midbasses in kicks
Illuminator 12mu tweeters and illuminator 3/4" ring radiators in pillars
Acoustic elegance IB15 subs in trunk wall
Pioneer p99 headunit

Lots of acoustical treatment here and there.

ernperkins
02-27-2012, 11:08 PM
Erin -

It would be interesting to see the bass decay repeated using some smallish 2-ways located in front of the kicks. It would help distinguish between an actual cabin mode and your kick enclosures. Easier said than done, I know....

I've been fighting the OmniMic and getting along with HolmImpulse and REW. You guys aren't making it any easier by showing all the neat things OmniMic can do :)

Regards....

ErinH
02-28-2012, 12:20 PM
Here's a picture of my friend's Nissan Murano. He built his dash to house his midbasses. This is the best bass decay I've seen yet.

http://i12.photobucket.com/albums/a224/pionkej/Murano%20Build%20Log/Johnsbasssweep.png

winslow
03-01-2012, 10:01 AM
I wonder if the number and locations of the midbasses have anything to do with it?

jbruner
03-04-2012, 08:35 PM
ErinH,

Thanks for sharing this informative anecdote. I think this thread deserves to be in the main forum for discussion. Many around here have recommended to high-pass the mains, but it is not always the best solution. I found an improvement in the (single sub) bass in the living room when I ran a 6.5" rear surround(40-ish F3) full range. The point is just try something, it might be better!