Summary
These three graphs show the measured audio frequency response of the HiFi stereo in my 2008 MCS. Each graph is at a different volume level (each one is 8 clockwise clicks louder on the stereo than the previous one.)

Caveats
Consider these casual measurements. I don't have a calibration file for my measurement mic, and I used the built-in audio of my laptop, not an add-on card. I also don't know what I'm doing. Time will tell how much that last factor matters.

Equipment
I used a Behringer ECM8000 microphone and a Behringer UB802 mixer as a preamp and to provide phantom power. This combination is pretty common among hobbyist speaker builders for having a reliably flat response curve. I'd obviously get better results with calibration.

The laptop is a Dell Latitude. Cabling was pretty run of the mill mic, 1/4 inch and 1/8 aux connectors.

Software
I used DSSF3. It's a pretty powerful piece of audio analysis software, but, as importantly to me, has an idiot mode for frequency response measurement. Equally important, it's free for 30 days.

Procedure
I ran the mic to the mixer using a balanced cable and set a reasonable output level on the mixer. Then I ran an 1/8 inch cable from the output of the laptop to the Aux in of the stereo and found a comfortable volume combination between the software output setting (basically, the default) and the MINI's volume ****.

I checked the purity of the signal by running a sine wave @ 1KHz and running the built-in oscilliscope of the software to check the waveform. The waveform wasn't perfectly steady, but was round at the top, so at least I wan't clipping.

I placed the mic so the tip was a couple inches in front of my right ear, facing forward. I pinned it between the top of the seat and the headrest (it's a long thin mic so it reaches forward.)

To help simulate the absorptive and reflective properties of a human body, I placed a 185lb dummy in the driver's seat (namely, me.)

Measurement
I ran the built in Frequency Response window of DSSF3, and the only thing I changed from default was to run a 10Hz to 20KHz sweep over 20 seconds. I also changed the displayed levels to 0 to -30dB because it filled the window better.

I clicked "Start" and got a graph I liked. Then I tweaked up the MINI's volume 8 clicks on the **** and did it again. Then I repeated one last time. I tried one more increase of 8 clicks but at that point, it started freaking out on me at certain frequencies, producing clicks and pops.

I repeated each sweep a couple of times to see how reproducible the results were, and was satisfied that I was taking at least something of a valid measurement.

Enough chat, the graphs already
Here they are. There's a noticeable flattening as I increased the volume.

A comfortable, quiet listening level. Assume that the curve below about 15Hz is a result of equipment or measurement limitations and is invalid:



8 clicks louder:


And 8 more clicks



Conclusions
Draw your own. I found it interesting how the shape changed with volume. I allow that my procedure, equipment and general understanding might be horribly flawed. If that turns out to be the case, I'll be happy to withdraw this post (which has now taken me longer to type than it did to conduct the entire test.)

 

Thanks,the graphs show It should sound better at the louder volume.Seems opposite of what you would think.
You also seem to show no funny issues. Maybe you can do the same for just the rear speakers since many claim there is something funny going on back there.

 

Thanks for posting your results.

It would be interesting to see what you get when you turn the front-to-back balance to run only the back speakers.

 

I'll do a faded measurement after my wife stops telling me how weird I am.

I should also measure my home stereo, just to validate that those dips aren't a function of my equipment or procedures.

The dips are about 90, 360, and 720Hz... multiples of 90.

 

Nice writeup. How subjectively loud are the 3 volume levels, and where are they compared to your normal listening level?

The response curve is actually much more flat than I expected. When you measure your home system, I predict that you'll find some crazy peaks and valleys. The car was at least a controlled environment that engineers could work with. At home, the response can vary wildly according to placement of speakers and listener, size/shape of room, furnishings, etc.

 

mslatter,

Nice work. Can you tell us what the bass, treble and 'aux level in' were set at? It might be interesting to play with these to see if you can get a flatter response on the low and high end and whether playing with any of these reduces the notches in the spectrum.

From what you have posted the frequency spectrum looks not too bad. I don't think there is any way to tell what level of distortion there is at any given frequency though. This isn't your problem, just that an amplitude spectrum won't tell you that.

 

Great question,I didnt even think about that.