Find also your MMM discrepancy weird though as it usually matched the averages of individual sweeps much better, could you maybe post a screenshot of your REW RTA settings?
Usually only the case below about 300-400Hz.
Find also your MMM discrepancy weird though as it usually matched the averages of individual sweeps much better, could you maybe post a screenshot of your REW RTA settings?
Side note: Surround mode is horrible. It sounds like a poorly executed phase based matrix up mixer. It further screws up the tonality, and feels like you're in a vague soup of boomy sound.
What are your thoughts and experiences on measuring the response of car audio systems? I've seen some papers outline a system of measurement using KEMAR heads / HRTFs. Unfortunately I do not have access to such equipment.
I am wondering what causes the discrepancy between the moving mic averages and the averaged sine sweeps. Is it correct to assume that the MMM is more representative of the overall response that I hear?
If you were to measure the response of a car system, what would you do?
Hey y'all. I recently completed a large set of measurements of the Bose sound system in the Porsche Macan. What are your thoughts and experiences on measuring the response of car audio systems?
In a "premium" preinstalled system there will be dynamic EQ, the amp will receive CAN signals with the speed or RPMs and it will increase a bass shelf while going fast and flatten it up while still/low speeds. In newer systems they even use a mic to get some feedback to the amp and adjust the EQ.I have played with car audio a fair amount. The biggest problem I have had is the noise floor of 70db. Too much bass at stops and low speeds, too much at higher, fatigue from higher listening levels to drown out the background.
The hard part; the part that will never be close to home audio is the other stuff. Imaging, focus, width, depth
Which measurements do you use to diagnose and correct problems in each of these areas, in your experience?
Perhaps, but it all adds up — especially with Porsche, where essential options (that almost everyone will want) often add 20-30% to the listed MSRP price. Sure, you can get a Porsche for $X MSRP, but if you want it to come installed with seats and a steering wheel, it will cost you extra (JK, but this feels not far from the truth).Macan is a utility thing. How much can you stuff in to it? Panamera is a high performance a luxury sedan. So audio is probably important. If you can afford Panamera, you can afford the Burmester upgrade.
And @hardisj what in your experience are the optimal number of speakers and locations in a typical, say, sedan.
(Mostly interested in your personal opinion/experience here )
To their credit, Porsche does offer 'badge deletion' as a free option. But I'm guessing that most owners want most non-owners to know that they are behind a GT2.Perhaps, but it all adds up — especially with Porsche, where essential options (that almost everyone will want) often add 20-30% to the listed MSRP price.
Nice work! First review of its kind that I've seen
Maybe the surround mode would sound better? Does it include upmixing? Car systems have to deal with a lot of crosstalk and comb filtering.
Not sure what accounts for the MMM results other than your sweep pattern.
@hardisj Do you have an opinion?
Get a loud exhaust, open the windows and forget the sound system.
I wonder, is it possible to actually play "true" 5.1 mch tracks here instead of just upmixing stereo music to "surround"?
Perhaps, but it all adds up — especially with Porsche, where essential options (that almost everyone will want) often add 20-30% to the listed MSRP price. Sure, you can get a Porsche for $X MSRP, but if you want it to come installed with seats and a steering wheel, it will cost you extra (JK, but this feels not far from the truth).
So the ~$8k Porsche charges for the Burmester means an extra 5-10% added to the price (even for the Panamera/Taycan), which is not trivial.
For example, in my area I very rarely see any models at the dealer with the Burmester option. As it turns out, audiophiles are rare, and many people would rather pay more for a prestige badge or aesthetic package (e.g. nonfunctional carbon fiber accents etc), than an upgraded sound system. So that’s what I see often reflected in the preconfigured models arriving at dealers.
Similarly, while I don’t see base models show up, I also don’t really see many fully optioned models show up around here. There’s always some package missing, whether it be Burmester or the extra self-driving features packages, or the thermal night vision packages. This applies to even the most expensive Porsche models, so I think this discounts the hypothesis that anyone buying a Porsche wants it fully optioned by default (even though as an audiophile, I obviously agree with you that it makes no sense to buy e.g. a Panamera or Taycan without the audio upgrade package).
Thanks for the insight. For my MMM, I used Pink Periodic 64k for stimulus. RTA settings were: infinite average, Rectangular window with 50% overlap, spectrum w/o smoothing.HRTFs are mostly important for spatial aspects. Especially in those systems featuring surround-sound type decoding. I saw a study that some grad students did using a Dyn Audio system in a Volvo that was pretty interesting but don't remember much about it other than they focused on ITD and IACC.
This gets back to the point of measuring with you in the seat vs not in the seat. I have used binuaral mics with me wearing them and sitting in the MLP. But that's only helped me identify issues that I had no chance of fixing. Such as side window reflections causing a peak or null. IMHO, for the average hobbyist, there is zero point in using anything other than the standard MMM for car audio.
What is the stimulus? And how are you measuring?
Sweeps use a different stimulus than RTA. The former uses a log sweep; the latter should be using pink noise (or pseudo pink noise). Also, make sure that if you are using pink noise that you are measuring the sample sample rate as you are outputting. Check the MLS (or FFT length). REW does this automatically when you use it to generate the signal, but if you are using a test disc with pink noise, you'll need to get the FFT length and make sure you tell REW to measure at this same rate.
For a beginner, this is what I recommend:
1) Have REW generate Pseudo Pink Noise ("Pseudo PN", is what they call it in the generator tab, IIRC).
2) Set MLS or FFT length to 64k (good overlap between enough averaging for bass and treble).
3) Set "Averages" to 32. That's a decent number.
4) Use the MMM. Move the mic around the headspace area of the MLP.
*This is the tricky part; if you use a house curve then sit in the back seat and wave the mic around. If you don't, then it might be easier to sit in the driver's seat and measure with the mic up. You will have to figure out what to adjust based on what you hear because you won't have a house curve to go by. My video I linked earlier discusses this.
5) Analyze the results
Now, you can do the above by individual drive unit, side (left vs right) or combined. Understand that these will all yield different results because of comb filtering effects. Even if you get both left and right sides matched *exactly the same* individually, the combined result will not be the sum; it will be some variation thanks to the cabin. I know it might not make sense. But do it and you'll see.
The left side response you posted makes perfect sense. You have the tell-tale "nearside null" (as I call it) at about 80Hz. Every car has this unless there is something very different about the installation. It's a modal issue you can do absolutely nothing about. It is why I recommend to cross the subwoofer up higher than most prefer because it fills in the null and actually provides "bass up front", contrary to what you would otherwise believe. Many in car audio try to run their midbass down as low at 50Hz to help make the bass more "up front" but that only serves to make the null more noticeable and push the bass toward the subwoofer in the trunk. I also made a video about that linked below.
Nice review! I had a Macan S with the Bose system a few weeks ago as a loaner and your impressions match mine exactly. Exaggerated peaky bass with rolled off subbass, weirdly harsh, glaring, and unnatural lower treble, and a generally scooped sound. I’m curious as to whether you have any measurements with linear mode being on? Would be nice to quantify exactly what it does to the frequency response.
IMO Porsche’s sound systems are pretty underwhelming in general, and one of the brand’s biggest sore spots. Overall, I would describe the sound system in the Macan as acceptable but not particularly high fidelity. It was better than the Bose system in my 911 though. Additionally, I rented a Cayman before and the base system in that was absolutely unacceptable.
On the other hand I find BMW does a pretty good job with both the base US (this is called the Hi-Fi option outside of US) and Harman Kardon options.