• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Measuring with a stereo microphone?

ppataki

Major Contributor
Joined
Aug 7, 2019
Messages
1,216
Likes
1,355
Location
Budapest
This is just to collect some thoughts on this subject....

I personally have been using a UMIK-1 measurement microphone to measure and optimize my systems (I guess I am totally not alone with this here :))
I use REW, Dirac or whatever comes in my way

I was wondering if it would make any sense instead to use a stereo microphone (not sure if there is any software adequately supporting it) for measurements (like a stereo in-ear microphone that is actually where our ears are.....)
Then potentially we could do a so-called 'true stereo' convolution (=left ear left/right channel + right ear left/right channel)
Also the two mics might pick up phase issues that a mono mic might not
I am just thinking out loud here....I might be totally wrong

Any thoughts/experience/comments would be very welcome
Thank you
 

GXAlan

Major Contributor
Forum Donor
Joined
Jan 15, 2020
Messages
3,866
Likes
5,953
There is a concept of a moving microphone for averaging. That last generation of Sony ES AVRs used stereo mics that were about 5 inches apart. A stereo mic with reasonable distance could in fact help you collect twice as much data in a shorter period of time.

A lot of audio science is based around mono testing (for good reason) but as a consumer, stereo music can make mediocre speakers sound better and as far as I know, there aren’t any preference scores going down when you listen in stereo compared to mono. That means for whatever budget or performance target you are shooting for, you always get a few bonus points once you listen in stereo.

The challenge is the 3-4kHz bump described by Dr. Toole and nicely
summarized by @Thomas_A.

“In the stereo triangle you will experience a dip around 2 kHz followed by a peak 3-4 kHz when using a speaker with flat frequency response. This effect is exaggerated in nearfield (and anechoic) conditions. The dip at 2 kHz can somewhat be filled with reflections from the room and the peak at 3-4 kHz can be damped with 1-2 dB dip.”
 

Jdunk54nl

Addicted to Fun and Learning
Joined
Aug 5, 2020
Messages
968
Likes
1,048
Location
Arizona
Yes, you could use stereo and a bunch of software supports multi mic input including REW pro version. Minidsp even makes a multi microphone array setup. This gets you a spatial average without having to move microphones around.

I like odd number of microphones when doing multi mic stuff. The odd one goes in the center of the MLP for phase information. This has been being done in car audio for years. About time home audio people start to catch up on the benefits of multi mic measurements rigs ;)

I also use SMAART and open sound meter (free version of smaart) that can get phase in real time.
 
OP
P

ppataki

Major Contributor
Joined
Aug 7, 2019
Messages
1,216
Likes
1,355
Location
Budapest
Yes, you could use stereo and a bunch of software supports multi mic input including REW pro version. Minidsp even makes a multi microphone array setup. This gets you a spatial average without having to move microphones around.

I like odd number of microphones when doing multi mic stuff. The odd one goes in the center of the MLP for phase information. This has been being done in car audio for years. About time home audio people start to catch up on the benefits of multi mic measurements rigs ;)

I also use SMAART and open sound meter (free version of smaart) that can get phase in real time.
Which multi mic are you using exactly?
Does SMAART and Open Sound Meter also support that apart from REW Pro?
What do you do with the measurements then? Do you do vector averaging or some other methods?
Have you tried producing a true-stereo convolution maybe?
Is multi mic giving any benefits vs just performing multiple measurements with a mono mic? (maybe phase info or anything)
Sorry for all these questions, just trying to understand if it would be worth upgrading....;)
Thank you
 
OP
P

ppataki

Major Contributor
Joined
Aug 7, 2019
Messages
1,216
Likes
1,355
Location
Budapest
That last generation of Sony ES AVRs used stereo mics that were about 5 inches apart.
Would be awesome to see some before/after measurements about that.... (not just FR but the other graphs too)
I guess this is wishful thinking from my side to be realistic :)
 

Matthias McCready

Active Member
Joined
Jul 2, 2021
Messages
209
Likes
273
A few thoughts:

1. Measurement mics are omni directional; something that our ears are not.

2. You can use multiple mics for measurement. Capsules will cause variations, so purchasing a mic with a calibration file is helpful. Please note that a capsules' frequency response will change over time, this is true even of expensive measurement mics that cost thousands. However with a functioning mic this is not a huge deal as the frequency variation which occurs in a room is much more significant. If you take a measurement, and then move that same mic over an inch, this usually causes a greater variance than between most functional measurement mics model/capsule differences. To the point your two ears never hear the same thing, and your brain is quite good at compensating for this.

3. If you are doing something such as time aligning you want a single measurement.

4. If you are doing general measurements to see what a loudspeaker is doing frequency wise in a given area than more measurement will give you a more accurate picture of what is going on. In a Rational Acoustics training I attended (the makers of Smaart) Jamie Anderson recounted that there is a diminishing return to multiple measurements. They performed a test in a venue where they collected 60 measurements in each area, and then averaged them. They kept dwindling down how measurements were used. 30 down to 20 and so on. The result of the testing was that: Two measurements is far superior to one, and that three measurements is better than two. Roughly though this is where the point of diminishing returns hit hard; the data given by three measurements and sixty were not significantly different.

5. Multi mic setups are all about speed, and potentially at the expense of accuracy, due to capsule variations. In live sound sometimes you get 20 minutes to do your measurements and tuning. My personal measurement rig is 3 mics, on a folding 4ft Manfrotto stand, with Triad & Orbit clips with a Cat Rat breakout and a 150ft tactical Ethercon (all 3 mics analog over 1 cable). It works great for large venues and moving quick; but I wouldn't bother with more than a single mic for home; where I presume you value accuracy over speed.

6. Measurements are not pretty to look at. Real-world "flat" is a far cry from what one would hope. A common response is to try and EQ every peak/dip. Jamie Anderson refers to this as "measurebation," and it does tend to sound a lot like comb filtering. What is prescriptive and corrective EQ for a dip in one area, might be a null an inch or a foot over. What is corrective in place is destructive in another; which is a general principle in audio, everything is a tradeoff. This is why averaging many different positions is important. In the live sound land (different from at home) the most skilled tuners tend to do 3-4 wide-Q filters at most. It is also important to realize that EQ is a hammer, it is not the correct tool to fix everything; it cannot fix room problems. I still do think it is quite useful, I do use Sonarworks on both of my main systems, but it is not a replacement for room treatment either.

Hopefully that helps you out :)
 
OP
P

ppataki

Major Contributor
Joined
Aug 7, 2019
Messages
1,216
Likes
1,355
Location
Budapest
I wonder if anybody here tried the true stereo convolution for room correction...
I am going to give it a try during the weekend; will report back my findings
 

DVDdoug

Major Contributor
Joined
May 27, 2021
Messages
2,916
Likes
3,831
You'll get differences in with a stereo mic but I doubt the extra information will be that helpful... If you play a high frequency tone, say 5-10kHz, and move your head a bit you'll hear huge variations as the direct & reflected waves combine in and out of phase (and a mic will measure/record the same thing). It happens with music too but it's natural and we're used to it so we don't really notice it. You can't really fix those wiggles in the frequency response because they are different if you move a couple of inches in another direction.

At lower frequencies (longer wavelengths) small location changes don't make any difference so a stereo mic will pick-up the same thing in the left & right channels. But two or more measurements with more distance and at different places in the room can be useful.

Also the two mics might pick up phase issues that a mono mic might not
We are mostly concerned with phase and wave cancelation of the acoustic waves and as-above it's not that useful at high frequencies.

Our ear/brain doesn't really hear phase difference between the left & right ears. The distance between our ears is about a half-wavelength at 1kHz so if you turn your head, pointing one ear toward the speaker, your left & right ears are 180 degrees out-of-phase but the sound isn't canceled in your brain. Or if you flip the polarity of one speaker you'll get "phase weirdness" and bass cancelation, but with headphones where the waves don't mix you won't hear a difference.

P.S.
I'm not sure if you can find a calibrated stereo measurement mic...
 
Last edited:

Jdunk54nl

Addicted to Fun and Learning
Joined
Aug 5, 2020
Messages
968
Likes
1,048
Location
Arizona
I wonder if anybody here tried the true stereo convolution for room correction...
I am going to give it a try during the weekend; will report back my findings
Two microphones will just give you a slight spatial average. I like to do 5 or 7 microphones if I am doing a head space. Otherwise it really isn't worth it compared to just moving a single microphone.

I like one in the center, one to the left, one to the right, one above, one below. Or setting up microphones in all listening positions and averaging all listening positions at once.

As was stated, the only real benefit is time-saving and your body is no longer impacting the measurements.

To the other questions:

Which multi mic are you using exactly? - I use a motu m4 or focusrite 18i20 depending on what I need to do. Then usually just dayton emm-6's since they are fairly cheap (although I am actually looking at cheaper xlr measurement microphone options since when they start to go bad, a replacement can add up). Sometimes when my partner is helping we will use his earthworks microphones, but not a huge difference between those and the emm-6's. My partner is MUCH MUCH smarter in this stuff, he has a degree in acoustics and understands this stuff at a level very few do (like all the math actually behind all of htis..

Does SMAART and Open Sound Meter also support that apart from REW Pro? Yes, they support multi-microphone averaging in real-time.

What do you do with the measurements then? Do you do vector averaging or some other methods? EQ and phase adjust. Phase is with a single MLP microphone but sometimes I will adjust phase based on different listening positions to get the best for multiple positions, much like Dirac and multi sub optimizer does, although they are better because it is a "brute force" and they can check multiple options much quicker than I can.

Have you tried producing a true-stereo convolution maybe? Sure, I have done two microphones, I don't see real benefit in this over moving microphone measurements and prefer moving a single mic compared to just using 2.

Is multi mic giving any benefits vs just performing multiple measurements with a mono mic? (maybe phase info or anything) - Time saving and repeatability. Since the microphones are constantly in the same exact spot, your measurements are easily repeatable.

Sorry for all these questions, just trying to understand if it would be worth upgrading....;)
Thank you
 

No. 5

Active Member
Joined
Nov 1, 2019
Messages
144
Likes
121
I was wondering if it would make any sense instead to use a stereo microphone (not sure if there is any software adequately supporting it) for measurements (like a stereo in-ear microphone that is actually where our ears are.....)
Then potentially we could do a so-called 'true stereo' convolution (=left ear left/right channel + right ear left/right channel)
I have long thought that a binaural or a sound field measurement is the next step forward in room analysis for the enthusiast. After all, our ability to “spacilize” what we hear is a massive part of enjoyment and preference in audio, and is one of the standout things that can make a room or setup good or bad. And all that happens (mostly) with level and time differences between the two ears. In fact, one metric of spaciousness is to measure how different the sound field is at each ear. All of that is lost on a single microphone channel.

Of course, all of that supposes that there’s freely available software that can do the analysis, or presents the data in a way that the user can draw conclusions.
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,521
Likes
37,050
A long involved read, but mitchco who is also a member posting here, has done some excellent work with room correction software, measurements and binaural measuring microphones.

 

Midnight Audiophile

Active Member
Joined
Jun 30, 2022
Messages
143
Likes
226
A long involved read, but mitchco who is also a member posting here, has done some excellent work with room correction software, measurements and binaural measuring microphones.

Thanks for that! I've put together a binaural recording head to do speaker comparisons and it's exactly the type of info I was looking for. Am I correct on speaker level matching that it will be best to go with pink noise and try to match the Db as close as possible? I'm making sure speakers and mics will be in the same position (and remove the unused speakers from the room).
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,521
Likes
37,050
Thanks for that! I've put together a binaural recording head to do speaker comparisons and it's exactly the type of info I was looking for. Am I correct on speaker level matching that it will be best to go with pink noise and try to match the Db as close as possible? I'm making sure speakers and mics will be in the same position (and remove the unused speakers from the room).
Yes, something like that. I think Harman uses pink noise, but filtered between 500-2000 hz for initial matching of level. 12 db/octave roll off below 500 hz and above 2000 hz. You can make such a signal in Audacity easily enough.
 

Midnight Audiophile

Active Member
Joined
Jun 30, 2022
Messages
143
Likes
226
Yes, something like that. I think Harman uses pink noise, but filtered between 500-2000 hz for initial matching of level. 12 db/octave roll off below 500 hz and above 2000 hz. You can make such a signal in Audacity easily enough.
It looks like that's the option I've been using in REW when I set my system's Db level (Pink random, Speaker Cal 500 to 2000Hz).
Screen Shot 2022-08-06 at 9.37.30 PM.png
 
Top Bottom