Room EQ Wizard Pro and miniDSP's New UMIK-X Microphone Array Make Multi-Channel Measurements a Reality

(October 20, 2020). Room EQ Wizard (REW) has launched a new Pro upgrade to its mega-popular REW software package. While still in its early stages, Pro’s first advanced feature is capturing multiple inputs simultaneously for both measurement and real-time analysis. Individual SPL metering is available for the first four inputs.

Measurements made using multiple inputs will deliver data showing an rms average of all captured channel magnitudes, with an option to view each channel individually or a span of individual channels collectively. REW Pro allows users to select which channels are the first and last to be captured; Java drivers on Windows limits channel capture to two channels, while macOS and Linux Java drivers offer as many channels as the interface supports.

The REW Pro Upgrade costs $100, requires REW V5.20 or later, and can be purchase by clicking here. It can also be purchased as part of a $550 miniDSP package that ships with the company’s all-new UMIK-X distributed microphone array. This array includes up to 16 microphone capsules for true real-time multi-channel measurements. Each UMA4 microphone module includes four capsules, and up to four UMA4s can be daisy-chained together with a single twisted cable. Because theater rooms can be quite large, the chain can be made up to 40 meters long and connects to a computer using a USB-A2B interface (16 channels, 24-bit/ 44.1kHz or 48kHz). And, of course, it’s fully compatible with REW Pro.

The UMIK-X kit includes the USB interface, four UMA4 microphone arrays, adjustable 1.2m stands, and a REW Pro license. You can purchase it by clicking here.

 

[Kal Rubinson]

Ideally, one might have a UMIK at microphone 1 in the center of the UMIK-X arrays. Apparently, ASIO4ALL will allow that with REW but I have not tried it.

EDIT: I think that I was not clear on what I was suggesting.
1. A single mic at the MLP.
2. Multiple 4-mic UMIK-X arrays (16 mics) on either side as well as above and below the MLP.

 

[Sonnie]

I know there are plenty who advocate moving the mic this much, but I've never been fond of the results for me personally. I prefer 3 measurements at ear level, center, 3" left, and 3" right... and on numerous occasions I have never moved the mic with excellent results. We done countless experiments and measurements, and come back to single mic position or the center + 3" placements during our evaluations and in the reviews conducted in my room. However, for someone trying to get a bunch of seats sounding better in a theater room, maybe this will benefit them.

I'll probably still try it anyway at some point down the line.

 

[AJ Soundfield]

1. A single mic at the MLP.
2. Multiple 4-mic UMIK-X arrays (16 mics) on either side as well as above and below the MLP.

I had perhaps incorrectly presumed this system as a way to (manually, rather than AutoEQ) do spatial averaging more efficiently, i.e. a 4 mic array (single) that was around the size of an average human head. This would allow for not over/under EQing bass peaks based on a single spatial position, something I think wise. Additional 4 mic "head" arrays could be used for other seating positions, again, to get more averaging, for the same.
No sure what you're suggesting above. Is this for presumed head movement at main LP, or just a wider spatial range of averaging for main LP, or...??

cheers

 

[AJ Soundfield]

I prefer 3 measurements at ear level, center, 3" left, and 3" right... and on numerous occasions I have never moved the mic with excellent results.

Measured, or aural results? Both?

cheers

 

[DanDan]

Placing one mic at nose centre works out well IMO. Two of the others are at ear positions and the one behind gives some 'depth' to the sweet spot, representing movement back and forward. Unfortunately I can see height being omitted.
Dirac are very firm about the need to include height variations in the sampling.

 

[Sonnie]

I've spoken with Dirac about adding the one mic non-movement (or a few inches left and right) to their instructions as an option that users can experiment with. There exist a sufficient number of users who have had good results not moving the mic beyond ear level, an to this day they still argue their results are better than moving the mic up/down... doing what I call the hokey-pokey measurements. There have been quite a few professional reviews where the mic was not moved beyond a 6" area and only 3-4 measurements taken, and the results were excellent. I've experimented and compared them until I thought Dirac might shut down my account because of excessive use (don't really think they would), but every time I come back to single mic placement sounding better for my two-channel listening.

 

[DanDan]

Unlike many inventors, Mathias of Dirac is very conversant with using the system. He is one of the few who completely concurs with my observation that a small delay to compensate for imperfect speaker distance/centre can be disastrous sonically. There are many who subscribe to the theory that the smaller the zone the better the 'correction', and see no point in sampling height variations. But Mathias/Dirac does recommend height variations. Given that Dirac is the best in class, I would simply follow his advice. But he does make a logical point that to make a spatially robust correction, you must sample the space. A friend over at NorthSlutz had a good go at testing the sampling procedures. https://www.gearslutz.com/board/stu...tware-dsp-optimising-results-small-rooms.html

 

[Sonnie]

I've followed his advise (tested it countless times), but I get better sound with a different method, and I'll have to stick with the different method. I can assure you, I am not the only one that does it and gets better results. I'm not going to argue with anyone about what they prefer, if it sounds better to them. The bottom line is do what sounds best to you, and not simply because someone thinks their process is better. Our ears have to be the final judgment. There is a reason people use Audiolense, Accuratesound and RoomShaper to get what they believe are superior results to Dirac Live, and in most cases they never move the mic even an inch for their measurements. People like different things, and we get to a neutral sound and tune it to what we like in different ways. I'm happy for anyone who can get to what they want, and it matters not to me how they get there. Don't try to hem folks in on one method... give them options to try and let them decide for themselves. Just because "John Doe" says it, doesn't mean it fits everyone.

 

[Kal Rubinson]

No sure what you're suggesting above. Is this for presumed head movement at main LP, or just a wider spatial range of averaging for main LP, or...??

Since the individual mics in each array are individual input channels, I see this as a way to collect discrete data from multiple microphone positions with a single sweep rather than with tedious and unrepeatable multiple sweeps/positions with a single mic.
Imagine one Umik-1 in the middle with a Umik-X 4-mic array above it and two more Umik-X 4-mic arrays flanking it or one Umik-1 in the middle with a pair of Umik-X 4-mic arrays on each side for a wider window.

 

[DanDan]

To be fair Mathias is the creator of Dirac Live, hardly a random 'John Doe'. The algorithms applied in Dirac are his. Essentially he is deciding what to Eq and how much. If we trust him in that, why not trust him in his advise on how to use his invention?

"I tend to take measurements spread out about 1 m side-to-side and 0.5 m up-down and 0.5 m front-rear, and I find that this is a general rule of thumb that works well for a single chair listening region. Five of these are at earheight plus minus 10 cm, and then the 4 remaining ones are spread out a bit further. But this is a rule of thumb. What’s important is that you spread them out as ”randomly" as possible (trying to capture the variability in the room response in your listening region plus some). The software does not know which measurement is which except for the first measurement which is the basis for the delay estimate, so in which sequence you take measurements 2-9 does not matter. And if you are unsure, I would just add a couple of more measurements. More measurements give more information, and the results will only improve. This is the essence of spatial robustness. You need to capture as much as possible of the room variations with your measurements, otherwise the algorithm has to guess what goes on in between the measurement points. Spatial robustness is all about making sure not to overcompensate or to compensate for colorations that are there in one measurement, but not in the vicinity. The first thing to understand is that a perfect inversion of the room response is impossible, as the impulse response and frequency response varies so much even in a small region. You have to make a correction which improves as much as your data and knowledge of acoustics tells you is possible without compromising results in any important position. This is simply impossible with a single-point measurement, because the laws of physics are cruel and won’t let you guess from one single data point what happens in other points. "

Just for clarity..... "they never move the mic even an inch for their measurements" Is this literal or are you referring to no variation of height?

 

[AJ Soundfield]

I see this as a way to collect discrete data from multiple microphone positions with a single sweep rather than with tedious and unrepeatable multiple sweeps/positions with a single mic.

Right...but over a much wider area of sampling than provide at a single "head" 4 mic array. This is for the main LP?
This much bigger average has been shown to....

Btw, what does Dirac et al have anything to do with this REW/Minidsp system????
The thread article and this https://www.minidsp.com/applications/subwoofer-tuning/umik-x-mso created my initial interest, but I must be missing something...

cheers

 

[Sonnie]

To be fair Mathias is the creator of Dirac Live, not 'John Doe'. Just for clarity..... "they never move the mic even an inch for their measurements" Is this literal or are you referring to no variation of height?

Yes... I am familiar with Mathias... but I was not directing my statement only to Mathias, but to anyone that might "advise" someone that there is only "one method" of achieving good results.

Read... there are literally "countless" Dirac users that do not move their mic from dead center ear level in any direction... and users that only move it left or right a few inches. These users generally sit in the same position for listening each time they listen, and/or have heard no improvements, and/or get inferior results when moving the mic. This is absolutely not new news... this has been going on since Dirac surfaced many years ago.

As far as other systems, when Mitch at Accuratesound asks for a measurement for setting up his services, which include uber hi-end systems, he only wants two measurements, left and right speaker dead center of ear level. When Thierry asks for measurements for his approach at Home Audio Fidelity, he only wants dead center, a few inches/cm left of center and right of center for left and right speakers. Granted, their systems do a little more than Dirac with some of their level of services, but there are numerous happy users who prefer their methods over Dirac Live. While I love Dirac and use it frequently, there are other avenues of achieving good sound (some claim better). I will likely end up using a different method for two-channel music listening and then use Dirac for movies/Atmos (multi-channel), as I am more critical of music than I am movies.

 

[Sonnie]

Btw, what does Dirac et al have anything to do with this REW/Minidsp system????
The thread article and this https://www.minidsp.com/applications/subwoofer-tuning/umik-x-mso created my initial interest, but I must be missing something...

cheers

miniDSP offers Dirac Live in several of their units, so miniDSP will likely refer to Dirac as the DSP system of choice, but the UMIK-X could be used with any DSP system, even their own system without Dirac Live.

 

[AJ Soundfield]

miniDSP offers Dirac Live in several of their units, so miniDSP will likely refer to Dirac as the DSP system of choice, but the UMIK-X could be used with any DSP system, even their own system without Dirac Live.

Yes I'm, aware MiniDSP also uses Dirac software, just saw nothing in the article or their page about this UMIK-X system having anything to do with Dirac et el. Whether it will be incorporated in future systems we'll see. Btw, I think you mean so called "Room Correction", not "DSP" which is simply any form of digital signal processing.
Yes, my intent would be to use it for what it is, good old fashioned.."EQ". Kinda like me. Albeit a more efficient way of spatial averaging.

cheers

 

[Sonnie]

I believe Dirac came up because of their recommendation to use spatial measurements, and I suppose it could be used with Dirac, as it is looking for a measurement, which this would provide.

And yes... I suppose some may use something other than DSP for room correction... and they might not even be "old fashioned".

 

[AJ Soundfield]

While I love Dirac and use it frequently, there are other avenues of achieving good sound (some claim better). I will likely end up using a different method for two-channel music listening and then use Dirac for movies/Atmos (multi-channel), as I am more critical of music than I am movies.

Good man, especially if your ears prefer C) over (the eyes have it) B). Not a "claim" either. If I had a dollar for every of those...
Btw, not to be outdone by lawyers, my consultation start at $900/hr. Toll free though, so call any time.

cheers

 

[Sonnie]

Good man, especially if your ears prefer C) over (the eyes have it) B). Not a "claim" either. If I had a dollar for every of those...
Btw, not to be outdone by lawyers, my consultation start at $900/hr. Toll free though, so call any time.

cheers

I think you left the decimal point out... $9.00/hr ... calling now!

 

[DanDan]

"Btw, what does Dirac et al have anything to do with this REW/Minidsp system????"

DL and Mini work hand in glove. DL favour multipoint measurement. This new system suggests that JohnPM and Minidsp also do. This new Mic system can do four at a time. I don't know if those dots can be joined or not. I kinda doubt it if each mic has an independent clock. I started out with FuzzMeasure, as it was the only Mac system. I used to recommend measuring at ear positions in a pseudo binaural process. But multiple ear positions, a pair for each 'head location' As it happens FM and the Mac OS can easily record multiple mics simultaneously. Toole and others use Dummy Heads and HATS systems. However, on consideration, unless the software were to sum my multiple pairs individually, including a HRTF, there seems little to recommend my method except the 'philosophical' fact that we did measure representative actual ear locations. DL and Sonarworks and others don't seem to care which sample came from where or when, except for the first centre one. But Sonar requires, can't remember 29 or 39? Dirac requires 9 and it used to be possible to use more if you wished by simply continuing. Not sure about current version.
@Sonnie, interesting that other systems use single point one two or three measurement. Thanks for the heads up. Presumably they cannot accept multiples, so there is no issue of not following the designers recommended use there. I hear you that many are using Dirac in ways other than intended, it brings to mind that many of us chose to turn the ADAM S3A vertical. This was later added to the Manual. I think it worth noting that Skols testing showed best results using a smaller than recommended zone, but not the single point. This was a highly treated Control Room though.

 

[AJ Soundfield]

Near certain this system is intended for multi-subs/LF.

The other systems Sonnie mentions methinks is 2ch "room correction", possibly full bandwidth from LP. Ouch.

 

[Sonnie]

I wouldn't necessarily say that John or miniDSP favor one measurement style over another, but they are offering (for a cost/fee) an easier method to accomplish spatial measurements. Quite worth it if you prefer spatial measurements. And as already stated, I'll certainly give it a try, as I am constantly testing and evaluating the alternative to what I prefer, and what I prefer changes from time to time. I try to keep an open mind about other methods, and respect those that choose to use other methods.

Yes... these services to offer multi-seat services, and one prefers multiple measurements, but with small mic movement. They can do the full Atmos corrections. However, I'm just not critical of movies like I am music, being that I am usually the only one in the room and not worried about the other seating locations.

 

[Sonnie]

Near certain this system is intended for multi-subs/LF.

The other systems Sonnie mentions methinks is 2ch "room correction", possibly full bandwidth from LP. Ouch.

These other services can do multiple subs, multiple speakers, time alignment unlimited, etc. https://accuratesound.ca/

Juice HiFi

 

[DanDan]

LOL, I just skimmed through a Sonarworks tutorial video. The tester hand held the mic for the close to speaker measurements. For the rest of the measurements his head was only a few inches from the mic.
Such proximity of the body to a mic can lead to errors of "up to 6dB around 400Hz", according to Bruel and Kjaer........

"Btw, I think you mean so called "Room Correction", not "DSP" which is simply any form of digital signal processing."
I would never intentionally use the term Room Correction. Generically I call all of these systems 'Speaker Controllers' Kinda tempted by a Meyer CP10 right now......

 

[Kal Rubinson]

Btw, what does Dirac et al have anything to do with this REW/Minidsp system????

Just another possible application of the new hardware.

 

[Nordo]

I know this is a bit off topic, but when placing my single mic at the centre of the MLP, I very carefully work out exactly where my ears would be when my head is resting comfortably into the headrest of the soft cushioned seat.
However I've now noticed that when watching a movie, I don't rest my head in the headrest, but it's located a good 2" off the cushion.
And because of the soft head cushion it means my ears are almost 4" from where the Mic was.
I'll allow for this when I do my next calibration.
Not sure if I will notice any difference.
Also I don't know yet where my head is located when I'm listening to music.

I thought I would mention this, as it may be a common mistake many people are making.
Perhaps this would be better off in the REW forum?

 

[DanDan]

I recommend try focussing your imaginary equilateral triangle a little distance behind your head/pillow. Obviously I would also recommend sampling at least at the two actual ear and centre locations, although not with you in the seat.
I hope the pillow is acoustically absorbent, not e.g. leather! http://arqen.com/acoustics-101/room-setup-speaker-placement/