Multi Mic Measurment Method

[1audiohack]

From a converstaion started in another thread.

Hi Thomas;

In my opinion a 12" midbass-midrange driver is one of the strengths and one of the weaknesses of the 4350. Dynamics are stellar, integration is difficult.

I tried the 1200Fe in attempt to lower the 15's into near sub territory but the dogbox for the 12 pretty much drives the results from about 200Hz and lower. The 1200's sounded nice so I just left them. Different EQ requirements of course. As a note, it looks to me like the 1200Fe and the 2020 use the same basic motor.

Several things stacked up to require changes to the 4350's. I excanged the 1501-1AL's back to the 1500AL's, inadvertantly locked myself out of the dbx4800, and was still never happy with the implementation of the 12's and it seemed a shame to have the TAD's just sitting in boxes so I hard reset the dbx and got started.

With the old (haha) 1500AL's the VLF required a little EQ work, the TAD's? None, none at all. A little out of band filter work on the top and they look like an electrical filter measurement.

The results are the best ever for the 4350's with a couple of audio buddies exlaiming that the 4350 system is my best system of all.

I also I used a different measurment method. I bought a new SoundCraft EMP6 six channel mixer. All analog, no effects. The sound and build quality belies its price. I have five precision measurment mics and typically set up one on axis and one about 45 degrees off axis but even outdoors the reflections, even just ground bounce requires interpretation and smoothing

TEF is a four port two channel measurment system so with the on axis mic in ch1 and the delay fixed there, the other three mics used in this session (up to six mics) can be anywhere in the sound field and with individual mic input level setting ability before being summed and sent to ch2 of TEF, I can weight the mics however I want and average the send to ch2.

This method made reigning in the 4350's so much easier and faster than any other measurment session I have ever conducted that I feel I should ducument it in a thread of it's own.

The board was less than any new precision measurment mic and I have high expectation that it will be even more valuable indoors.

Side benefit of a board with two additional stereo inputs used with TEF, I listened to music through the 4350's the entire time I tested them.

I will get back to this with some documentation in the near future. For now suffice to say this is the method I will use going forward. The results were phenominal as were the time savings.

My best to all,
Barry.

[srm51555]

I will get back to this with some documentation in the near future. For now suffice to say this is the method I will use going forward. The results were phenominal as were the time savings.

Thanks for sharing!

[pos]

From a converstaion started in another thread.

Here is the aforementioned thread



Hi Barry

Nice method, somewhat similar to what trinnov does (but they use a tight array IIRC).
Have you tried the other MMM method, where the first M stands "moving" instead of "multiple" ?

One drawback I see with the multiple mic approach is that each mic will have a different path length to the source, so time information will not be usable (just like with the moving mic method), and also the summed response will include comb filtering artifacts from those delay differences.
You also cannot use individual calibration files, so you need good mics to start with, or at least a set of mic with very similar response curves that can be corrected at once.

The ability to weight each mic on the fly is really nice, and with real time analysis like what smaart does it must be really convenient and fast! Does TEF let you do real time analysis?
I'd be very interested in reading more about that approach, including photos!

Regarding your 4350, why did you replace your 1501AL-1 with those old school 1500AL ?
And have you tried the 2020H in there?
I bet you have, and I am curious to read your impression and comparisons with the TM1201.
(Oh, and did I mention you ought to try current drive on that driver? )

[1audiohack]

HI Thomas;

I haven't set up anything moving, speaker or mic. TEF will automate for polar responses and such but I haven't delved into that, yet.

Losing the time information on the channel two mics while regretalble is not the end of the world for me. I use TEF channel input number one for the one axis mic and set the filter delay there. Any other mics are summed on the board and the time information there is ignored. This preserves the time information for alignment.

I know you well understand the issues of combfiltering but let me get this out for context:

As far as the comb filtering issue, I tried the if you can't beat them, join them theory. For a single enclosure multi way system I tylicaly set up at ≈9M or 10' with my on axis mic looking into the HF horn (not the UHF). Without taking serious precautions like a pile of fiberglass in between the mic and DUT, even outdoors there will be a calcuable ground bounce notch and the resultant comb filtering. This makes the integration of cross over and EQ difficult, even with much smoothing. With another mic on the ground at the same distance and the two averaged the picture is much clearer. Another mic off axis at about the same distance at a different height and the picture is clearer still.

Using the primary mic direct to the measurment aparatus and the others wighted and summed to channel two of the aparatus with both traces simultainiously on the screen made the integration process a breeze.

The mics I use, four are a weel matched set and the other two are a matched pair. You are right though, lesser mics would be an issue.

I have SMAART V8 and the RTA is spectacular. RTA in TEF is rudimentary but usefull indeed.

I will get some pictures and screen shots mext go, hopefully this weekend.

The 1501-1Al's will go in the Everest Cabinettes if I ever get the horn parts done.

2020? Yup, I liked it a lot. Like the 2012 and 2123 I love the attack and sound of them. A good snare hit will crack your teeth. In the end I like the 2020 best horn loaded, that's where they really shine.

The TM1201 definitely has a flavor I have not heard in a 12" driver. It might be the most neutral 12 I have heard to date. The more I listen to it the more I like it. It takes me a while to really get to know them however.

I have yet to experiment with the current drive option. I should get going and try it.

Off topic question for you, have you done anything with LIR filters? I borrowed a Linea Research amp and just got into it before I had to give it back. Guess I will have to save for one.

All the best,
Barry.

[pos]

Off topic question for you, have you done anything with LIR filters? I borrowed a Linea Research amp and just got into it before I had to give it back. Guess I will have to save for one.

Hi Barry

I have no experience with LIR.
With today DSP power (from BSS, miniDSP, and others) I'd stick with time-domain FIR: that let you do everything, including correcting phase shifts from the driver's natural rolloffs (what is the point of having linear-phase electrical filters if the resulting acoustical ones are not?), and with no additional delay over the one implied by the FIR itself (which can go down to 0 if a minimum-phase response is the target).
In the end everything is in the FIR design process, and it is entirely predictable and reliable.