If a picture is worth 1000 words - first adventures with REW

[old_bassist]

I ran three tests - one using the Galaxy meter's foam windscreen plus the furry screen from my Tascam handheld recorder, then a test with just the foam windscreen and finally a test with just the furry.

Here is the equipment I used to generate the frequency response curves.

Here is one of my "Half-a-TLAH-Pro" columns/line arrays. Eminence 8s and Goldwood piezo horns, with crossovers done to Bill's exact specs.

[DJPhatman]

No offense, old_bassist, but your results are useless. If you wish to test your cabs, take them into an open field, with no buildings, fences, anything within 50' (~15.25m) of the speaker. Where you have it placed, you are measuring the reflections of the fence, the buildings, the half wall, etc.

Did you check your meters' calibration with REW? Especially using the USB interface, which "colors" the signal. Calibrate the mic and the USB interface before you measure.

[Bill Fitzmaurice]

No offense, old_bassist, but your results are useless.

I wouldn't say that, it shows what he can expect with the speaker in that placement. But since he's measuring not just the speaker but every reflection off every object in the vicinity, including the floor and table, it's not a definitive half-space result.

[ejh2854]

If one were measuring a low-frequency cab- a T39, for example- in an open grassy yard... how far off the ground should it be raised?

[Bill Fitzmaurice]

If one were measuring a low-frequency cab- a T39, for example- in an open grassy yard... how far off the ground should it be raised?

It should not be raised. The mic should be placed about an inch from the ground at least 2 meters away. That will give a ground plane measurement, which is the equivalent of half-space up to the baffle step frequency. The baffle step on a T39 would be no lower than 300Hz, so there would be no need for a second measurement to determine the half-space result above the baffle step.

[ejh2854]

Thanks for the info- exactly what I needed.

Ed

[old_bassist]

Curves from three measurement operations using new setup

Thank you for the feedback gentlemen. I'm downtown so I can't implement all of your recommendations. I didn't mention that I used REW's built-in calibration file for my Galaxy CM-140 when I did my first round of tests yesterday. I used the same cal file today and also made these changes:

• Created a cal file for my audio interface. The Behringer UCA202 looks to be amazingly flat from 20 Hz to 20 kHz.
• Placed the speaker cabinet on its back, pointing straight up and placed the SPL meter directly above the center of the baffle, 30" from the front of the baffle.

My thinking was that the speaker would now pick up almost entirely direct sound with minimal influence from nearby reflective surfaces. I also ran three tests in a row to minimize the impact on ambient city noise.

Like yesterday, the graphs are smoothed to 1/12 octave. Hey, there are 12 musical notes in an octave, right?

Thanks again.

[Bill Fitzmaurice]

• Placed the speaker cabinet on its back, pointing straight up and placed the SPL meter directly above the center of the baffle, 30" from the front of the baffle.

That's way too close, the individual driver wave fronts need at least three times that before they integrate into a single cohesive wavefront. Also, you'll get an Allison Effect dip where the baffle is 1/4 wavelength from the ground. You eliminate that by taking a ground plane measurement, then splicing the two results. You can't get a full range 1/2 space result in one take unless you dig a hole in the ground for the cab to sit in and then back fill it so that the baffle is flush with the ground.
As for smoothing, if the intent is for EQing smooth it to the same 1/3 octave that your EQ will adjust to. To see what you hear, make it 1/6 octave, that's how precise your ears are. To drive yourself nuts seeing all sorts of stuff that you can't hear use finer than 1/6 octave.