What accessories or hardware to be used for speaker measures

Bemopti123 · « **on:** 2 Jan 2006, 05:37 am »

I am interested in running some diagnostics about frequency extension and other features of my horn loaded speakers.
What do you recommend in terms of equipment to do this?
I have an Apple Powerbook I can use for this and have even found a freeware that does some sort of analysis. The Powerbook has a mic input, so, is it possible to use something like this....

http://www.dreer.ch/fft.html

or

would I need more accessories such as a USB microphone

or

should I do a Behringer DSP equalizer job?

I would like to hear from people and the equipment they use to measure speaker response.

Thanks

warnerwh · « **Reply #1 on:** 2 Jan 2006, 06:07 am »

The Behringer units can do this with the Behringer ECM 8000 microphone. This mike btw is very accurate. I use a DEQ 2496, a wonderful tool. These with the mike can be purchased for about 300 on the net if you look. Of course there are more expensive alternatives but the price jumps fast. If you bought this unit you can also use it to help in room correction.

Remember the average room has about a plus or minus 15 db variation so don't be too shocked at those measurements as many people are when they see them. If you decide to go this route you need to make sure you know what you are doing if you use it for room correction too. Good luck. This is a very fun and earlightening experience.

ctviggen · « **Reply #2 on:** 2 Jan 2006, 02:23 pm »

Here's one:

http://www.purebits.com/speaker-design-fy.html

ETF will measure room response, but you can use it to measure speaker response provided you take the speakers outside.

http://www.etfacoustic.com/

There's also CLIO, directed more toward speaker measurements:

http://www.cliowin.com/

There are more.

The FFT program you cited would simply take a digital signal and perform an FFT on it, which would give you an estimate of the frequency response of a portion of the digital signal. What the programs above typically do is attempt to detemine an "impulse response" (really a misnomer as the input isn't an impulse) of a speaker/room and then use a version of an FFT to determine a frequency response. Maximum Length Sequences are used to create something similar to the frequency response of an impulse (i.e., all frequencies represented evenly). All frequencies should be represented and the impulse response would have all frequencies. Alternatively, low frequency analysis is performed by using sine waves from low frequency (20 Hz) to higher frequency. This provides the frequency response. Note that this method will include the room response (as will the MLS measurement) unless the speakers are taken outside. For the MLS measurement, they recommend for inside measurements that you put the speaker in the middle of the room and try to cut off the portion of the impulse response due to the room. Personally, if I had the option and ability, I think taking the speaker outside and pointed upward would be the best and provide the best response.

samplesj · « **Reply #3 on:** 2 Jan 2006, 02:23 pm »

You can get a calibrated measurement mic from Doug at Acoustisoft for $325. While its pricer than the ECM8000 it is calibrated so an individual unit quirks are compensated by any tools that can read the calibration file. It also includes a mic preamp so you don't have to hunt for phantom power to feed it.

I also like ETF as a tool, but that adds another $150.

Jeremy

tonyptony · « **Reply #4 on:** 5 Jan 2006, 04:00 pm »

Jeremy, this mic sounds interesting. Can you recommend a good PC based analysis program that would be able to read the calibration file?

I've always wondered about the PC based options, though. How good could they be if the signal from the mic is going through a low end sound card?

samplesj · « **Reply #5 on:** 5 Jan 2006, 09:02 pm »

I know ETF will. ETF is the software that Doug/Acoustisoft created (thats why he sells mics).

As far as the sound card, there are some cheap (~$60) USB cards that measure pretty well for the specs that matter for speaker measurement. I think Doug has a couple of recommendations.

I'm not sure, but I think the mics may have had a recent price increase.

Sintz · « **Reply #6 on:** 5 Jan 2006, 09:15 pm »

It doesn't really matter what program you use to measure as long as you're using a calibrated mic. Like warnerwh noted, the Behringer units work well (I also use a DEQ2496 with an ECM8000 mic). The ECM8000 mic is calibrated but requires phantom (12V) powering, which the DEQ2496 can provide.

Also note that a unidirectional mic should be used if possible to avoid heavily altered frequency response of omnidirectional mics have (no good if you're measuring responses at your listening position, which comes from all around you, not just in front).

Hope that helps.

samplesj · « **Reply #7 on:** 5 Jan 2006, 09:18 pm »

Quote from: Sintz

It doesn't really matter what program you use to measure as long as you're using a calibrated mic. The ECM8000 mic is calibrated but requires phantom (12V) powering, which the DEQ2496 can provide.

You can purchase calibrated ECM8000 from certain individuals that will measure them, but in I didn't think in general it was calibrated from the factory.

Calibration referers to a finding the deviations of that specific mic with a highly accurate known good mic as a reference (big $ to buy and big $ to keep accurate to that degree). Technically any mic should even have to be recalibrated periodically.

Sintz · « **Reply #8 on:** 6 Jan 2006, 06:05 pm »

Quote from: samplesj

Quote from: Sintz
It doesn't really matter what program you use to measure as long as you're using a calibrated mic. The ECM8000 mic is calibrated but requires phantom (12V) powering, which the DEQ2496 can provide.

You can purchase calibrated ECM8000 from certain individuals that will measure them, but in I didn't think in general it was calibrated from the factory.

Calibration referers to a finding the deviations of that specific mic with a highly accurate known good mic as a reference (big $ t ...

You're completely correct. I should have written "use as linear mic as possible, like an ECM8000..."
Thanks.

ctviggen · « **Reply #9 on:** 6 Jan 2006, 06:28 pm »

The spec sheet --if to be believed -- that comes with the ECM8000 shows the ECM8000 basically ruler flat from 20H-20kHz. I never bothered getting mine calibrated or getting a calibration file, as the inherent error in a program such as ETF way overwhelms the data from the mic.

Watson · « **Reply #10 on:** 6 Jan 2006, 07:40 pm »

There's a generic calibration file for the ECM8000 available here:
http://archive.avsforum.com/avs-vb/history/topic/479846-1.html

It's fairly flat up to around 15kHz, then it starts to misbehave.

ctviggen · « **Reply #11 on:** 6 Jan 2006, 09:05 pm »

Thanks for that file! I'll have to see if that data is on mine. I see what they've done -- they make the errors look small by choosing large values on the scale. Bastages! I'm going to have to figure out how to add this to ETF (though I'm primarily interested in low frequency data).