speaker measurements

[kendt]

I was wondering how expensive and difficult it is to do my own speaker measurements. I really would like to see for real what my speakers are doing

[Svein_B]

It does not need to cost much money.
(but it will cost some of your time to understand how, and what you measure).

I suggest you download ARTA
The [free] demo version is fully functional, only lacking the possibility to save compleate measurement files to disk.
Using a PC with built-in soundcard and a microphone you can start practicing today!

Then you need a decent microphone.
The virtually no cost solution is a Panasonic WM-61A electret capsule at $2 a piece. The mic needs a battery and a couple of resistors for power feeding, as described here: Powering Microphones

There are also a few threads on thyis subject in the Loudspeaker section.

SveinB

[SY]

Do you have a basic understanding of acoustics and transform methods? If so, the hardware and software tools are quite cheap. A hundred dollar soundcard, some freeware, a calibrated mike capsule (don't know what those go for these days but can't be much more than $40-50), $20 worth of preamp and you're good to go.

If you don't have a good understanding of these measurements, you'll find that the hardware and software tools are remarkably powerful ways to generate tons of useless data. I'd recommend reading Joe d'Appolito's book on loudspeaker measurement while you experiment.

[Geoff H]

For a mic, I use the line out from a RS SPL meter to the line in on the sound card. It's not a B&K, but usually I am after comparitive data, not absolute.

Once you understand the software, it's pretty easy to do. Using the screen print - paste to a image editor, paint will do, you can save the results for future reference.

An onboard sound chip will get you going, an extra card will take some of the processing burden from the CPU, and sometimes give better s/n ratios. Keep it as far as possible from video and network cards.

Be wary of your noise floor. I was getting strange results one night. The chorus of frogs and crickets had raised the noise floor to 68 dB SPL! So much for using the backyard in lieu of anechoic chamber.

Regards,
Geoff

[Ron E]

The RS SPL meter is fairly accurate from ~100-1000Hz. It rolls off at the low end, and it has a peak and a sharp drop above.

My RS SPL meter response:

Hz......SPL
1000 +0.6dB
3000 +2dB
6200 +5dB
10000 0dB
20000 -18dB

Other RS SPL meters have measured quite similarly. Look at the "ETF Acoustic" page for a calibration they did. Very similar to mine.

[Geoff H]

Thanks Ron. Those figures must be close to my meter. The chart that came with it is well off. I don't need super tweeters as most of my "tests" indicate. Trust the ears.

Geoff.

[MarkMcK]

If you don't have a good understanding of these measurements, you'll find that the hardware and software tools are remarkably powerful ways to generate tons of useless data. I'd recommend reading Joe d'Appolito's book on loudspeaker measurement while you experiment.

First, a search will turn up many suggestions for testing applications.

Second, I second and extend sy's comment. If you are going to play around with a modeling or simulation application, they you need to understand acoustical theory. If you are going to go it by "ear," then you need to understand acoustical theory.

I do not recommend going it by "ear." Musical instruments have been designed by ear for a long time. Very few people have ever been capable of being good at it. The great ones numbered in the handfuls.

As there are thousands of not understood measurements lying around, there are many thousands of home built speakers lying around that are inaccurate and bad sounding.

[Geoff H]

Kendt.

One problem with running tests is analysing the results. You could spend hours worrying about the data, when its not the speakers, but the room. If you have a good ear for detail, and subject the ear to live music, you will know if the sound is approaching the real thing.

If you're chasing better sound, that's when need to draw on your acoustical knowledge.

"As there are thousands of not understood measurements lying around, there are many thousands of home built speakers lying around that are inaccurate and bad sounding"

Does that really matter if the speakers provide hours of enjoyment for the owner?

Is there an area where you think the speakers are lacking, or just plain curious?

And if you collect the data, be careful what you compare it to. I have seen published data using graphs, where one axis is not marked.

BTW, the free version of RTA, is limited to 1 octave sampling. You'll outgrow it real fast. ARTA is going on that PC before I make any further measurements.

Geoff.