Best (Affordable) Method To Capture a Speaker's Frequency Response?

[Spells]

I have a unique speaker that I find particularly useful. A cheap thing. I want to know what it is that makes it so useful to me. I have thought to measure its frequency response, out of interest.

Best way to do it accurately?

[resistorman]

I don't know much about this, but some questions occur to me. Is it just the speaker itself, or is a cabinet involved? It seems to me that frequency response is just part of the story?

[adrian_k]

I think “speaker” should be taken to mean the whole system (cabinet, driver, tweeter, crossover, port etc).

Personally I think this is a can of worms. Frequency is not whole story, there is also the time domain stuff. Then there are the problems of measuring just the speaker, and not the room response. Especially at low frequencies.

But I guess it might be fun to try something like Room EQ wizard (free software) with a flat response mic, in a very well treated room, and compare your main monitors with this speaker. It might give some clue as to what is different about this speaker, but I doubt it will be “accurate”.

[Wonks]

If you really want to know about it then you’ll want to know about transient response and harmonic distortion as well.

I’d refer to one of Phil Ward’s monitor speaker reviews in SOS magazine for all the tests that will tell you something useful about why you like it.

‘Speaker’ can mean different things to different people, so to start with are we talking about a single speaker chassis free of any enclosure, a single speaker in an enclosure, or two or more speakers in an enclosure?

Do you listen directly on-axis, or is it off to one side?

How well acoustically treated is the listening environment and whereabouts in your room is the speaker located and how is it mounted? Room acoustics can do a lot to skew frequency response results so to do it properly you need a very flat response room.

I did a spate of the ‘opposite’ measurements, using a single Genelec monitor to measure the relative frequency response of different (mainly low-cost) stage mics against an SM58.

It was good for a relative comparison, but the speaker’s low-frequency roll-off affected all the results, and dips and peaks in the frequency response at the same points in all the graphs indicated either room standing wave nulls or peaks, or areas where the monitor itself didn’t have a very flat response. Without repeating the tests in a different room, you can’t separate room from speaker effects.

But I could simply look at the difference between the results and say mic A has 3dB more gain than mic B at this frequency, and a lot more bass response.

So it’s pretty easy to do a comparative test against an existing monitor, but if you want to a comprehensive test of its responses in the frequency and time domains, plus a measure of its distortion, then it becomes a lot harder.

You’ll want a very wide frequency band measurement mic with a calibration file that can be used to linearise the results obtained.

You’ll need an acoustically treated room such that it’s flat across the frequency range you want to measure. Note that Phil Ward’s reviews don’t go into too much detail at the very low frequencies because even his own home test environment isn’t laboratory perfect so he will indicate the low frequency reach, but not comment in too much detail on the exact very low frequency performance.

Most monitor speaker frequency responses are measured in a ‘free space’ environment, i.e. raised off the ground and away from walls that will give increased bass response figures. You’d need to be able to mount the speaker in such a location on a stand that didn’t affect its performance e.g. not located in the middle of a much larger support platform.

That measurement space will ideally be empty of anything else but the speaker, speaker stand, mic and mic stand, to avoid reflections from other equipment affecting the results.

You can if course monitor in-situ (which could be right in the corner of the room), but that will make it hard to compare with test results from monitors measured in free-space environments.

You’ll also need suitable test software and an audio interface with a known flat response.

So if you really want to know all about the speaker, you may be best sending it off to a lab to be tested. But if you want to roughly know how different it is from your existing monitors, then buy a cheap measurement mic and just go for it.

Position it the same distance from the speaker each time (about 3 feet/1m away is common). For a single driver soeaker, orca coaxial speaker, get the mix as on-axis as you can. For a two-driver speaker, you’ve got the choice of recording on-axis with the tweeter (especially if the tweeter is normally at ear level), or halfway between the tweeter and woofer. For a three-way speaker, then either the tweeter position or the rough centre of all three drivers. Or measure both the tweeter and mid-driver positions and compare the results.

[MarkOne]

There aren’t many room reflections in a field.

You could perhaps drive out to a big empty space and run the speaker off a battery powered amp.

Most DAWs have a test tone plugin, or there is a free one from blue cat audio.

A beheringer ECM 8000 measurement mic isn’t lab standard, but it’s £30.

Set your mic and speaker up on some kind of. Non resonant platform well off the ground on a really still day, and record your sweep.

Of course even on a really still day there will be the sounds of nature

Failing that you could see if a nearby university physics department has an anechoic chamber they’d be willing to rent.

[Spells]

Food for thought! Thank you all.

I can see it'll be hard to get an accurate read. I like the idea of the uni / lab route!

[Hugh Robjohns]

Hi believe Phil Ward uses Fuzz Measure for his testing:

https://www.rodetest.com/

(It only runs on Macs).

Acoustic analysers like this normally have a time windowing function so that if you test the speaker with an impulse or brief chirp, you can exclude anything picked up by the mic after so many milliseconds — thereby rejecting first reflections from nearby boundaries. This helps in less than perfectly anechoic rooms, but also constrains the measurement's LF bandwidth. No free lunches!

Testing on a high pole out of doors is a good substitute, especially in the dead of night when most birdies stop singing. Ply the neighbours with lots of booze before hand.....

[Tomás Mulcahy]

Testing on a high pole out of doors is a good substitute, especially in the dead of night when most birdies stop singing. Ply the neighbours with lots of booze before hand.....

[Spells]

All noted !

[Sam Spoons]

Testing on a high pole out of doors is a good substitute, especially in the dead of night when most birdies stop singing. Ply the neighbours with lots of booze before hand.....

Or wait for winter and a thick fog?

[Roland Harris]

No need for a pole: grass + earth is pretty unreflective. I just take speakers outside to my 'poor man's anechoic' chamber when needing to measure for crossover design etc.: if you scroll down to the second to last photo in this blog post of mine you will see this in action! https://drbadphil.com/speakers-exponent ... ered-horns
Cheers,
Roland

[Martin Walker]

No need for a pole: grass + earth is pretty unreflective. I just take speakers outside to my 'poor man's anechoic' chamber when needing to measure for crossover design etc.: if you scroll down to the second to last photo in this blog post of mine you will see this in action! https://drbadphil.com/speakers-exponent ... ered-horns
Cheers,
Roland

What a fascinatingly detailed and infromative website you have Roland!

[tea for two]

https://drbadphil.com/speakers-exponent ... ered-horns
Cheers,
Roland

They are from Dr Who. Enjoyed reading this.
Also the link to B&W Nautilus. Inspired from a Snail : curves to reduce inside speaker box resonances.
https://m.youtube.com/watch?v=0FKyfTsyY ... VzIHN0b3J5

https://www.soundonsound.com/reviews/blueroom-minipod
Offshoot from B&W.

[Wonks]

No need for a pole: grass + earth is pretty unreflective. I just take speakers outside to my 'poor man's anechoic' chamber when needing to measure for crossover design etc.: if you scroll down to the second to last photo in this blog post of mine you will see this in action! https://drbadphil.com/speakers-exponent ... ered-horns
Cheers,
Roland

Though the drivers of the speaker used in that instance are about the distance from the ground that I’d have arranged small monitors (e.g. KH80s) to be located on stands (or a pole).

But I personally wouldn’t use a very tall pole as any very small benefits in reduced reflected sound are far outweighed by the difficulties in maintaining a consistent measuring position and a non-resonant pole.

I’ll point out that there is no single type of ‘ground’ and that there will be very different absorption and reflection coefficients between hard-baked earth with almost no vegetation on, and soft well-watered turf.

Has anyone looked at ‘green wall’ absorption panels made from turf?

[Roland Harris]

What a fascinatingly detailed and infromative website you have Roland!

Thanks Martin!

[Roland Harris]

They are from Dr Who. Enjoyed reading this.

Thanks!

In terms of where exponential horns went after the Nautilus, of course designer Laurence Dickie went on to Vivid Audio, where he has much advanced the idea (and with much better drivers). But this is a bit off-subject!

Cheers,

Roland

[Roland Harris]

Though the drivers of the speaker used in that instance are about the distance from the ground that I’d have arranged small monitors (e.g. KH80s) to be located on stands (or a pole).

Fair enough Wonks: indeed, the drivers weren't at ground level even if the speaker wasn't raised. Having said that, with present (soggy) ground conditions here I'd be surprised if anyone could raise a reflection off my lawn however near!

Cheers,

Roland

[Spells]

No need for a pole: grass + earth is pretty unreflective. I just take speakers outside to my 'poor man's anechoic' chamber when needing to measure for crossover design etc.: if you scroll down to the second to last photo in this blog post of mine you will see this in action! https://drbadphil.com/speakers-exponent ... ered-horns
Cheers,
Roland

Wow! Your speakers look incredible!

Out of interest, what would you charge if I were to send you my little speaker for some frequency plot graphs to be done?

[Roland Harris]

Sorry Spells for the delayed reply: I was away, then not well. Thanks for the kind thoughts on my speakers.
Mmmm, I don't think I'm the person to start offering (my homely or, rather, garden-based) speaker plots: I am sure you could find better or, indeed, do it yourself as well, if not better, than me - using REW (free) and an inexpensive Dayton Audio EMM-6 measurement mic.
Cheers,
Roland

[ajay_m]

While I'm not quite sure how to implement the details here, one thing I've always thought of is that, suppose you have an ok-ish mic in a (probably) untreated or semi-treated room. Then you would like to capture the impulse response of a speaker and somehow compensate for the room and mic.

One interesting approach seems to be to have an ideal impulse source, and the simplest one I can think of is a spark gap. It's relatively trivial to build a small circuit that repetitively triggers a spark across a small gap, producing pretty much close to an ideal impulse source. My old gas boiler has a little circuit like this that runs on a 9V battery and produces a nice constant 'zap zap zap' when you're lighting the boiler front fire (yes, I said it was old!).

Now if we record this through our mic we are eliminating the speaker at this point. Our spark gap comes very close to an ideal impulse source.

We then repeat with the speaker reproducing the impulse instead of our spark gap. It ought to be possible for some clever software to then take the 'delta' and figure out what colouration the speaker is adding, since the other factors (room and mic) remain constant in that scenario.

[Hugh Robjohns]

Might be worth reading up on MLSSA audio testing which has been the standard speaker measurement technique for years now.

http://www.mlssa.com/pdf/MLSSA-Brochure.pdf

[Spells]

Many thanks all! Will have a read up