I've read of a high voltage discharge through a spark gap used to generate a good, wide-band acoustic impulse.How about a gun? You know, like a starter pistol - if it's the room that interests you and not the speaker. That's how a lot of impulse measurements of rooms are done.
That's what I wanted to do, until I found out you have to have a permit for a starter pistol in this stupid county.
Just one of many questions. I've got a nice little DC motor (that's easy to speed control . . . it has a separate set of windings that serve as a tachometer) that's suitably quiet . . . finding a fan that produces uniform broad spectrum "noise", and the correct speed to spin it . . . not obvious to me . . .
I suspect that was "around' 1mW, not "exactly" 1 mW . . .
Probably a synchronous motor. That way it would run at a constant speed locked to power line frequency and independent of power line voltage fuctations.
Maybe. AC induction motors running with constant load (which the fan would be) are very close to immune to normal line voltage changes . . . and "constant speed" within a few percent even with variable load. It could well be a synchronous motor, but that would not be a design requirement. There might be an advantage to a variable speed (adjustable) motor, as it would allow for optimizing the noise profile of the fan. I'm sure there are studies somewhere of the physical mechanism of noise production ("quiet" fans being a marketable commodity), I just don't know where to find them . . .
I don't understand why you need a flat, constant noise source to measure room response. Can you explain that?
Back then they needed the fan because they weren't able to measure all 1/3-octave bands simultaneously, so they had the tedious process of measuring decay in each band separately. A known noise source was a shortcut for them to save time. Now you can measure decay in all bands simultaneously from either an impulse source (balloon pop is what we use, but starter pistol works too) or a build up and cut of pink noise. As long as you have sufficient energy in all bands, their relative levels don't matter. You can look at the decay in each band independently to determine your room response.
Back then they needed the fan because they weren't able to measure all 1/3-octave bands simultaneously, so they had the tedious process of measuring decay in each band separately. A known noise source was a shortcut for them to save time. Now you can measure decay in all bands simultaneously from either an impulse source (balloon pop is what we use, but starter pistol works too) or a build up and cut of pink noise. As long as you have sufficient energy in all bands, their relative levels don't matter. You can look at the decay in each band independently to determine your room response.
A balloon pop! Will have to try that.
Looking at the MacIntosh fan, it has an end plate with a hole. Maybe that's how it was calibrated?
How would you use a noise source like this? I understand the impulse measurements because I've used them, I'm not sure about the continuous noise source, tho some software will convolve and impulse from a noise signal.
Looking at the MacIntosh fan, it has an end plate with a hole. Maybe that's how it was calibrated?
How would you use a noise source like this? I understand the impulse measurements because I've used them, I'm not sure about the continuous noise source, tho some software will convolve and impulse from a noise signal.
Balloons are light, inexpensive, simple to operate, and allowed on all aircraft (as far as I know).
Once I was measuring a room's response on an out of town job and I forgot to bring a tripod, meaning I had to hand-hold my meter for the test. Popping a balloon is usually a 2-handed job, plus I didn't want to pop the balloon right next to my meter, so I had a challenge on my hands.
I happened to be measuring a kitchen, so I set the balloon on the island, took a few steps back, started my meter, and gracefully threw my pencil at it point first like a dart. It worked... pop! Except I had a second impulse in my data from the sound of the pencil hitting the floor. Luckily that happened more than a second after the 60 dB decay I needed so it didn't effect my measurement.
Once I was measuring a room's response on an out of town job and I forgot to bring a tripod, meaning I had to hand-hold my meter for the test. Popping a balloon is usually a 2-handed job, plus I didn't want to pop the balloon right next to my meter, so I had a challenge on my hands.
I happened to be measuring a kitchen, so I set the balloon on the island, took a few steps back, started my meter, and gracefully threw my pencil at it point first like a dart. It worked... pop! Except I had a second impulse in my data from the sound of the pencil hitting the floor. Luckily that happened more than a second after the 60 dB decay I needed so it didn't effect my measurement.
What meter do you use? How does it display the room's absorbtion frequency response? Can it measure less than third octave bandwidths (identify narrow resonances)?
One doesn't need (as if it's the only way) such a source, but it appears to be a convenient and easy way. Doesn't measuring the reverberant impulse decay require a quiet room?
On that day I was using a Larson Davis 824. It does not have filters narrower than 1/3-octave, but it does do FFT. For general room acoustics, 1/3-octave is plenty.
Quiet enough for at least 5 dB, preferably 10 dB of headroom. I'm not aware of any acoustic measurements that don't require low enough background levels, unless you're measuring ambient.
Well, that looks like a nice enough gadget, but beyond the means of most DIYers. A constant power noise source, however, should give the same (or even higher resolution) results when used with measuring equipment and software that many of us already have, at (potentially, anyway) a much lower cost.
And the continuous output of an acoustic noise source (fan or otherwise) should make such a test much less sensitive to ambient noise . . .
What is the measurement equipment and software DIYers already have? Do DIYers typically have sound level meters from the 1970's and 1980's?
The reason the calibrated broadband noise source is more convenient is it prevents you from having to shoot RTs for each band individually. Instead, you can check constant levels in each band and, assuming your source is indeed calibrated, you can see the relative response of the room in each band quickly.
I can't imagine that's a technique that's being used anywhere seriously today when it's so easy to measure all bands simultaneously in the time domain. And I really can't see how it would be more accurate than measuring decay times. It requires that you have a perfectly flat broadband noise spectrum. Is that even possible? You're putting a lot of faith into a piece of equipment to use that technique.
I'm not aware of any kind of audio measurement equipment, professional or otherwise, made in the last 20 years, that requires you to step through bands one at a time to make a measurement. You're really solving a problem that no longer exists.
It's not at all a question of the right equipment being out of reach of your average DIYer. If you can afford a laptop, you can afford far more processing power than you need for 1/3-octave, 1/12-octave, or even very narrow band FFT.
REW does perfectly valid room response measurements and it's free. Hardly in the realm of technology only available to professionals. Add a SLM with an audio passthrough and you've got a complete, reasonably accurate system.
Where are you doing this test? I don't have a lab, all of the RT measurements I've taken were in the field. It's really not that hard to get 10 dB of headroom in a typical residence in the middle of the day. You don't need to be in a bunker or an acoustic lab. Just turn up the volume if you need to. You can even use RT30, if that's what it takes.
One other thing to consider is that a frequency vs. magnitude response at one location in a room isn't the response of that room (unless the room is perfectly reverberant or perfectly dead) even when the sound source's response is "backed out". It's only the magnitude response at *that* location and will be different at another location. This is independent of sources and measurement technique.
Thanks,
Chris
Thanks,
Chris
Which is why ISO 3741 dictates that you take measurements at various locations.
Yes indeed. And why everybody uses impulses and laptops to sum 'em. Too easy these days and too hard any other way. (I'm the exception - still have a 1/3 octave RTA on hand.)
Thanks,
Chris
Last edited:
I commonly see references to SoundEasy (and associated equipment), ARTA, and the new OmniMic system . . . while a few "old timers" still use "previous generation" software.
I doubt it (although I still have my (very) old Radio Shack analog meter). Why do you ask?
Yes . . . a simple "spectrum analyzer" (they're even "built in" to some audio gear) gives an immediate visualization.
I haven't suggested that it would be "more accurate" . . . if impulse response measurements are accurate then the results should be the same. There are often different ways of measuring the same thing . . . where results differ (if they do) there's usually something interesting to explore.
No more than impulse response measurements require "perfect" impulse (is the balloon "perfect"?) or either technique requires "perfect" microphones. "Good enough" is . . . well . . . good enough.
There's that "perfect" again . . . REW is "perfect"? I expect instead that it's "good enough" for most purposes, within its limits, which it, like every other measuring technique, has. I don't see a good $20 (or whatever) broadband noise power source as threatening it, or the technique, but as complementing it. Maybe I'm missing something (it has happened before), but I'm not seeing why some people find the very idea of a constant power noise source so threatening . . .
I think your understanding is a bit off here. It doesn't matter if the source frequency spectrum is flat for a decay measurement. Each band decays at its own rate and it doesn't matter whether it all starts out at the same level.
The nature of the source doesn't matter, provided it can excite all of the frequency bands you're interested in. Impulses are useful because they have energy across the spectrum. Balloon pops are convenient, as are starter pistols. One time, in a pinch, I banged a couple of 2x4's together because it was all that was available.
Another common method is to excite the room with pink noise (or similar) and cut the noise source off, then watch the decay in each band in the time domain. Your built up noise doesn't need to be even across the spectrum, it just needs to have enough level with respect to ambient (in that band).
In contrast, using a calibrated sound power source requires your source to be special and reliable. You have to have faith in your source that it has even, known sound power in every band, which is very difficult to achieve. Why go through all the trouble when you can buy a pack of balloons for $1.19 at Walgreens and have better results?
When you put something in quotation marks it's supposed to be exactly what the person said, and it's definitely not supposed to change the meaning. I did not say "perfect," I said "perfectly valid." You could drop the word perfectly and the sentence still has the same meaning.
Exactly what I'm saying.
I don't see a $20 noise source with known sound power in each band as existing.
You are. Let's just walk through the process and maybe you'll be able to see the problem.
Suppose you discover a fan on Craig's List that looks similar to the one in the article you posted and you buy it for $20. Then you take it to your test area, fire up your RTA system (whatever it is), and turn on your new source.
You see and measure the spectrum it creates in your room in 1/3-octave bands. 65 dB, 68 dB, 69 dB, 66 dB... etc. OK. So what do you do with that information? How can you use these numbers to determine the room response? Based on what you know so far, you can't.
What you need to know is the sound power of this fan in each band before you can derive some kind of meaning from your measurement. Just knowing the relative sound power from band to band (and you don't even know that) isn't good enough, you need to know the actual per-band sound power. So how do you know that? You have to measure it.
And how do you measure it? Well, you take some RT60 measurements of your room to determine the room response in each band, then use that information to determine the sound power of your fan according to ISO 3741. But why bother with the fan in the first place if you have to measure the room response anyway?
Why not just cut the fan out altogether and use the RT60 measurement to quantify the sound power of the speaker you're interested in in the first place? Your results are going to be more accurate, less work, and less money.
What I think you're missing is the reason they were using the calibrated noise source in the first place. In the old days, you couldn't measure an entire spectrum at once. You had to measure 50 Hz, then measure 63 Hz, then measure 80 Hz, and so on. It took a long time just to capture a single spectrum, plus you had to assume that whatever sound you were trying to measure stayed constant.
Now consider how much more time needed to be spent doing RT measurements in this way. Measuring multiple decays in each 1/3 octave band in the time domain every time you wanted to check room response would have been extremely time consuming. When someone walked in with a gizmo with known sound power, they jumped all over it because now all they had to do was take a series of sound level readings to get the room response, or an approximation of it that was close enough for their purposes . But they had to have faith in the reported sound power of this device.
Fast forward to today. Now we can use free software and/or inexpensive hardware to measure 24 1/3-octave bands simultaneously, even in the time domain. Plus we have software that does the RT calculation for us automatically. There's nothing to do but hit run, pop your balloon, and hit stop. Or, if you have your measurement device hooked up to a source, you just hit run and let it run its own test, where it plays noise, cuts it, and measures the response. You can even use a sine sweep now, which supposedly offers even better results (REW does this).
So the impetus for having the calibrated source is gone. There's no longer any point to it because it doesn't save us any time. In fact, it adds time because it causes us to perform measurements that we don't actually need to do. It also adds expense because it causes us to purchase something we don't need.
Last edited:
Here's something I've never been completely clear on. With the old system you could use pink noise and get the response via filtration. Albeit slowly. You could also do a sine sweep, with or without filtration. With the new system you can do impulse or use the pink noise and do an FFT. Is the output of all four methods theoretically identical?
The reason I ask is let's say there was a big brass church bell in the room. As the sine sweeps through the resonance, the bell starts building up a tone. You'll see it in the sweep. This doesn't happen with random noise.
The reason I ask is let's say there was a big brass church bell in the room. As the sine sweeps through the resonance, the bell starts building up a tone. You'll see it in the sweep. This doesn't happen with random noise.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.