Middle Ground?
I have been following this 'debate' with interest
I can see that a 10M / 100 W test might give more accurate results than a 1M / 1W test in the ideal world, but I would think that for the majority of DIYaudio people this is not going to happen!
An accurate 10m test would require a vast open flat area, preferably without neigbours to annoy.
The 1 watt / 1Metre (or should we say the 2.83V / 1Metre test) is the standard by which we compare speakers for efficiency.
From what has been stated so far, an increased distance should help with accuracy, so can we find some middle ground.
Looking at my garden I would think I could reasonably achieve 2 or 3 metres, but not much more. Would this be better than the 1Metre test?
I calculated
Watts as Metres squared
Volts as the square root of (Watts x Impedance)
I put this into an Excel spreadsheet, I see that a 32M / 1024W test would be interesting😀
Regards
Martin
Impedance 8 Ohms
Metres Watts Volts
1, 1, 2.83
2, 4, 5.66
3, 9, 8.49
4, 16, 11.31
5, 25, 14.14
6, 36, 16.97
7, 49, 19.80
8, 64, 22.63
9, 81, 25.46
10, 100, 28.28
I have been following this 'debate' with interest
I can see that a 10M / 100 W test might give more accurate results than a 1M / 1W test in the ideal world, but I would think that for the majority of DIYaudio people this is not going to happen!
An accurate 10m test would require a vast open flat area, preferably without neigbours to annoy.
The 1 watt / 1Metre (or should we say the 2.83V / 1Metre test) is the standard by which we compare speakers for efficiency.
From what has been stated so far, an increased distance should help with accuracy, so can we find some middle ground.
Looking at my garden I would think I could reasonably achieve 2 or 3 metres, but not much more. Would this be better than the 1Metre test?
I calculated
Watts as Metres squared
Volts as the square root of (Watts x Impedance)
I put this into an Excel spreadsheet, I see that a 32M / 1024W test would be interesting😀
Regards
Martin
Impedance 8 Ohms
Metres Watts Volts
1, 1, 2.83
2, 4, 5.66
3, 9, 8.49
4, 16, 11.31
5, 25, 14.14
6, 36, 16.97
7, 49, 19.80
8, 64, 22.63
9, 81, 25.46
10, 100, 28.28
I think for comparing prosound subs, the standard measurement needs to be done at a greater distance than two or three meters. Ten meters seems to be the distance most agree on as being most useful. The wavelengths being measured are on the order of ten meters, so it seems to me that is a good minimum distance for reference level testing. Ten meters is also convenient because at 100 watts, it calculates back to being equivalent to the 1W/1M figure.
On the other hand, I think that Josh makes a compelling argument that there is at least some value in the closer measurements, especially for smaller subs like those designed to be used for home hifi and home theater. Those will almost never be listened to at greater distances. It may make some sense to include measurements made at a couple meters. I think that's what you're saying too.
What is important to me is the understanding that while we all consider the 10M/100W value to be equivalent to the 1W/1M value, they're equivalent onnly in the sense that they calculate the same. In practice, the SPL at 1W/1M is almost never the same as the 10M/100W figure.
I have taken the position that the distant measurement is the "right" way to do it, since it reduces the path-length/acoustic-center error problem. I consider it to be more accurate. It does give the value that is most accurate in terms of the inverse-square law, letting you know what the subwoofer will perform like at distances. I consider it to be the most useful metric for prosound subwoofers.
I don't see any reason not to also include closer measurements. It might be interesting to see measurements done at various distances. But I do think prosound hornsub measurements should be measured at 10 meters for comparison with other subs.
On the other hand, I think that Josh makes a compelling argument that there is at least some value in the closer measurements, especially for smaller subs like those designed to be used for home hifi and home theater. Those will almost never be listened to at greater distances. It may make some sense to include measurements made at a couple meters. I think that's what you're saying too.
What is important to me is the understanding that while we all consider the 10M/100W value to be equivalent to the 1W/1M value, they're equivalent onnly in the sense that they calculate the same. In practice, the SPL at 1W/1M is almost never the same as the 10M/100W figure.
I have taken the position that the distant measurement is the "right" way to do it, since it reduces the path-length/acoustic-center error problem. I consider it to be more accurate. It does give the value that is most accurate in terms of the inverse-square law, letting you know what the subwoofer will perform like at distances. I consider it to be the most useful metric for prosound subwoofers.
I don't see any reason not to also include closer measurements. It might be interesting to see measurements done at various distances. But I do think prosound hornsub measurements should be measured at 10 meters for comparison with other subs.
This is the crux of our difference of opinion.
Josh Ricci and I share the same opinion.
It is important to measure the distance to a sub’s acoustic center to get time alignment correct, but the inverse square drop in level starts happening from the cabinet mouth or baffle, not the acoustic center.
My and Phil’s tests presented in post #237 prove that the inverse square drop in level starts happening from the cabinet mouth or baffle, not the acoustic center, the results are unambiguous.
You have made several statements like in #254 claiming “several dB difference” between 1 and 10 meter results, but have never actually methodically doubled the distance and measured the results as we did.
If you want to continue to claim that two meter tests are invalid, and that “the acoustic center is the location where the inverse-square law begins to form” you should provide the data that shows the “several dB difference” between 1 and 10 meters and points in between for your 12Pi.
To make your test more valid, extend it out to 32 meters as we did, comparing horn load to front load.
Since your contention is “the acoustic center is the location where the inverse-square law begins to form” you may also want to measure at the mouth and the acoustic center too.
Art Welter
Martin, it’s not that the 1w/1m measurements are useless but 100w/10m is much more functional to PA cabs. With 1w/1m measurements you have to know how to interpret them in relation to 100w/10m. If you want to dive into specific frequencies or precision within a few decibels it is not done to compare. For overall comparison where this precision isn’t an issue it can be used for PA, in my view. For instance, tolerances in electric values between drivers of the same type, can become a bigger issue. Also 100w/10m measurements can be questioned since most of them don’t provide additional information (wind figures, temp, background noise figures, driver tolerances etc). 3 meter measurements already provide an acceptable error rate to compare in detail. So for every distance there are advantages and disadvantages.
Besides all points that were mentioned there is another issue. Most leading PA manufacturers optimise their LF PA drivers (especially their suspension) for high loads and high excursion. With 1w/1m measurements the driver’s suspension is often not as linear compared to higher ratings. Some can even show spikes as part of these mechanical factors. Personally I prefer 1w/1m and 100w/1m for instance, before you go to 10 meters. These extra steps at 1 meter can give information in relation to power compression figures and characteristics in suspension/cone.
Besides all points that were mentioned there is another issue. Most leading PA manufacturers optimise their LF PA drivers (especially their suspension) for high loads and high excursion. With 1w/1m measurements the driver’s suspension is often not as linear compared to higher ratings. Some can even show spikes as part of these mechanical factors. Personally I prefer 1w/1m and 100w/1m for instance, before you go to 10 meters. These extra steps at 1 meter can give information in relation to power compression figures and characteristics in suspension/cone.
Last edited:
Art, you're just plain wrong. I'm honestly surprised you persist in making statements like that.
I don't believe you should speak for Josh, saying "Josh Ricci and I share the same opinion." He has repeatedly said that he understands the difference between the 1W/1M and 10M/100W values. In fact, he has measured different values, same as I have.
Josh's argument appears to be that both the actual 1W/1M and the 10M/100W measurements would be valuable. I kind of agree with him there.
My point has always been simply that you cannot compare a 1W/1M measurement with a 10M/100W measurement, unless the acoustic centers of the microphone and the DUT are precisely known in the 1M measurement. When going from face to mic, the 1W/1M measurement will almost always be different than the 10M/100W measurement. I have also suggested that prosound subs should probably be compared at 10M/100W, since this reduces path-length/acoustic-center error.
Here are the opinions Josh Ricci and I agree on and you disagree:
Josh writes in #221
“I agree with Weltersys here.
The diaphragm may be further inside the horn but that is completely irrelevant to the listener or end user other than perhaps for setting delays between cabs. When you simulate something in HR or another program it calculates the acoustic performance at some specified distance from the final radiation point of the cabinet, not from a driver sitting 3m or whatever inside. “
From #230
“Wayne see my last post please. I simply disagree that in the case of a basshorn that the acoustic center matters at all as regards its real world use and performance. I also disagree that 2 meter or 1 meter measurements are flawed for basshorns or other subs....
The horn affect on the driver output doesn't mature fully and take shape until it reaches the mouth and exits the cab anyway. “
From #232 Josh writes:
“You seem to be implying that the long path from the driver to the mouth of the horn is already engaging the inverse square law such that if the acoustic center of the FLH is back 2.5m in the horn then the output should only drop by 12db at 10m instead of the 20db a direct radiator would see which would place the FLH at advantage at long distances and disadvantage or on equal footing, depending on your view point at shorter distances like 1 meter. I have not seen this born out in any measurements yet. The ones I have seen seem to trend the other way.”
The “path-length/acoustic-center error” is an audio myth that you continue propagating, similar in age to the myth that still floats around that clipping (“square waves”), rather than too much average power, is responsible for burning up speakers.
Having learned both those myths early in my career, it took a lot of education to finally drop the square wave myth, and it was not until I actually measured did I fully let go of the path-length/acoustic-center error myth, even though empirical results had shown it false for years.
Your contention that “the acoustic center is the location where the inverse-square law begins to form” is wrong, if you bothered to test you would understand your error.
I again suggest you do just that, rather than continue to insist with no proof of your own something that is demonstrably right is wrong.
Art Welter
Dude, go away. You're just trying to talk your way out of a corner and it's annoying to the rest of us that are looking at real stuff.
Please look at the “real stuff" presented in post #237 and explain why the truth annoys you .
These arguments have been discussed at some length over the past few pages. I think there is agreement among all of us here (except maybe Art Welter) that there is a difference between 1W/1M values and 10M/100W values. Look at the previous page from yesterday - See post 256 and those that follow for what I consider to be something of a conclusion to the acoustic center discussion.
Wayne,
I completely agree that a speaker will respond differently to 1 watt than 100, or 10,000 watts, regardless of the measurement distance.
But Josh Ricci and I (and many others not bothering to write) have a fundamental disagreement with you, as pointed out in #266.
Back to that discussion that for some reason annoys you so much, I agree that longer measurement distances would reduce “path-length/acoustic-center error” if it existed, but it does not.
You have not discussed the information presented in post #237 at all, and neither your post #256 or any others have a shred of data to support your “path-length/acoustic-center error” myth as it relates to measurement distance of subwoofers.
Man up, discuss the data presented instead of hiding from it.
Art Welter
I really don't know how to respond to you, Art. I don't like our talks being so ugly. I hate it when messageboard discussions get like that.
If you want me to respond to why I think you find the concept of the acoustic center to be a myth, my answer is, I don't know. I think you're wrong. If you want me to respond to why you have reported data that shows errors, my answer is, I don't know that either. But I do know it's not the first time I've seen you post data and/or interpretations that don't make sense.
You know, the disagreement Josh and I had was really one of semantics. I said I thought the 1W/1M measurements of a prosound subwoofer were meaningless. What I really meant was they were meaningless for comparison with 10M/100W measurements, because of the acoustic-center ambiguity problem in the closer measurement. But his point is that even though this is so, he feels the 1W/1M measurement is valid, on its own merits. Not so much for comparison with 10M/100W values, but for comparison with other 1W/1M values. I can respect that, even agree with it.
You are saying you think the whole concept of the acoustic center is a myth. I find that to be a very sophomoric argument, one that I just cannot respond to in any way. It's like you just dig your heels in and declare the Earth is flat. I don't know how to have a discussion with you about that.
Originally Posted by weltersys
I agree that longer measurement distances would reduce “path-length/acoustic-center error” if it existed, but it does not.
You have not discussed the information presented in post #237 at all, and neither your post #256 or any others have a shred of data to support your “path-length/acoustic-center error” myth as it relates to measurement distance of subwoofers.
The concept of an acoustic center, or acoustic source is most certainly not a myth.
The fact that most loudspeakers have acoustic centers that vary with frequency is a primary reason that arrays of loudspeakers often have terribly ragged polar response.
Your concept that the path length of a horn will change the distance from the mouth where the inverse square law applies to some further distance in front of the horn is the concept we disagree on.
Josh Ricci, and others agree with me, and disagree with you on this particular point.
To avoid semantic problems, let’s attempt a civil discussion on the specific point in bold above we disagree on.
If you found errors in the methodology or data presented in post #237, please point them out, and please explain how you would interpret the data differently from me.
Art Welter
I agree that longer measurement distances would reduce “path-length/acoustic-center error” if it existed, but it does not.
You have not discussed the information presented in post #237 at all, and neither your post #256 or any others have a shred of data to support your “path-length/acoustic-center error” myth as it relates to measurement distance of subwoofers.
Wayne,
The concept of an acoustic center, or acoustic source is most certainly not a myth.
The fact that most loudspeakers have acoustic centers that vary with frequency is a primary reason that arrays of loudspeakers often have terribly ragged polar response.
Your concept that the path length of a horn will change the distance from the mouth where the inverse square law applies to some further distance in front of the horn is the concept we disagree on.
Josh Ricci, and others agree with me, and disagree with you on this particular point.
To avoid semantic problems, let’s attempt a civil discussion on the specific point in bold above we disagree on.
If you found errors in the methodology or data presented in post #237, please point them out, and please explain how you would interpret the data differently from me.
Art Welter
There are a lot of things that affect the acoustic center, and path length most certainly is one of them. It is obvious in the phase matter, which I think you agree on. It is also obvious in the amplitude matter, if you think about it. Maybe not in all cases, but certainly in many of them.
A horn creates an impedance transformation by matching the wavefront expansion with its physical area expansion. The cross-section of the horn forms a pie-slice that is the radiating angle. So it is already developing an expanding wavefront before it even leaves the horn.
At the horn throat, pressure is highest, and it is gradually reduced as the wavefront travels down the horn. In fact, in a conical horn, the pressure attenuation down the horn is precisely the same as a spherical wave in free-space, exactly following the inverse-square law. So you can easily see that the acoustic center of a traditional horn is pretty much where the radiating diaphragm is.
There are some extra complications, of course. The electro-mechanico-acoustic properties make the impedance vary from resistive to reactive, especially at the low end. This is the matter of phase. This is even more pronounced in an undersized horn, as is the case for most basshorns. But in groups, the horn can be made to be uncompromised, so it will act as a true, full-size horn.
So it's not so simple to say "the acoustic center is at the mouth." It may be in some configurations - perhaps a tapped horn is one of them. Maybe some pipes are too. Certainly speakers mounted on spherical cabinets are - even more than that - their acoustic centers are actually in front of the cabinet. So there are lots of possibilities for acoustic center positions - It all depends on the configuration, shape and size of the speaker system.
A horn creates an impedance transformation by matching the wavefront expansion with its physical area expansion. The cross-section of the horn forms a pie-slice that is the radiating angle. So it is already developing an expanding wavefront before it even leaves the horn.
At the horn throat, pressure is highest, and it is gradually reduced as the wavefront travels down the horn. In fact, in a conical horn, the pressure attenuation down the horn is precisely the same as a spherical wave in free-space, exactly following the inverse-square law. So you can easily see that the acoustic center of a traditional horn is pretty much where the radiating diaphragm is.
There are some extra complications, of course. The electro-mechanico-acoustic properties make the impedance vary from resistive to reactive, especially at the low end. This is the matter of phase. This is even more pronounced in an undersized horn, as is the case for most basshorns. But in groups, the horn can be made to be uncompromised, so it will act as a true, full-size horn.
So it's not so simple to say "the acoustic center is at the mouth." It may be in some configurations - perhaps a tapped horn is one of them. Maybe some pipes are too. Certainly speakers mounted on spherical cabinets are - even more than that - their acoustic centers are actually in front of the cabinet. So there are lots of possibilities for acoustic center positions - It all depends on the configuration, shape and size of the speaker system.
A tapped horn’s acoustic origin is not at the mouth, it is generally pretty close to the path length.
Many users incorrectly think they don’t need delay to align TH with front loaded cabinets.
In tapped horns, bass reflex and regular horn cabinets, all the impedance transformation the cabinet provide is finished at the mouth/baffle interface with outside air.
Your concept that the path length of a horn will change the distance from the mouth where the inverse square law applies to some further distance in front of the horn is still the concept we disagree on.
Both the Growler and the C horn I used have similar horn expansion to the 12Pi, though they have shorter path length, they are fairly long horns.
Phil’s data says both the Growler and the SRX 718 both have almost identical drop off rates that are almost identical to the theoretical inverse square law drop off.
If the internal path length of the horn made any difference to what happens after the sound waves have left the enclosure, it would have been obvious in our tests.
Again I’ll ask If you found errors in the methodology or data presented in post #237, please point them out.
Art Welter
The acoustic center of a horn is in the throat, not at the mouth, and not some distance in front of the cabinet. If you disagree with that, you disagree with basic horn theory, not with me. This is in all of the basic acoustic texts.
The implication is that horns of various internal lengths present a path-length/acoustic-center offset. When you measure at a fixed distance, what is the point of origin you should measure from? The acoustic center is the point of origin. Determination of this point is non-trivial. But in most cases, it is certainly not the mouth.
That's why the 10M/100W measurement is so useful. It makes greater tolerance for path-length error. Even if there's several feet difference between the distance of one sub's acoustic center and another, the fact that the measurement microphone is 33 feet away swamps the differences. That's the whole point of this path-length/acoustic-center discusson and what brought us into it.
As I said in my last post, in a horn, pressure is highest at the throat, and it is gradually reduced as the wavefront travels down the horn. In a conical horn, the pressure attenuation down the horn is precisely the same as a spherical wave in free-space, exactly following the inverse-square law. So you can easily see that the acoustic center of a traditional horn is pretty much where the radiating diaphragm is.
To be a little more specific, the acoustic center is where the planar wave from the radiating piston transforms into a spherical wavefront. In a conical horn, this happens at the throat. In a quadratic or catenary (OS) horn, it happens very nearly at the throat, where the curvature bends the wave from planar to spherical. In most other horns, it happens within the horn path, usually near to the throat in a properly sized horn.
There is one condition that makes the mouth become the place where the plane wave transforms into a spherical wave. That is a straight pipe. In this case, the wavefront from the radiating piston stays planar through the entire length of the pipe, only becoming a spherical wave at the exit.
Of course, there are lots of other variables to juggle too - all the other electro-mechanico-acoustic properties that come into play. But in general, I think you'll find those conditions to be true for pipes, tapered ducts and horns.
It just isn't as simple as saying "the source is at the mouth". In some cases, this may be true, but it is rare. In a true horn, it is usually closer to the throat than the mouth. Only in a non-tapered pipe (or very mildly tapered pipe) does the mouth become the place where the wavefront transforms into a spherical wave. Even then, if the pipe is tapered at all, then mouth diffraction becomes a low frequency condition; As frequency rises, the acoustic center shifts further down the throat, because the mouth no longer functions as a diffraction slot.
Even in direct radiators, we see that the acoustic center is rarely at the baffle. In most cases, it is behind the speaker, giving lower SPL readings than expected at close range. But in some cases, like oddball cabinet shapes, it can be in other positions. In a spherical cabinet, for example, the acoustic center is in front, ahead of the baffle.
The implication is that horns of various internal lengths present a path-length/acoustic-center offset. When you measure at a fixed distance, what is the point of origin you should measure from? The acoustic center is the point of origin. Determination of this point is non-trivial. But in most cases, it is certainly not the mouth.
That's why the 10M/100W measurement is so useful. It makes greater tolerance for path-length error. Even if there's several feet difference between the distance of one sub's acoustic center and another, the fact that the measurement microphone is 33 feet away swamps the differences. That's the whole point of this path-length/acoustic-center discusson and what brought us into it.
As I said in my last post, in a horn, pressure is highest at the throat, and it is gradually reduced as the wavefront travels down the horn. In a conical horn, the pressure attenuation down the horn is precisely the same as a spherical wave in free-space, exactly following the inverse-square law. So you can easily see that the acoustic center of a traditional horn is pretty much where the radiating diaphragm is.
To be a little more specific, the acoustic center is where the planar wave from the radiating piston transforms into a spherical wavefront. In a conical horn, this happens at the throat. In a quadratic or catenary (OS) horn, it happens very nearly at the throat, where the curvature bends the wave from planar to spherical. In most other horns, it happens within the horn path, usually near to the throat in a properly sized horn.
There is one condition that makes the mouth become the place where the plane wave transforms into a spherical wave. That is a straight pipe. In this case, the wavefront from the radiating piston stays planar through the entire length of the pipe, only becoming a spherical wave at the exit.
Of course, there are lots of other variables to juggle too - all the other electro-mechanico-acoustic properties that come into play. But in general, I think you'll find those conditions to be true for pipes, tapered ducts and horns.
It just isn't as simple as saying "the source is at the mouth". In some cases, this may be true, but it is rare. In a true horn, it is usually closer to the throat than the mouth. Only in a non-tapered pipe (or very mildly tapered pipe) does the mouth become the place where the wavefront transforms into a spherical wave. Even then, if the pipe is tapered at all, then mouth diffraction becomes a low frequency condition; As frequency rises, the acoustic center shifts further down the throat, because the mouth no longer functions as a diffraction slot.
Even in direct radiators, we see that the acoustic center is rarely at the baffle. In most cases, it is behind the speaker, giving lower SPL readings than expected at close range. But in some cases, like oddball cabinet shapes, it can be in other positions. In a spherical cabinet, for example, the acoustic center is in front, ahead of the baffle.
Last edited:
Not to be argumentative, but the big boys never did measure things at 1M. They always measured at a distance and then calculated what the SPL would be at 1M.
1M is only recently a standard, before that it was 4' (in the USA). EIA sensitivity was 1mW at 30' (if I remember correctly), figures like 48dB were common.
http://archives.telex.com/archives/EV/Speakers/EDS/Sentry%20IA%20&%20IIA.pdf
1M is only recently a standard, before that it was 4' (in the USA). EIA sensitivity was 1mW at 30' (if I remember correctly), figures like 48dB were common.
http://archives.telex.com/archives/EV/Speakers/EDS/Sentry%20IA%20&%20IIA.pdf
I'm no expert with horns or horn theory, far from it, but just trying to think this through in my head - surely there is a different situation depending on the size of the mouth relative to the frequencies being reproduced, because this will affect what happens to the expansion rate of the wave after it leaves the mouth.
For example, if you are reproducing frequencies whose wavelength are small relative to the mouth size, such as midrange on a large mouth conical horn, the wave will have expanded at a constant rate all the way from the source as you describe, but then when it leaves the confines of the horn, at those high frequencies the wave will not bend around the edge of the horn and radiate at 360 degrees to illuminate a full sphere, instead it will continue more or less on its way at the same expansion rate illuminating the same area of a sphere that the physical geometry of the horn dictated.
In this case you are right - inverse square law fall off begins at the acoustic centre where the diaphragm is, because there has been a constant expansion of the wavefront throughout the horn through which continues more or less unchanged after exiting the mouth. (A CD waveguide I assume would have a similar effect as the wave exits the mouth, although the apparent acoustic centre might be slightly offset from the driver since it's not a conical expansion)
However what happens at bass frequencies ? As large as your 12 pi sub is, the mouth is only 1.14m high x 0.71m wide (sorry, I can only think in metric 😀 ) which is still relatively small compared to bass wavelengths. Even as high as 150 Hz the longest dimension of the mouth is still only a half wavelength, let alone at much lower frequencies like 30-50Hz, where the mouth dimensions are a very small fraction of a wavelength.
I would argue that over most of the bass region as soon as the wave exits the mouth it will immediately expand out in 360 degrees in both axes (ignoring for the moment ground effects) to illuminate a full sphere, because of the omni-directional nature of bass frequencies being reproduced by what is still essentially a small source relative to the wavelengths.
If all the wavefront expansion occurred at the mouth - as in the example you gave of of a straight pipe with the woofer at the far end, then the mouth of the pipe is unambiguously where inverse square law expansion must begin.
In the case of your 12 pi, although there is significant expansion of the wavefront within the horn (after all if there weren't, it wouldn't be a horn) I would argue that an even larger expansion of the wavefront in a bass horn of this size must occur as the wave exits the mouth.
At the very least there is going to be a transition from 2pi to 4pi at the mouth, as the opening of the front face of a cabinet can't by definition be any wider than half space. If the solid angle of radiation at the mouth is less than 2pi the transition to 4pi at the mouth will be an even greater increase in expansion.
This surely must bring the effective beginning point of inverse square law fall off well forward of the driver, but still slightly behind the mouth. If the majority of the total expansion to reach 4pi space occurred at the mouth, this would put the effective point of inverse square law fall off quite close to but slightly behind the mouth.
In other words saying that inverse square law fall off effectively begins at the driver on a low frequency horn is wrong, but so is saying that it begins exactly at the mouth - it must be somewhere in between, depending on the horn flare rate and its dimensions relative to the wavelength, thus it will also change with frequency.
Or am I missing something really obvious ? The irony is that if I'm right this is yet another reason to say that for maximum certainty of measurement accuracy its necessary to measure further away, because the effective point of inverse square law fall off is not some nice well defined easily measurable physical location like the driver diaphragm or the mouth, but rather a virtual point somewhere between the two, depending not only on horn taper and dimensions, but also on test frequency.
Thus, direct comparisons of two systems with largely different horn tapers and/or sizes at a close distance could be significantly in error due to uncertainty about where the effective beginning of inverse square law fall-off is. This is in addition to the other things already discussed - like the fact that you don't want to measure close enough to be in the near-field where the driver/mouth is too large to be a point source. (Since inverse square fall-off only asserts itself fully when you're far enough away for the source to approximate a point...)
Everything still points towards measuring more than a few metres away for physically large pro-sound woofers to get certainty and accuracy in measurement, particularly when comparing radically different designs against each other.
I don't think anyone is suggesting that home DIY'ers should be measuring the bass response of their creations at 10 metres. The whole discussion is centred around rather large pro-sound bass systems, the type of systems which you might see a dozen of at an outdoor concert, and which present significant challenges to measure compared to what you or I might have in our living rooms. We're talking a couple of orders of magnitude difference here.
A few factors come into play here - In general, the larger a speaker or speaker system is, the further away you have to measure it to maintain accuracy, for many reasons, especially multi-way systems.
The bigger drivers are, the further away you have to be for them to approximate point sources, the larger cabinets are and driver spacing is, the further away you have to be before the different drivers "blend" properly, and so on. A small 2 way book shelf with a 5" woofer can be measured accurately at one metre, a large floor standing 3 way with a 12" woofer cannot.
Then there is the dynamic issue. If you're designing a home hi-fi system your microphone can probably handle the maximum output of the system at 1 metre without clipping or compression. (if its decent) With a huge pro-sound system designed for outdoor use, probably not 😉 If you want to measure the dynamic performance such as compression and distortion at very high levels, you'll need to get further away.
If you want to measure bass response in a home hi-fi design, near-field measurements are probably your best bet, if you're using a direct radiator system.
It's relatively easy to do a close mic'ed measurement of the woofer and a port, (if any) scale the two responses based on cone/port areas, then adjust for baffle step, and get a pretty accurate response up to 200Hz. You don't even need to move the speaker from it's normal playback position as room effects are reduced down to almost a line thickness on the graph...
A near-field measurement is not something that is applicable to a horn-sub though.
It's really important to realize though, that although speaker efficiency (actually sensitivity, as they are not the same thing) is nowadays quoted at 2.83v / 1 metre, that doesn't automatically mean that it should be measured at that distance regardless of driver type or size.
The calculated sensitivity figure of a driver at 1 metre is based on the premise that the driver approximates a point source. For example if you measure the T/S parameters of a driver you can then calculate the reference efficiency, "no", and from that calculate the sensitivity, (SPL at a given distance on axis) assuming radiation into half space, and if you use 1 metre you'll get the typical sensitivity figure. A lot of software such as WinISD Pro will do this for you, which is quite convenient.
However, all such calculations assume the driver is a point source with inverse square law fall off, but when you get close to a large driver this is not the case. If you were to measure a 1" tweeter at 1 metre your measurement would be very accurate. If you were to measure a 15" woofer at 1 metre you'd be in error.
To get a sensitivity figure which is correct regardless of driver size, you need to measure it "far enough" away based on it's size and the required accuracy, then calculate back to a 1 metre SPL figure.
Depends what you're measuring and how. If you're talking about a large 3 way system then yes, measure it at 2-3 metres instead of 1 metre, although you now have the problem that if you're doing gated measurements, your usable window size gets smaller the further away the microphone is for a given height off the ground - no free lunches here unfortunately.
This another of those arguments which could go on and on and get nastier as well. It seems to me that Wayne Parnham and Weltersys are actually talking about two different things and if that is the case there should be no argument at all.
Wayne is discussing acoustic centres of various enclosures and to a layman like me what he is saying makes sense. Weltersys is saying that a FLH and a box type enclosure can both be measured the same way for comparison purposes with valid results. Again, to me, his figures from tests that he or somebody else has made show this is so, in those cases.
Gentlemen, I would say you both seem to be correct in what you each discuss,therefore I suggest you shake on it and let it go as they are not the same thing.
jamikl
Wayne is discussing acoustic centres of various enclosures and to a layman like me what he is saying makes sense. Weltersys is saying that a FLH and a box type enclosure can both be measured the same way for comparison purposes with valid results. Again, to me, his figures from tests that he or somebody else has made show this is so, in those cases.
Gentlemen, I would say you both seem to be correct in what you each discuss,therefore I suggest you shake on it and let it go as they are not the same thing.
jamikl
- Status
- Not open for further replies.