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. The final use of the cab and the listeners to it, do not care what extra distance the drivers may have inside one cab over another. If you set a direct radiator of the same enclosure shape and size next to a FLH and put a listeners chair 10m away, you would not move the chair 3m closer to the FLH to make it a "fair" or "equal" listening distance since the drivers are further in the enclosure. Place the listener chair at 8m,5m,2m or even 1m it does not matter. You would not move the subwoofer closer or move the seating in a large venue or arena because one sub has the drivers further in the cabinet, why would you do that with a microphone for performance measurements? That makes absolutely no logical sense to me. What matters is the acoustic performance that arrives at the listening position after the sound leaves the cabinet.
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. The final use of the cab and the listeners to it, do not care what extra distance the drivers may have inside one cab over another. If you set a direct radiator of the same enclosure shape and size next to a FLH and put a listeners chair 10m away, you would not move the chair 3m closer to the FLH to make it a "fair" or "equal" listening distance since the drivers are further in the enclosure. Place the listener chair at 8m,5m,2m or even 1m it does not matter. You would not move the subwoofer closer or move the seating in a large venue or arena because one sub has the drivers further in the cabinet, why would you do that with a microphone for performance measurements? That makes absolutely no logical sense to me. What matters is the acoustic performance that arrives at the listening position after the sound leaves the cabinet.
Maybe there is some confusion between an apparent acoustic source and an acoustic center, which are two different things.
The concept of an apparent acoustic source is useful in analysis of directivity and polar charts. It is usually used in context of an HF horn with a diffraction slot, which, when measured off-axis causes a re-radiated wavefront to define a new apparent source location. It is a point where the wavefront starts to curve and expand. So, for example, a horn with a diffraction slot an inch or two forward of its diaphragm, might have an apparent source that's an inch or two forward of its acoustic center, when measured off-axis. On-axis, the apparent source location is the same place as the acoustic center. Some horns are astigmatic, meaning they have a different apparent source location in the vertical and horizontal planes.
The acoustic center is approximately equal to the diaphragm location. The acoustic load and other (electro-mechanico-acoustic) reactive features can shift this forward or back somewhat, as is shown in the phase. Horns are generally resistive, so they tend to have phase that's pretty close to zero. But still, the point is that phase comes in to play and can shift the acoustic center from the physical location of the radiating diaphragm.
The acoustic center of a basshorn is behind the mouth. The internal path length sets the distance back from the mouth where the acoustic center is found. This location is the sound source distance to the listener, and is easily found by measurement. It is what is important when setting delay, to match the mains with the subs.
The concept of an apparent acoustic source is useful in analysis of directivity and polar charts. It is usually used in context of an HF horn with a diffraction slot, which, when measured off-axis causes a re-radiated wavefront to define a new apparent source location. It is a point where the wavefront starts to curve and expand. So, for example, a horn with a diffraction slot an inch or two forward of its diaphragm, might have an apparent source that's an inch or two forward of its acoustic center, when measured off-axis. On-axis, the apparent source location is the same place as the acoustic center. Some horns are astigmatic, meaning they have a different apparent source location in the vertical and horizontal planes.
The acoustic center is approximately equal to the diaphragm location. The acoustic load and other (electro-mechanico-acoustic) reactive features can shift this forward or back somewhat, as is shown in the phase. Horns are generally resistive, so they tend to have phase that's pretty close to zero. But still, the point is that phase comes in to play and can shift the acoustic center from the physical location of the radiating diaphragm.
The acoustic center of a basshorn is behind the mouth. The internal path length sets the distance back from the mouth where the acoustic center is found. This location is the sound source distance to the listener, and is easily found by measurement. It is what is important when setting delay, to match the mains with the subs.
Wayne,
As tb46, and Josh Ricci point out, what is inside the cabinet (the location of the radiating source) is irrelevant other than for setting delay times between cabinets.
As you point out, the apparent source can change in location in certain types of horns, shifting from the diaphragm to the diffraction slot in the throat.
Horn measurements are always taken at a specified distance from the mouth, not from the diaphragm or diffraction slot.
The apparent source location does not change the fact that the sound emanating from the horn falls off at the inverse square law after exiting the mouth.
At any rate, if you don’t believe the results of my and Phil’s experiment, you can easily test a front loaded cabinet and a horn loaded cabinet yourself.
You will find, as we did, that regardless of the cabinet type (as long as the cabinet frontal dimensions do not exceed the measurement distance) the low frequency output will conform to the inverse square law, as measured from the front of the cabinet .
Art Welter
Just to be clear - My position is that prosound subwoofer comparisons be made using measurements where the microphone is placed ten meters from the mouth. That's what I suggested earlier, and I did so because I do not think comparison with measurements taken closer than that are very valid.
However, I also think it is important to understand that accurately identifying the position of the true acoustic source isn't as simple as defining it to be the mouth. We can all agree to measure at ten meters from the mouth, and in fact, that's what I personally have always done. But I do believe it is important to understand that the acoustic center is rarely actually at the front face of the loudspeaker cabinet. There are many reasons I think it is important to understand that. Path length averaging for accurate SPL measurements is just one of them.
The differences in acoustic centers between various enclosure types make the most difference in measurements taken close to the mouth. An example is the one I mentoned earlier, where a microphone is placed two meters away from the cabinet. In a direct radiator or a tapped horn, the microphone will be two meters away from the acoustic center, but in a horn with three meter path length, putting the microphone two meters away from the cabinet results in it being five meters away from the acoustic center. Even if we all agree to measure some equal distance from the mouth, it is important to realize that the distance to the acoustic center is different in each situation.
- A note on the concept of acoustic center, by Jacobsen, Figueroa and Rasmussen
Last edited:
12Pi Vs Xoc1 TH18
Another interesting comparison;
12Pi = 929Litre
Vs
Xoc1-TH18 loaded with a B&C 18SW115 = 393,54Litre
That makes the Xoc1 42% of the volume of the 12Pi. Comparison is based on maximum displacement from both drivers which makes the power compression figures superfluous. While the Xoc1 is significant lower in sensitivity it shows how the B&C driver can make up by its much higher displacement.
Another interesting comparison;
12Pi = 929Litre
Vs
Xoc1-TH18 loaded with a B&C 18SW115 = 393,54Litre
That makes the Xoc1 42% of the volume of the 12Pi. Comparison is based on maximum displacement from both drivers which makes the power compression figures superfluous. While the Xoc1 is significant lower in sensitivity it shows how the B&C driver can make up by its much higher displacement.
Respectfully, I think you're still missing the point. I still get the feeling your assumption is that with more power, the efficiency difference can be made up. But I don't see it. I mean, if your proposed woofer could handle ten times the power without compression, maybe then you'd have something. But you're talking about what amounts to pretty small differences in power levels, and pretty large differences in efficiency levels. The two just don't add up.
Wayne, I truly believe you are honestly convinced and not just marketing your product, even when everybody can see all your links to your commercial website that are questionable. It’s not I’m missing your point(s), but it looks like you don’t feel comfortable by analysing step by step. You can’t take out efficiency and forget about max displacement, power compression, power ratings and all other arguments that limit your 12Pi system. The total SPL is the sum of all factors involved of which efficiency is one of them, and yes it’s an important one.
About the efficiency of TH’s, so far you haven’t proved anything other then the principles of FLH that nobody was or is questioning. That’s the reason why nobody is willing to respond to that. You also haven’t proved the TH’s can’t deliver. By stating the horn principles only work in FLH’s because your 12Pi uses a bigger horn mouth, doesn’t prove anything about the TH besides it’s not a traditional horn and that is old news.
Meanwhile, other members here have proven, their 1w/1m measurements did line up with their 100w/10m measurements. They also proved the differences in sensitivity in conjunction with directional radiation. Also many members showed that their measurements do line up pretty well with the predictions from HornResp. Even if their measuring methods aren’t correct in the eye of 'scientifically accepted methods', there is still enough evidence you can’t push aside by qualifying it as non-prove or not relevant to the discussion. Besides it is disrespectful, you could only qualify it IF you prove otherwise, which you haven’t.
That doesn’t mean the 12Pi is a bad design, on the contrary. I stated that from the beginning and I’ll repeat it again. I really think if you use the same working principles in combination with drivers that can deliver the maximal potential of today’s innovations; it would be a very difficult cab to beat in stacks.
Last edited:
Frankly, I'm saddened to see this kind of post. I've put a LOT of R&D resources - both time and money - into the 12Pi hornsub design. It's a sophisticated basshorn, and offers a high level of performance for those that need it. And all the development detail, history and supporting documentation is out there for everyone to see. It isn't just Hornresp models, although those are there, along with the history of development that preceded the design, and extensive testing results after the fact.
So anyway, my replies:
Woah, now what exactly is questionable? The 12Pi hornsub plans are available to anyone for the asking, and all the support documentation is online. It's as much a DIY project as any of the others here, it's just been really pushed to the max with respect to design and testing.
What seems much more questionable to me are the people that have come into this thread - where someone asked about the 12Pi hornsub - promoting their own designs, some even openly offering them for sale. Most of these subs have just been barely developed, and are nowhere near mature, having only basic models or very rudimentary measurements.
I'm not sure what the rules are on this site, but I would expect they at least say something to the effect that if you hijack a thread about one design to shill your own stuff, that's in bad form.
The same could be said for you, but the facts of the physics fall sharply on my side of the argument. You can't focus solely on excursion or thermal limits and forget about efficiency. Efficiency is very important for max SPL, arguably more important than any other metric. If you have one design that provides 40% efficiency and another that provides only 10% efficiency, it's really hard to make up the difference with higher-power woofers. It's not like we're talking woofers with a ten-fold power difference, we're talking maybe 25%. That's nothing. You have to double the power to get even a 3dB increase, which still doesn't cover much ground when compared with a system having significantly higher efficiency. The lower efficiency system wastes it's extra power in heat loss.
Plenty have responded, but if anyone takes the position that a small box with very little taper can match the efficiency of an uncompromised basshorn, then they're on the wrong side of the argument. It's not a matter of tapped horn verses traditional horn - it's a matter of flare rate and mouth size large enough to become uncompromised in a group. If you optimize the horn so that it can be used this way, you can achieve maximum efficiency. If you don't, you can't.
This is physics, basic horn theory 101. If you study the material, you'll see that this is not controversial stuff. There should be no disagreement here.
I have seen no 10 meter measurements of any of the other boxes promoted in this thread as an alternative for the 12Pi hornsub, so how would you know who has "proven" what?
I'll tell you this, it's a rare hornsub that measures the same SPL at 1M with 1W input as it does at 10M with 100W input. The numbers are all over the place at this close distance, depending on the characteristics of the sub. They're all different. The only way the 1W/1M measurement will equal the 10M/100W measurement is if the acoustic center is at the face, which is rare.
The 1W/1M values and the 10M/100W values are considered to be equivalent because they calculate back to being equivalent, but as the link in my last post shows, the close-mic'ed measurements are much more likely to be in error. A 10M/100W measurement is much more accurate because it reduces the acoustic-center/path-length error. One meter and two meter measurements of a hornsub are basically worthless for SPL reference levels.
Honestly, I don't know whether to thank you or punch you because your "compliments" are peppered with accusations and innuendoes. I understand that English isn't your first language, so I'll let it go at that. But I will assert that the 12Pi is an advanced design, using cooling that is vastly superior to any of the drivers you've mentioned, because of its use of cooling plugs which wick the heat out of the core. None of the drivers you've mentioned have that feature, and so do not cool the core as well.
Further, your Hornresp models are not even close to "proof". I trust Hornresp a lot, and depend on it for initial development of every horn I make. But the models just take what you feed them, and until you have a battery of acoustic measurements of a physical design with real wood and woofers - done at 10 meters distance using calibrated measurement equipment - then you cannot come in this thread and compare them with the tests I have done.
So anyway, my replies:
Woah, now what exactly is questionable? The 12Pi hornsub plans are available to anyone for the asking, and all the support documentation is online. It's as much a DIY project as any of the others here, it's just been really pushed to the max with respect to design and testing.
What seems much more questionable to me are the people that have come into this thread - where someone asked about the 12Pi hornsub - promoting their own designs, some even openly offering them for sale. Most of these subs have just been barely developed, and are nowhere near mature, having only basic models or very rudimentary measurements.
I'm not sure what the rules are on this site, but I would expect they at least say something to the effect that if you hijack a thread about one design to shill your own stuff, that's in bad form.
The same could be said for you, but the facts of the physics fall sharply on my side of the argument. You can't focus solely on excursion or thermal limits and forget about efficiency. Efficiency is very important for max SPL, arguably more important than any other metric. If you have one design that provides 40% efficiency and another that provides only 10% efficiency, it's really hard to make up the difference with higher-power woofers. It's not like we're talking woofers with a ten-fold power difference, we're talking maybe 25%. That's nothing. You have to double the power to get even a 3dB increase, which still doesn't cover much ground when compared with a system having significantly higher efficiency. The lower efficiency system wastes it's extra power in heat loss.
Plenty have responded, but if anyone takes the position that a small box with very little taper can match the efficiency of an uncompromised basshorn, then they're on the wrong side of the argument. It's not a matter of tapped horn verses traditional horn - it's a matter of flare rate and mouth size large enough to become uncompromised in a group. If you optimize the horn so that it can be used this way, you can achieve maximum efficiency. If you don't, you can't.
This is physics, basic horn theory 101. If you study the material, you'll see that this is not controversial stuff. There should be no disagreement here.
I have seen no 10 meter measurements of any of the other boxes promoted in this thread as an alternative for the 12Pi hornsub, so how would you know who has "proven" what?
I'll tell you this, it's a rare hornsub that measures the same SPL at 1M with 1W input as it does at 10M with 100W input. The numbers are all over the place at this close distance, depending on the characteristics of the sub. They're all different. The only way the 1W/1M measurement will equal the 10M/100W measurement is if the acoustic center is at the face, which is rare.
The 1W/1M values and the 10M/100W values are considered to be equivalent because they calculate back to being equivalent, but as the link in my last post shows, the close-mic'ed measurements are much more likely to be in error. A 10M/100W measurement is much more accurate because it reduces the acoustic-center/path-length error. One meter and two meter measurements of a hornsub are basically worthless for SPL reference levels.
Honestly, I don't know whether to thank you or punch you because your "compliments" are peppered with accusations and innuendoes. I understand that English isn't your first language, so I'll let it go at that. But I will assert that the 12Pi is an advanced design, using cooling that is vastly superior to any of the drivers you've mentioned, because of its use of cooling plugs which wick the heat out of the core. None of the drivers you've mentioned have that feature, and so do not cool the core as well.
Further, your Hornresp models are not even close to "proof". I trust Hornresp a lot, and depend on it for initial development of every horn I make. But the models just take what you feed them, and until you have a battery of acoustic measurements of a physical design with real wood and woofers - done at 10 meters distance using calibrated measurement equipment - then you cannot come in this thread and compare them with the tests I have done.
10M/100W measurements
Very good observation & points 🙂
Indeed, & also helps to average out ALL contributions from vents etc, which can be more than one, & placed in various positions away from the driver/s.
Except when having to "attempt" comparing against others measured that way. So until the "industry" agrees on 10M/100W measurements, it looks like we're stuck with 1M's !
Very good observation & points 🙂
Indeed, & also helps to average out ALL contributions from vents etc, which can be more than one, & placed in various positions away from the driver/s.
Except when having to "attempt" comparing against others measured that way. So until the "industry" agrees on 10M/100W measurements, it looks like we're stuck with 1M's !
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. Granted I have only measured 3 tapped horns and no FLH's yet but I have measured some very large cabs. There was not a dramatic difference in any cab yet. You mention a 1 m measurement of a FLH basshorn actually being closer to 5m to the acoustic center while a front loader is 1m. Why do we care about this with respect to performance as used? Are you suggesting that we place the microphone 2 meters inside the FLH to compare with the direct radiator? The response and loading of the horn on the drivers output isn't even developed yet or incubated fully. What matters is the response at the microphone at whatever distance is being used from the cab and that the measurements between different cabs are relative and under the same conditions.
If intrepid sound provider A is comparing new subs to put in a venue, sub 1 a 6th order bandpass and sub 2 a FLH both of which are 22.5x45x36 and he knows that the meat of his primary audience is located at 10m from the subs why does he care that there is 4m of extra path to the drivers inside the FLH other than for setting delay times? In what way should this information be used in the comparison of the cabs? It does not physically change the placement of the subs or the audience. 😕 Isn't what we care about the performance that the cab delivers to the listener/microphone. I simply do not understand why we should care about the extra path length inside the cab when looking at performance applied at the listener position. I also do not understand why when the horn sub must have this extra length to work its magic to begin with. The horn affect on the driver output doesn't mature fully and take shape until it reaches the mouth and exits the cab anyway.
If intrepid sound provider A is comparing new subs to put in a venue, sub 1 a 6th order bandpass and sub 2 a FLH both of which are 22.5x45x36 and he knows that the meat of his primary audience is located at 10m from the subs why does he care that there is 4m of extra path to the drivers inside the FLH other than for setting delay times? In what way should this information be used in the comparison of the cabs? It does not physically change the placement of the subs or the audience. 😕 Isn't what we care about the performance that the cab delivers to the listener/microphone. I simply do not understand why we should care about the extra path length inside the cab when looking at performance applied at the listener position. I also do not understand why when the horn sub must have this extra length to work its magic to begin with. The horn affect on the driver output doesn't mature fully and take shape until it reaches the mouth and exits the cab anyway.
Last edited:
To anyone that believes the position of the acoustic center is unimportant, or that you can accurately measure SPL of any radiator shape/configuration at any arbitrary distance, no matter how small, I encourage you to study the matter further.
It is best to measure at a fairly large distance, I'd say on the order of the wavelength scale, at least, in order to reduce the acoustic-center/path-length error. You can get a 1W/1M value at distances greater than 1M, you just have to calculate back. That is an accepted practice.
If you measure at too short a distance, your results will be ambiguous, unless you know with a great degree of accuracy the position of the acoustic centers both of the DUT and of the measurement microphone.
That's why the longer distance works best, it makes the requirement for accuracy less stringent, sort of a tolerance deal. Instead of needing accuracy to the centimeter, like it would be for a close measurement, a more distant measurement allows some tolerance. The distance between the (acoustic centers of the) source and receiver might only need to be known within a meter or two, for example. That then allows for each device to be measured from mouth to microphone, without really having to care about the position of the acoustic centers. At a large distance, where the error could only amount to something under 1dB, it ceases to be an issue.
That's why it is best to measure hornsubs at 10 meters, and that's why a 1 meter measurement is pretty much useless. I mean, it would be fine if you knew precisely the position of the acoustic centers of both the DUT and the microphone.
I'm afraid I find little point in discussing this further - It's an academic matter that I think is probably not well served by continued "argument" here. But I will again reference one of many documents written on the subject, for those that might be interested.
It is best to measure at a fairly large distance, I'd say on the order of the wavelength scale, at least, in order to reduce the acoustic-center/path-length error. You can get a 1W/1M value at distances greater than 1M, you just have to calculate back. That is an accepted practice.
If you measure at too short a distance, your results will be ambiguous, unless you know with a great degree of accuracy the position of the acoustic centers both of the DUT and of the measurement microphone.
That's why the longer distance works best, it makes the requirement for accuracy less stringent, sort of a tolerance deal. Instead of needing accuracy to the centimeter, like it would be for a close measurement, a more distant measurement allows some tolerance. The distance between the (acoustic centers of the) source and receiver might only need to be known within a meter or two, for example. That then allows for each device to be measured from mouth to microphone, without really having to care about the position of the acoustic centers. At a large distance, where the error could only amount to something under 1dB, it ceases to be an issue.
That's why it is best to measure hornsubs at 10 meters, and that's why a 1 meter measurement is pretty much useless. I mean, it would be fine if you knew precisely the position of the acoustic centers of both the DUT and the microphone.
I'm afraid I find little point in discussing this further - It's an academic matter that I think is probably not well served by continued "argument" here. But I will again reference one of many documents written on the subject, for those that might be interested.
- A note on the concept of acoustic center, by Jacobsen, Figueroa and Rasmussen
Last edited:
Wayne what is the acoustic center of your 12pi? Would it not be near the mouth of the horn not back inside near the drivers? The entire point of this discussion hinging on whether you can get an accurate 1 or 2m measurement of a large FLH. 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. If this is not what you are implying with the acoustic center being different in FLH's then I have gotten the wrong idea and I apologize. Do you have a 10m/100w measurement and also a 1w/1m of a horn sub to compare?
You mention that it is accepted practice to back calculate from a 10m 100w measurement to a 1w/1m one, yet imply that the reverse is not true and 1 or 2m measurements are flawed for hornsubs. If your back calculated 1w/1m result does not match what a real 1w/1m measurement would be within reason then what is the point or value in it as it is not the real result? If it does match well then the back calculation is valid and reverse calculation from 1m to 10m is also valid.
You mention that it is accepted practice to back calculate from a 10m 100w measurement to a 1w/1m one, yet imply that the reverse is not true and 1 or 2m measurements are flawed for hornsubs. If your back calculated 1w/1m result does not match what a real 1w/1m measurement would be within reason then what is the point or value in it as it is not the real result? If it does match well then the back calculation is valid and reverse calculation from 1m to 10m is also valid.
As I said, when you measure at a large distance, the acoustic-center/path-length error is reduced. It becomes insignificant at large distances.
It's a matter of percentages. If the path length error is +/-2M but the measurement distance is 10M, the most you can be off is 20%. That's only about 1.5dB, which is not too bad. At most, the error would be 1.5dB, one way or the other. But if the measurement distance is only 2M, then you can be off by as much as 100%. That would give a window of accuracy of +/-6dB, which is not very good.
That's why it is best to measure hornsubs at 10 meters, and that's why a 1 meter measurement is pretty much useless. I mean, it would be fine if you knew precisely the position of the acoustic centers of both the DUT and the microphone. But even without knowing the precise positions of the acoustic centers, you can get pretty good accuracy as long as the measuement distance is great enough.
It's a matter of percentages. If the path length error is +/-2M but the measurement distance is 10M, the most you can be off is 20%. That's only about 1.5dB, which is not too bad. At most, the error would be 1.5dB, one way or the other. But if the measurement distance is only 2M, then you can be off by as much as 100%. That would give a window of accuracy of +/-6dB, which is not very good.
That's why it is best to measure hornsubs at 10 meters, and that's why a 1 meter measurement is pretty much useless. I mean, it would be fine if you knew precisely the position of the acoustic centers of both the DUT and the microphone. But even without knowing the precise positions of the acoustic centers, you can get pretty good accuracy as long as the measuement distance is great enough.
I guess I will leave it there.
I have simply not noticed a large variation from the expected inverse law in practice with 1m to 10m measurement comparisons but as I said I have not measured any FLH's myself yet.
I have simply not noticed a large variation from the expected inverse law in practice with 1m to 10m measurement comparisons but as I said I have not measured any FLH's myself yet.
I understand, but perhaps your view is (naturally) guided by your experience. You said you were usually measuring the same kinds of enclosures, which probably have a similar characteristic with respect to the position of the acoustic center relative to the face.
You might be interested in that paper I referenced in my earlier post. It's kind of an important concept, if your goal is to get accurate SPL/w/m measurements. That's not the only paper - there are lots of them - but that one is pretty short and concise, and illustrates the issue with some examples.
I have measured a lot of direct radiating cabinets and a lot of hornsubs. They're definitely different. In fact, sometimes, the 1M and 2M measurements are so far off, you would think the equipment was all out of whack. But then move out a distance and the numbers come more in line.
I measure up close for low-power distortion figures, for amplitude response (when accurate SPL isn't really needed) and for finding the position of lobes and nulls. Even those will seem to move some at distances, but a lot of my stuff is for home hifi anyway. That's a different situation than prosound. I'm generally more interested in how it will act at one or two meters, maybe three or four, but almost never at 10 meters. Still, I know that if I want an accurate SPL/w/m figure, I prefer to measure at some distance because of the accuracy issue.
If you measure enough, you'll see some weird numbers at close distance. One speaker is pretty much the same at 1W/1M as it is at greater distance. So obviously its acoustic center is pretty close to the baffle. Another is much lower than expected, so its acoustic center is further back. Occasionally you'll even find speakers that are louder than expected at 1M.
What gets complicated is when the acoustic centers of two subsystems (like woofer and tweeter) are real far apart. Then you have the situation that you can match the sensitivities at some distance, but then move closer to the speaker and the driver with the closer acoustic center becomes louder. I try to design speakers that don't act that way when used for hifi, because they'll always be used with the listener relatively close to the speaker cabinet.
You might be interested in that paper I referenced in my earlier post. It's kind of an important concept, if your goal is to get accurate SPL/w/m measurements. That's not the only paper - there are lots of them - but that one is pretty short and concise, and illustrates the issue with some examples.
I have measured a lot of direct radiating cabinets and a lot of hornsubs. They're definitely different. In fact, sometimes, the 1M and 2M measurements are so far off, you would think the equipment was all out of whack. But then move out a distance and the numbers come more in line.
I measure up close for low-power distortion figures, for amplitude response (when accurate SPL isn't really needed) and for finding the position of lobes and nulls. Even those will seem to move some at distances, but a lot of my stuff is for home hifi anyway. That's a different situation than prosound. I'm generally more interested in how it will act at one or two meters, maybe three or four, but almost never at 10 meters. Still, I know that if I want an accurate SPL/w/m figure, I prefer to measure at some distance because of the accuracy issue.
If you measure enough, you'll see some weird numbers at close distance. One speaker is pretty much the same at 1W/1M as it is at greater distance. So obviously its acoustic center is pretty close to the baffle. Another is much lower than expected, so its acoustic center is further back. Occasionally you'll even find speakers that are louder than expected at 1M.
What gets complicated is when the acoustic centers of two subsystems (like woofer and tweeter) are real far apart. Then you have the situation that you can match the sensitivities at some distance, but then move closer to the speaker and the driver with the closer acoustic center becomes louder. I try to design speakers that don't act that way when used for hifi, because they'll always be used with the listener relatively close to the speaker cabinet.
Agreed.
There's not just differences in the path length to the acoustic centre to worry about either, (when comparing bass horn vs bass reflex for example, where the acoustic centre would be unfairly closer to the mic than a basshorn at "1 metre") most people seem to be completely overlooking the fact that inverse square law does not hold once you get "too close" to a driver.
Inverse square law fall off with distance only applies to a point source.
When you're a long way from a driver (many times it's diameter) it approximates a point source, but as you get close to a circular/square driver it starts to approximate an infinite plane source, where in theory there is no fall off in level with distance. (And an infinitely long line source would show a 3dB change per distance doubling, compared to 6dB for a point source)
If we measured an imaginary bass diaphragm which was 5 metres in diameter, at 50cm from the middle of the diaphragm, and then again at 1 metre, we'd find virtually no difference in level - certainly no 6dB inverse square law reduction, despite the distance doubling.
This is the same reason that "near-field" frequency response measurements of bass drivers works so well - if you measure 5mm away from the middle of a 12" woofer, you'll get almost exactly the same result as at 10mm, because the size of the driver relative to the measurement distance approximates the "no fall-off" characteristic of an infinite plane source.
If inverse square law was in effect this close to the diaphragm there would be far too much uncertainty to the absolute SPL with small changes in distance to make it accurate. (Not to mention induced 2nd harmonic distortion in the measurement due to SPL change on in and out half cycles)
Without this gradual transition from point source to infinite plane source, it would also imply that the SPL at the driver diaphragm was an asymptote to infinity, which it clearly isn't. 😉
This is also an issue when designing a multi-way system with large differences in driver diameter - although all drivers will converge on inverse square law at some far enough distance, larger drivers will start to deviate from inverse square law first as you get closer.
Measure an 8" woofer and a 1" dome tweeter at 5 metres (near enough to infinity for those driver sizes) and adjust the networks so their levels at the crossover frequency are perfectly matched, then measure again from just 0.5m and the woofer output won't have increased as much as the dome - it will appear to have less output at the crossover frequency than the tweeter, even if you measure both drivers separately on axis to avoid off axis errors.
It's because the dome is still approximating a point source more closely than the woofer at that distance, and the woofer is deviating significantly from inverse square law. Any measurements taken "too close", relative to the size of the largest driver will be in error, in addition to any baffle step effects.
If we were designing a crossover, would we design it so the drivers drive levels were correctly matched at 0.5 metres or at infinity ? The latter obviously, because (a) we don't listen at 0.5 metres, and (b) the rate of fall-off of the two drivers will converge after some distance when both drivers are small enough to approximate a point source, after this they will track.
For the same reason, when doing comparative SPL measurements of two bass systems that have different cone areas/sizes, greatest accuracy will be achieved by measuring at a relatively far distance, (the further the better, as long as good SNR is achieved) and calculating back to a theoretical 1m SPL figure by assuming an inverse square law change.
Again - with a huge bass bin nobody is going to listen to it at one metre, so the true inverse square fall-off and SPL at a distance is what matters, not the SPL at one metre, which as a comparative figure is a theoretical construct anyway, as it's too close for a large bass bin to be a true point source.
As you mention, measuring further away also dramatically reduces any errors caused by unaccounted for acoustic centre differences - so if you truly want to accurately compare two large systems with different acoustic centre depths and/or different cone/mouth areas, you can only do so measuring at a far distance such as 10 metres.
Wayne,
Phil Lewandowski measured a FLH and a front loaded cabinet in singles and pairs, two different frequencies, six distances from one to 32 meters, a total of 90 tests.
The inverse square law would predict a 30.1 dB variance between one meter and 32 meters.
Phil’s tests average a 31.428 dB drop from 1 to 32 meters for the front loaded JBL SRX 728 units.
His tests average a 30.875 dB drop from 1 to 32 meters for the horn loaded JTR Growlers.
The difference in level change between the two cabinet types was only .553 dB over a distance from one to 32 meters, and the FLH varied from predicted by less than one dB.
My tests comparing BR and FLH gave similar results.
My field experience (over 50% outdoors) over the last forty years comparing BR to FLH with path lengths anywhere from one to three meters long also confirms Phil’s results, that is internal path length has nothing to do with the inverse square law reduction once the sound exits the horn mouth.
If this were not true, a speaker using a HF horn and a front loaded cone could only be balanced in output on axis at one distance.
From the “Conclusions” in this article:
“A note on the concept of acoustic center
IV. CONCLUSIONS
The idea of replacing a real, extended source by an
equivalent point source from which outgoing wavefronts appear
to diverge is deceptively simple, and various procedures
for determining the position of such a source give in general
different results. The most useful approach, in reciprocity
calibration of transducers as well as in testing the quality of
anechoic rooms, would seem to be the one that gives the
position from which the inverse distance law applies, as seen
from positions in the region of concern.”
For single cabinet tests with a mouth as large as the 12Pi, a two meter test would be adequate, while a one meter test may not be indicative of distant measurement due to mouth area, not driver location.
The larger the mouth area or cabinet frontal area, the more “outgoing wavefronts appear to diverge” from a single point source.
Rather than continuing your unfounded guesses and observations with no supportive data, I invite you to measure the 12Pi and front loaded cabinets with an equal frontal area outdoors at various distances, and post your results.
As you and I agree, measurement trumps simulation or conjecture.
You will find for a cabinet with a frontal area the size of the 12Pi, that two meters will give a measurement position from which the inverse distance law applies, and one meter may introduce a small error when compared to 10, or 16, or 32 meters.
Art Welter
Attachments
It is true, of course, that you can find the acoustic center with measurements. That was the point of the document I referenced. But what also is the point is the variety of loudspeaker configurations and conditions you'll find where the acoustic center is not at the front face.
You would have us believe that you have measured a lot of different loudspeaker systems, and that all of them had acoustic centers precisely at the mouth and/or baffle.
If that were the case, I could only think that either (a) you have not measured very many loudspeakers or (b) that your measurement equipment or techniques were not very good, perhaps both.
You would have us believe that you have measured a lot of different loudspeaker systems, and that all of them had acoustic centers precisely at the mouth and/or baffle.
If that were the case, I could only think that either (a) you have not measured very many loudspeakers or (b) that your measurement equipment or techniques were not very good, perhaps both.
Absolutely. The mistaken notion that SPL measurements of subwoofers can be made at 1M or any other arbitrary close distance is probably what has skewed the results some of these guys have gotten more than anything else.
Believe it or not, I have personally made thousands of indoor and outdoor sound system measurements and looked over the shoulder of others doing thousands more.
Seldom have I found the acoustic center from a time alignment standpoint precisely at the baffle of a front loaded driver, and never for a horn mouth, TH included.
I have provided large scale sound for events covering distances in excess of 200 meters outdoors and in stadiums using BR and FLH cabinets, sometimes using both as I noticed in the old EV advertisement from the December 1990 Pro Sound news, photo taken at the Minnesota State Fairgrounds.
As an aside, a stock Lab 12 outperforms the EVX-150A woofers I touted as the “best” back then, and the output of the BC18SW115 makes either of those speakers seem pale by comparison.
Some things have improved from the good old days, LOL.
The MSF was an interesting “proving ground”, although we provided the sound (and light) systems for most of the Grandstand shows, various front and horn loaded systems from Clair Brothers, Electrosound, Showco, Turbosound, Malcolm Hill, Meyers Sound, and many others were brought in to the same venue. Walking the stands (I probably put on a hundred miles a season) one could get a very good idea of how well various cabinets work singly and in array, and at some very long distances.
My job has always been to make whatever system available sound the best, and know what tool is best for the particular job, which takes a lot of measurement, using both machines and ears, and shoe leather.
The acoustic center, as measured by “time of flight” has nothing to do with the point at which the inverse square law can be applied to the output of a horn or BR sub, the comparative data in post 237 clearly shows it works from as little as one meter from the mouth.
This can be verified outdoors away from buildings using any (12Pi size and under) horn and any front loaded device tested at various distances with the SPL adjusted for both to the same level at one meter, then recording SPL at doubling of distances to 32 meters, or doing the reverse, checking level at 32 meters and working back to one meter.
An exception to the above as Simon (DBMandrake ) points out would be would be testing “an imaginary bass diaphragm which was 5 meters in diameter, at 50 cm from the middle of the diaphragm, and then again at 1 metre, we'd find virtually no difference in level “.
While I agree that testing at 10 meters is a good thing for testing subs, your implying that tests done at distances closer than that are invalid is ludicrous in the absence of actual supportive data.
If you have any actual data to support your assertion that single sub tests done at less than 10 meters are invalid, post it.
With all due respect, as they say put up or shut up.
Art Welter
Attachments
- Status
- Not open for further replies.