Hey guys.
I'm glad the Horn Honk thing caught your interest - but woudn't it be better to discuss it over in This Thread?
There seems to be enough going on here, already.
I'm glad the Horn Honk thing caught your interest - but woudn't it be better to discuss it over in This Thread?
There seems to be enough going on here, already.
Opinion on Gunness work?
Would you like to comment on Gunness Mr. Geddes? EAW approach
Would be interesting.
Would you like to comment on Gunness Mr. Geddes? EAW approach
Would be interesting.
Getting back to the topic of OBs with previously stated objectives....
Do you think that an adequatley chosen 21'' could deliver proper dynamics / linearity down to -- say -- upper mid 30s without the need of a sub ? Can the WR integration and driver coloration be addressed by e.g. proper driver selection ?
I am thinking about the aforementioned 21LW1400 physically decoupled from the OB and driven by a plate amp with a 4th order LR LPF set to about 200Hz. Suggestions for other readily avail 21'' highly appreciated (the Executioner-X21 sure looks nice, but sadly unobtainum for those on the other side of the pond).
Lynn, I know you stated in the beginning of this thread -- and multiple times after that -- that the last two octaves would be driven by a Rythmik or similar sub, but such less ambitious design variant might appease the "middle of the road" crowd with smaller rooms, smaller budgets and lower SPLs reqrmnts...
That is, in case you care about them / such a design offshoot, of course...
Lynn Olson said:
Do you think that an adequatley chosen 21'' could deliver proper dynamics / linearity down to -- say -- upper mid 30s without the need of a sub ? Can the WR integration and driver coloration be addressed by e.g. proper driver selection ?
I am thinking about the aforementioned 21LW1400 physically decoupled from the OB and driven by a plate amp with a 4th order LR LPF set to about 200Hz. Suggestions for other readily avail 21'' highly appreciated (the Executioner-X21 sure looks nice, but sadly unobtainum for those on the other side of the pond).
Lynn, I know you stated in the beginning of this thread -- and multiple times after that -- that the last two octaves would be driven by a Rythmik or similar sub, but such less ambitious design variant might appease the "middle of the road" crowd with smaller rooms, smaller budgets and lower SPLs reqrmnts...
That is, in case you care about them / such a design offshoot, of course...
It's a pretty simple allocation between dynamics vs bass depth. To expand a bit (I'd illustrate this if my big G5 computer were up and running, but it is still awaiting repair), a closed-box monopole speaker has a 12 dB/octave increase in excursion vs frequency - above Fs. It's this area-under-the-curve that limits dynamics and creates the majority of the IM distortion we have to contend with.
The situation is a bit worse with dipoles - in the region between the 1/f transition and Fs, in an equalized, flat-response system excursion increases at a 18 dB/octave rate, not merely 12 dB as it is with a monopole.
The area under the curve is much greater than a monopole, and this limits dynamics and creates more distortion. The only cure is more surface area - and more Xmax, and lower distortion, don't hurt. But the biggest useful gain is in simple surface area, since it doesn't require extremely long voice coils, exotic magnetic drive, and associated lowered efficiency.
One extreme is the Bastanis, with a very high crossover (220 Hz) to a monopole bass unit. Linkwitz leans on equalization, amplifier power, H-baffles, and the lowest-distortion drivers he can find. One corner of the triangle that hasn't been as well explored is a large amount of surface area that also takes full advantage of the floor image.
The key point here is the spectral distance between the lower frequency limit (of the dipole) and the 1/f transition frequency. The greater this distance, the more surface area and/or equalization is going to be required. No free lunch here.
If the baffle is big enough for a movie theater, the 1/f frequency essentially becomes low enough to treat the speaker as an infinite baffle - literally, wall-sized. If the baffle is this petite, decorator-friendly little thing, well, it's going to need an awful lot of equalization, or a very high subwoofer crossover frequency.
Equalization has the terrible drawback of tremendous driver excursion, many times worse than an equivalent monopole system, There are very real limits to what can be achieved with exotic low-distortion drivers, or even servo feedback.
Equalized H and W-baffles increase the effective baffle size, but at a cost: restricted bandwidth, since the H and W-baffles suffer from the same annoying box modes - as well, a box, a monopole. Linkwitz sidesteps this by sharply restricting the bandwidth of the H-baffles so they cut off before reaching the first mode. If you want a compact dipole speaker with any degree of LF extension, I can't see any other solution than a narrow-bandwidth, equalized, H or W-baffle. A Linkwitz system, in other words.
The same tradeoffs are on the horn side of the fence, too. A truly low-excursion, high-dynamic deep-bass horn is the size of a single-car garage. The more typical approach transitions to a direct-radiator driver anywhere between 200 and 800 Hz, depending on whether there are midbass horns or not. A big speaker like the Klipschorn takes advantage of the room corners to expand the mouth size, and operates the bass horn right through the cutoff region.
It's not very pretty: big baffles, big cone area, big horn size, or smaller sizes and more amplifier power (and more excursion in direct proportion). The old AR-1W vs Klipschorn or E-V Patrician tradeoff.
Dipoles make this tradeoff worse, not better. It's Hoffman's Iron Law with an extra twist. At some point you have to stop swimming upstream and turn around. The "turn-around" is a subwoofer of some type, either heavily equalized W-baffles or less heavily equalized monopoles.
Our friend johninCR has come up with a novel, transitional solution, combining a partial open-baffle with something like a TQWT. A split-path rear wave, in other words. No measurements so far, but he reports good results.
The situation is a bit worse with dipoles - in the region between the 1/f transition and Fs, in an equalized, flat-response system excursion increases at a 18 dB/octave rate, not merely 12 dB as it is with a monopole.
The area under the curve is much greater than a monopole, and this limits dynamics and creates more distortion. The only cure is more surface area - and more Xmax, and lower distortion, don't hurt. But the biggest useful gain is in simple surface area, since it doesn't require extremely long voice coils, exotic magnetic drive, and associated lowered efficiency.
One extreme is the Bastanis, with a very high crossover (220 Hz) to a monopole bass unit. Linkwitz leans on equalization, amplifier power, H-baffles, and the lowest-distortion drivers he can find. One corner of the triangle that hasn't been as well explored is a large amount of surface area that also takes full advantage of the floor image.
The key point here is the spectral distance between the lower frequency limit (of the dipole) and the 1/f transition frequency. The greater this distance, the more surface area and/or equalization is going to be required. No free lunch here.
If the baffle is big enough for a movie theater, the 1/f frequency essentially becomes low enough to treat the speaker as an infinite baffle - literally, wall-sized. If the baffle is this petite, decorator-friendly little thing, well, it's going to need an awful lot of equalization, or a very high subwoofer crossover frequency.
Equalization has the terrible drawback of tremendous driver excursion, many times worse than an equivalent monopole system, There are very real limits to what can be achieved with exotic low-distortion drivers, or even servo feedback.
Equalized H and W-baffles increase the effective baffle size, but at a cost: restricted bandwidth, since the H and W-baffles suffer from the same annoying box modes - as well, a box, a monopole. Linkwitz sidesteps this by sharply restricting the bandwidth of the H-baffles so they cut off before reaching the first mode. If you want a compact dipole speaker with any degree of LF extension, I can't see any other solution than a narrow-bandwidth, equalized, H or W-baffle. A Linkwitz system, in other words.
The same tradeoffs are on the horn side of the fence, too. A truly low-excursion, high-dynamic deep-bass horn is the size of a single-car garage. The more typical approach transitions to a direct-radiator driver anywhere between 200 and 800 Hz, depending on whether there are midbass horns or not. A big speaker like the Klipschorn takes advantage of the room corners to expand the mouth size, and operates the bass horn right through the cutoff region.
It's not very pretty: big baffles, big cone area, big horn size, or smaller sizes and more amplifier power (and more excursion in direct proportion). The old AR-1W vs Klipschorn or E-V Patrician tradeoff.
Dipoles make this tradeoff worse, not better. It's Hoffman's Iron Law with an extra twist. At some point you have to stop swimming upstream and turn around. The "turn-around" is a subwoofer of some type, either heavily equalized W-baffles or less heavily equalized monopoles.
Our friend johninCR has come up with a novel, transitional solution, combining a partial open-baffle with something like a TQWT. A split-path rear wave, in other words. No measurements so far, but he reports good results.
This Is Better
But genuinely appreciated by yrs truly.
A vivid demonstration that horn theory has plenty of controversy in it, seventy-plus years after the first theater and studio-monitor horns. It also says volumes about the sound quality of contemporary constant-directivity horns in movie theaters. I'm enough of an old-timer to prefer the sound back in the Todd-AO 70mm mag-track days.
gedlee said:
But genuinely appreciated by yrs truly.
A vivid demonstration that horn theory has plenty of controversy in it, seventy-plus years after the first theater and studio-monitor horns. It also says volumes about the sound quality of contemporary constant-directivity horns in movie theaters. I'm enough of an old-timer to prefer the sound back in the Todd-AO 70mm mag-track days.
Member
Joined 2003
First - The Gunness approach is interesting, I've read this before. The measurements get my attention the most as they do seem to show an effect that I would say could well be HOM. But I disagree that HOM can be controlled with DSP. The manner in which this is suppossedly done is not at all apparent from the write-up which is much more of a marketing document than an engineering one. EAW has never really done an engineering description of this technology.
Second - I don't have a tried and true way of measuring the HOM, but I do have a very good handle on how they are generated and hence what one can do in the design to minimize them. I have focused on minimizing them. Theoretically, the HOM would not be readily apparent in the frequency domain, but should be in the time domain, i.e. the impulse response. This is in essence what Gunness has done - looked at the frequency domain in time slices (also known as time-frequency analysis in the literature) - another form of waterfall plot, but with much better resolution in time and frequency. Resonances and HOMs will both appear as time delayed signal content, but will difer in that resonances will be minimum phase while HOMs will not. I am not sure that this difference would be readily apparent in the Gunees plots or if it is even important. I have found that a properly designed waveguide and driver (with foam plug), properly EQd, will have an almost perfect impulse response (actually a doublet since there cannot be any LF content) at any angle with virtually no tail - a very compact impulse response in time.
Second - I don't have a tried and true way of measuring the HOM, but I do have a very good handle on how they are generated and hence what one can do in the design to minimize them. I have focused on minimizing them. Theoretically, the HOM would not be readily apparent in the frequency domain, but should be in the time domain, i.e. the impulse response. This is in essence what Gunness has done - looked at the frequency domain in time slices (also known as time-frequency analysis in the literature) - another form of waterfall plot, but with much better resolution in time and frequency. Resonances and HOMs will both appear as time delayed signal content, but will difer in that resonances will be minimum phase while HOMs will not. I am not sure that this difference would be readily apparent in the Gunees plots or if it is even important. I have found that a properly designed waveguide and driver (with foam plug), properly EQd, will have an almost perfect impulse response (actually a doublet since there cannot be any LF content) at any angle with virtually no tail - a very compact impulse response in time.
Lynn Olson said:
IMHO a notable omission from that list is John Kreskovsky NaO speaker with the damped U-frame woofer. His study on the interaction btw dipole bass and in room response was a real eye opener for me, together with the problems in attempting to presurize a room below the room fundamental.
Unfortunately it seems that one has to accept the back side of the OB coin, just like you detailed (thanks again)
John's waveguides definitively match the "think outside the box" approach of this thread... Looks promising.
Again, thanks for the input. Best regards and a speedy recovery (we are all eagerly waiting for the first builds, afterall...
).Florian
I did an AES paper on the subject of monopole, dipole and cardiod responses in rooms You can see it in the journal some years back.
The paper referenced above must be in error since it shows a flat response to DC from a dipole. This is not possible as the front and rear radiation must cancel and hence the DC response must go to zero. I am surprised that this was missed.
The paper referenced above must be in error since it shows a flat response to DC from a dipole. This is not possible as the front and rear radiation must cancel and hence the DC response must go to zero. I am surprised that this was missed.
NaO Speaker
IIRC this project is basically a knockoff of the Orion John starter putting some damping materials on the mid-hi panels, it seems he is trying a different woofer arrangement than the H used by Linkwitz.
I don't think he adds anything beyond what Linkwitz does on his site.
This thread is a lot of fun to follow. Building on what Linkwitz does and developing a High Efficiency OB that used cone area instead of EQ is very interesting.
IIRC this project is basically a knockoff of the Orion John starter putting some damping materials on the mid-hi panels, it seems he is trying a different woofer arrangement than the H used by Linkwitz.
I don't think he adds anything beyond what Linkwitz does on his site.
This thread is a lot of fun to follow. Building on what Linkwitz does and developing a High Efficiency OB that used cone area instead of EQ is very interesting.
Earl,
very interested to have you here. A few questions:
1- if HOM's can't be measured in a straightforward way but only posited (and removed) by mathematical reasoning, how do you correlate HOMs' influence to listener experience? Do you have perceptual data on this?
2- how do you differentiate between possible HOM and possible mouth reflection issues causing a specific listener impression? For instance, you use throat plugs to attenuate HOMs and listener impression improves. But would such a throat plug not also attenuate a possible mouth reflection, by terminating and damping the throat end?
3- if I understand correctly, HOMs are multiple near-throat reflections caused by a wavefront that is non-perpendicular to the horn wall. Your design attempts to solve this by proper horn geometry and throat plugs. But, related to 2- , HOM's are likely to involve much shorter time-domain aberrations than mouth reflections.
ad 3-, a: Basically HOMs would occur in the narrow throat area. So say for a 25 mm throat HOMs could cause say 1 to 10 cm path length difference for a wave reflected several times in throat vicinity. That gives a time delay of 30-300 uS.
ad 3-, b: Mouth reflections OTOH would travel forth and back between throat and mouth. So for say, a 15 cm long horn, a mouth reflection going back to the throat and back out twice, would cause a minimum of 15x4=60 cm path length difference between direct and reflected sound, for a time delay of 1.7 ms. This time delay order of magnitude seems much more in line with thresholds of audibility of time delays (0.6-0.7 ms from reference in post #1018 ), than the HOM time delay order of magnitude.
The reason why I ask is to determine whether try and error time in waveguide projects should better be used on mouth termination or on throat plugs.
very interested to have you here. A few questions:
1- if HOM's can't be measured in a straightforward way but only posited (and removed) by mathematical reasoning, how do you correlate HOMs' influence to listener experience? Do you have perceptual data on this?
2- how do you differentiate between possible HOM and possible mouth reflection issues causing a specific listener impression? For instance, you use throat plugs to attenuate HOMs and listener impression improves. But would such a throat plug not also attenuate a possible mouth reflection, by terminating and damping the throat end?
3- if I understand correctly, HOMs are multiple near-throat reflections caused by a wavefront that is non-perpendicular to the horn wall. Your design attempts to solve this by proper horn geometry and throat plugs. But, related to 2- , HOM's are likely to involve much shorter time-domain aberrations than mouth reflections.
ad 3-, a: Basically HOMs would occur in the narrow throat area. So say for a 25 mm throat HOMs could cause say 1 to 10 cm path length difference for a wave reflected several times in throat vicinity. That gives a time delay of 30-300 uS.
ad 3-, b: Mouth reflections OTOH would travel forth and back between throat and mouth. So for say, a 15 cm long horn, a mouth reflection going back to the throat and back out twice, would cause a minimum of 15x4=60 cm path length difference between direct and reflected sound, for a time delay of 1.7 ms. This time delay order of magnitude seems much more in line with thresholds of audibility of time delays (0.6-0.7 ms from reference in post #1018 ), than the HOM time delay order of magnitude.
The reason why I ask is to determine whether try and error time in waveguide projects should better be used on mouth termination or on throat plugs.
gedlee said:
Dr. Geddes,
Below the room fundamental I believe a dipole is flat to DC. Down there the baffle isn't relevant and the room dominates, but since the wave no longer fits in the room the phase relationship of the front and rear waves remains constant. The problem is the sharp drop before getting there.
Florian,
My WG's were borrowing some basic concepts from the Summa. I just wanted to try a constant directivity OB using a physical control instead of the typical route of multi drivers and complex XO's, which attempt CD but fall short. I feel there's a lot of uncharted territory in physically manipulating the rear wave to make it combine however you want with the main output in the front.
I don't believe in pure dipole. If you walk all the way around any live event amplified or not, it doesn't sound the same in the rear. While live amplified music comes out of boxes, in our rooms I feel some rear output from an OB helps to add a better sense of space. When you take OB to the extreme of fully dipole, then it can't sound real because the real thing isn't dipole. The real thing sounds more like a dipole with a very shallow sloped low pass filter on the rear output or like a box with a portion of its output rearward (maybe a bipole with a similar very shallow low pass filter on the rear).
The idea to which Lynn referred is what I use in the bass region. That is using a Helmholtz slot as a filter to extract a portion of the very low frequency content of the rear wave and send it on a much longer path (about 2m). That portion is sufficiently delayed so it actually reinforces the front output, instead of cancelling it, down to below 30hz, thus offsetting much of the dipole cancellation in the low bass region.
I've been inactive in my speaker building hobby while I play around with hot air engines, but that pursuit has lead to an idea that is way way outside the box. If I can get wide enough bandwidth and maintain low distortion, it could really shake up the speaker world. It's a means of amplifying the output after is leaves the driver, giving an 3 or 4" driver the ability to put out more bass than a big subwoofer, while maintaining a very small construction. I know the process works, because I shook my whole house with a 4" driver, a few litres of construction, a 60hz tone (I'm trying to output 60hz AC), and only a fraction of a watt input to the driver. The question is whether I can make it more than just a "one noter".
johninCR said:
This sentence is difficult to understand. Here is a thought experiment, perhaps you can point out where it differs from the reality you have in mind.
Take a driver that we know we can hear at some non-zero frequency. Put it in a perfectly sealed box, in a sealed room. At any frequency below the box and room resonances the response must be flat - a given current in the coil will lead to the same motion of the cone and same presure change in the room, independent of frequency.
Now put the same driver on an open baffle - I'm baffled as to how it can pressurise the room as the air just goes from front to back. The response surely is that of a zero at zero frequency.
I know rooms are not sealed etc. which is why the response of a sealed box does fall off towards DC in reality, but I don't see how that helps a dipole.
Apologies in advance if I misconstrue your meaning.
Ken
The "pressure mode" of a room is directly excited by the net volume displacement (or volume velocity) of the source, which for a dipole is zero.
Perceptive questions on HOMs:
The subjective aspects were studied using a simulation of an HOM on a wave file that was auditioned over headphones. Lots of subjects in a double blind protocal. To the extent that the "model" was an acurate representation of the HOM (this I have medium confidence in), the results are applicable to the discussion here. They were reported at the last AES and are completely consistant with results reported by Toole and Moore in similar studies.
HOM's do not travel along the axis of the waveguide, but at angles to it. No waves travel perpendicular to the walls - thats the boundary condition.
The foam plug does in fact attenuate both internal reflections as well as HOMs (which are also internal reflections but not normal to the throat and mouth). So the use of the foam plug does not only affect the HOMs as you suggest. Thats why the test noted above was performed - to sort out one type of aberation from another.
The difference between a reflection and an HOM - which is a form of diffraction - is that the reflection is minimum phase while the diffraction is non-minimum phase. This means that the peak energy of a reflection arrives at the same time as the main signal (there will be a tail), but the peak energy of the HOM and/or diffraction arrives delayed in time (plus a tail). This is a small but very distinct difference, especially when subjective aspects are taken into account. The ear is far less sensitive to the minimum phase aberation than the nonminimum phase one, but more significantly, the minimum phase one is mostly level independent while the nonminimum phase one is highly level dependent. The ear masks the diffraction (HOM) more at lower SPL levels than at higher ones. This makes diffraction sound like nonlinear distortion in that it "distorts" (subjectively) higher SPL waveforms while not affecting lower level ones.
This has a profound impact to our understanding of "distortion" in that a linear system that has diffraction will sound like a nonlinear system. Its the ear that is nonlinear here, not the system.
Your data on the perception of delays is consistant with typical understanding, but is contradicted (more like refined actually) by several studies, one of which is my own. We dealt with delays on the order of .2 ms and found substantial subjective effects, but typically only at higher SPLs. What you need to factor in here is that most studies of delay perception are done at a single SPL (usually fairly low in fact) and it is well know from Toole and Moore that these effects are level dependent. .2 ms is quite audible on a loud passage, but inaudible on a softer one. This aspect has been overlooked in virtually all studies of the perception of reflections and diffraction. Moore noted the effect, but never went further in his investigation. Toole noted it also, but only for non-musical signals and never followed up with a study of level as a principle variable. Our recent paper did three variables - SPL level, delay time, and delayed signal level. Our results were completely consistant with all previous results where they overlaped, but we found the strongest effect with SPL, the variable which others have virtually always held constant.
Mouth refelctions are coherent with the main signal, which is why they are minimum phase. They will not be strongly level dependent, while the HOMs will be.
Factor all of this into your analysis and I think that you will come up with the same conclusions that I did in my designs.
Thanks for the great questions.
Perceptive questions on HOMs:
The subjective aspects were studied using a simulation of an HOM on a wave file that was auditioned over headphones. Lots of subjects in a double blind protocal. To the extent that the "model" was an acurate representation of the HOM (this I have medium confidence in), the results are applicable to the discussion here. They were reported at the last AES and are completely consistant with results reported by Toole and Moore in similar studies.
HOM's do not travel along the axis of the waveguide, but at angles to it. No waves travel perpendicular to the walls - thats the boundary condition.
The foam plug does in fact attenuate both internal reflections as well as HOMs (which are also internal reflections but not normal to the throat and mouth). So the use of the foam plug does not only affect the HOMs as you suggest. Thats why the test noted above was performed - to sort out one type of aberation from another.
The difference between a reflection and an HOM - which is a form of diffraction - is that the reflection is minimum phase while the diffraction is non-minimum phase. This means that the peak energy of a reflection arrives at the same time as the main signal (there will be a tail), but the peak energy of the HOM and/or diffraction arrives delayed in time (plus a tail). This is a small but very distinct difference, especially when subjective aspects are taken into account. The ear is far less sensitive to the minimum phase aberation than the nonminimum phase one, but more significantly, the minimum phase one is mostly level independent while the nonminimum phase one is highly level dependent. The ear masks the diffraction (HOM) more at lower SPL levels than at higher ones. This makes diffraction sound like nonlinear distortion in that it "distorts" (subjectively) higher SPL waveforms while not affecting lower level ones.
This has a profound impact to our understanding of "distortion" in that a linear system that has diffraction will sound like a nonlinear system. Its the ear that is nonlinear here, not the system.
Your data on the perception of delays is consistant with typical understanding, but is contradicted (more like refined actually) by several studies, one of which is my own. We dealt with delays on the order of .2 ms and found substantial subjective effects, but typically only at higher SPLs. What you need to factor in here is that most studies of delay perception are done at a single SPL (usually fairly low in fact) and it is well know from Toole and Moore that these effects are level dependent. .2 ms is quite audible on a loud passage, but inaudible on a softer one. This aspect has been overlooked in virtually all studies of the perception of reflections and diffraction. Moore noted the effect, but never went further in his investigation. Toole noted it also, but only for non-musical signals and never followed up with a study of level as a principle variable. Our recent paper did three variables - SPL level, delay time, and delayed signal level. Our results were completely consistant with all previous results where they overlaped, but we found the strongest effect with SPL, the variable which others have virtually always held constant.
Mouth refelctions are coherent with the main signal, which is why they are minimum phase. They will not be strongly level dependent, while the HOMs will be.
Factor all of this into your analysis and I think that you will come up with the same conclusions that I did in my designs.
Thanks for the great questions.
I find this a very interesting thread! Oddly it was about 1.5 years ago, given of take that I started really becoming interested in OB speakers, Lynn’s Karna amp and about the same time visited Dr. Geddes’ home to hear his Summas in their natural habit. Now here we are mixing them altogether. What does this have to do with the discussion? Nothing, just a humorous factum.
Having heard the Summa at Dr. Geddes’ home, I can fully attest to the lack of horn-honk exhibited by the Summa. His speaker was also the first to demonstrate to me the dynamic capabilities of hi-eff speakers. I’ve had my wheels spinning ever since, but haven’t actually created anything yet.
My background is in statistics and I did a masters in financial mathematics. I am familiar with Dr. Taleb from the financial mathematics text I own of his. I am also quite familiar with his criticisms of financial practitioners such as I, and the work of many behavioral financiers on the role of human psyche in investing and other activities. But what can I say, I am an empiricist.
I think it is true also in the DIY community that many/most of us tend to be empiricists. Not in the sense of throwing s*** against the wall to see what sticks, but letting our ears (and our psyche?) guide our studies. What we will accept as proof of concept would undoubtedly fall short in any academic circle, but that isn’t the point either.
I know I am quite content to spend money and effort on building electronics, such as the amps of Lynn’s design, because my experience is proof enough for me of the existence of amp differentiation.
I could be pulling the wool over my own eyes, but who cares? I am satisfied and proved it to myself enough to be satisfied. I however, don’t have faith in my ability to discern every minutia some say they can. That is ok too. I also don’t have as much faith in my short term discernment ability as I do in my more prolonged listening evaluations.
I, like so many others here, do not have the time, energy, patience, funding and desire to collect statistical data to prove each of our beliefs. That would simply take too long. Instead many of us simply use a more empirical approach to design. I like to control where I can and measure where I can, but lacking those, it ultimately comes down to hearing and deciding, flawed as that may be.
Just as I’ve heard things about your Summas that I really enjoyed and inspired me, I’ve heard things done by dynamic dipoles that I’ve not heard other speakers do. I just try to observe as much as possible, keep an open mind and an inquisitive mind on the how and why’s, loosely grounded in theory and let that be my guide.
I understand where you objected to Lynn’s subjective comments on your RMAF demonstration but I believe, given the audience here, and Lynn knowing his audience, he implicitly made a discernment that most here believe that electronics do make a difference and moreover, *typically* mass-market electronics are the bottom of the barrel.
I believe many of us have experienced a weak correlation between price and performance in electronics. I certainly have witnessed many exceptions, but most of the outliers were high priced failures rather than low priced stars.
Having heard the Summa at Dr. Geddes’ home, I can fully attest to the lack of horn-honk exhibited by the Summa. His speaker was also the first to demonstrate to me the dynamic capabilities of hi-eff speakers. I’ve had my wheels spinning ever since, but haven’t actually created anything yet.
My background is in statistics and I did a masters in financial mathematics. I am familiar with Dr. Taleb from the financial mathematics text I own of his. I am also quite familiar with his criticisms of financial practitioners such as I, and the work of many behavioral financiers on the role of human psyche in investing and other activities. But what can I say, I am an empiricist.
I think it is true also in the DIY community that many/most of us tend to be empiricists. Not in the sense of throwing s*** against the wall to see what sticks, but letting our ears (and our psyche?) guide our studies. What we will accept as proof of concept would undoubtedly fall short in any academic circle, but that isn’t the point either.
I know I am quite content to spend money and effort on building electronics, such as the amps of Lynn’s design, because my experience is proof enough for me of the existence of amp differentiation.
I could be pulling the wool over my own eyes, but who cares? I am satisfied and proved it to myself enough to be satisfied. I however, don’t have faith in my ability to discern every minutia some say they can. That is ok too. I also don’t have as much faith in my short term discernment ability as I do in my more prolonged listening evaluations.
I, like so many others here, do not have the time, energy, patience, funding and desire to collect statistical data to prove each of our beliefs. That would simply take too long. Instead many of us simply use a more empirical approach to design. I like to control where I can and measure where I can, but lacking those, it ultimately comes down to hearing and deciding, flawed as that may be.
Just as I’ve heard things about your Summas that I really enjoyed and inspired me, I’ve heard things done by dynamic dipoles that I’ve not heard other speakers do. I just try to observe as much as possible, keep an open mind and an inquisitive mind on the how and why’s, loosely grounded in theory and let that be my guide.
I understand where you objected to Lynn’s subjective comments on your RMAF demonstration but I believe, given the audience here, and Lynn knowing his audience, he implicitly made a discernment that most here believe that electronics do make a difference and moreover, *typically* mass-market electronics are the bottom of the barrel.
I believe many of us have experienced a weak correlation between price and performance in electronics. I certainly have witnessed many exceptions, but most of the outliers were high priced failures rather than low priced stars.
johninCR said:
So you have measurements to show what is happening with your concept? Any data to show Earl and the rest of us?
This is your belief?
This is supported by measurements? Could you share them with the DIY community?
I was hoping your recent inactivity was tied to being so busy measuring all your previous concept/ideas/claims. Apparently this wasn't the case?
John, if you are a fan of Earl's work, perhaps you've seen this wonderful free work he has shared with us
http://www.gedlee.com/downloads/Chapter4.pdf
cheers,
AJ
Dr. Geddes,
So, with such fine delay measurements that are measurably audible, what does this indicate for OB speakers? Is it possible that the audibility of specific delay is frequency dependant as well? Not just SPL dependant?
Also, what were the subjective effects? How did it change the sound? Is that what creates a "flattened" perception of reproduced sound? Is this what causes the lack of depth and width that is inherent to live performance? Is it a lack of 3 dimensional wave propulsion and propagation?
Last question; how much of this discussion is a result of the recording equipment's inability to properly capture the dynamics of the real performance? It's almost like we're trying to get the speaker to make up for what the recording inefficiencies have created. Is this the real culprit?
TEH
By the way; I'm sure I can speak for all of us and say it is such a pleasure to have all of you professional designers/engineers contributing.
Lastly, I don't think any of you or "us" should have any illusion here. Lynn really has no need for us. I think all of this has to be viewed as a brain storming process that allows him to open up to different possibilities. That is the most admirable trait of a "learner". It is the pinnacle or your profession; to finally realize that you still know less than there is yet to learn.
I'm just feeling so fortunate to be able to see this discussion, right here, right now. What a learning experience!
Teh said:
Yes, frequency is another aspect that is quoted in the literature and that we studied, but I did not describe. There tends to be a peak of perception of short delays arround 4 kHz. Below 1 kHz. the perception falls off fast to the normally quoted times for room reflections.
Subjective effects - well here I am just hypothesizing because we didn't do this in or study. But, I suspect, in accordance with Blauert that these very short time delays affect localization more than timbre. How this relates to your subject terms I am not sure.
One thing that I have noticed (subject comments only here) is that Summas are far more telling of bad recordings than other speakers. In fact they can make a really bad recording sound really bad. They don't mask a thing - if the tank is dirty and the fish are dead, you can smell it (see my other posts on objectives)!
But a really clean well done recording is like a live performance - listen to Linda Ronstat on "Adiue False Heart" for an example of an exemplary recording. Unfortunately there just aren't that many really good recordings and very few from years ago. The equipment just wasn't that good, and you can hear it. Poor loudspeakers mask a lot of bad habits and bad equipment. Quite honestly, I suspect that older recordings probably sound better on older designs because the tended to mask the errors of the day. But a good modern recording on a first class loudspeaker (OK and amp!!) is far better than anything that I heard years ago. - Again my own opinion here.
AJ, while your questions about John's work are reasonable ones (I'd like to see some of those measurements myself), I think it would be helpful to use a more professional tone.
I'm skeptical of the idea that one can (at this point) derive a useful correlation from rear-wave sound and listener satisfaction. I would point to my own analogy elsewhere about trying to image an object with a limited number of light sources (whether point sources, cardioid, or dipolar) in a room with mirrored walls. But if there's real data indicating otherwise, I'm... ummmm... all ears.
I'm skeptical of the idea that one can (at this point) derive a useful correlation from rear-wave sound and listener satisfaction. I would point to my own analogy elsewhere about trying to image an object with a limited number of light sources (whether point sources, cardioid, or dipolar) in a room with mirrored walls. But if there's real data indicating otherwise, I'm... ummmm... all ears.
One more thing:
As I study the detrimental effects of diffraction I can't help but be concerned with the diffraction at the baffle edge of an OB.
Has anyone ever tried using a donut shape with the piston at the hole opening. This is a dipole but with no (well low) edge diffraction. - Remember you heard it here!!
As I study the detrimental effects of diffraction I can't help but be concerned with the diffraction at the baffle edge of an OB.
Has anyone ever tried using a donut shape with the piston at the hole opening. This is a dipole but with no (well low) edge diffraction. - Remember you heard it here!!