Frank,
In my experience Cinema sound varies from dreadful to excellent, with most in the mediocre range. Sometimes different screens in the same complex have wildly different sound quality.
The nearest multiplex to us has absolutely awful sound - way too loud, (even for someone who likes loud music) and an extremely ragged tipped up frequency response. (The treble is tipped up at least 6dB from the bass in my estimation, to make it extremely "bright")
I have no idea what speaker designs they're using, but to me it sounds exactly like extreme uncontrolled cone breakup in the upper midrange 2-5Khz. If you were to measure the high frequency response I can guarentee it would look like a mountain range with a dreadful CSD.
THAT's what makes your ears bleed. I actually take my iPod ear buds with me so I can put them in to attenuate the level and cut down the upper mids a bit to stop my ears bleeding, but I still don't enjoy the experience.
At the other extreme is our local IMAX which has absolutely fantastic sound quality. No tipped up treble, very good tonal balance through the range with good bass and treble extension, NO cone breakup harshness in the upper midrange at all. Loud but effortless, smooth, unfatiguing to listen to, in fact I would go so far as to say that the sound quality is as good as a very good home system even on music, but with the power for a movie experience.
So clearly it CAN be done right, and I doubt that the IMAX uses all direct radiators...
I suspect it comes down to simple economics and the problem that the general public just don't care about sound quality.
As long as the explosions and shouting are loud they're happy...
Last edited:
Dr. Geddes,
I understand you're feeling misunderstood and hurt in a way.
I am not taking sides and personally I could not care less for
HOM's or horn issues and implementing pro drivers in boxes.
I guess being European I will never fully grasp the idea why anyone
would want to buy or build such sensitive boxes in general for home
use naturally.
My speakers are 84 dB/ 2,83 V and with an amp of 100W/4 ohms
I am pretty much able to get permanent hearing damage in my listening
room if I choose so, not to mention local authorities knocking on my
door. LOL
Anyway, by looking at your web site, I did notice one thing. Taking into
account all what it takes to build such products, you are not a greedy person.
Other manufacturers base their pricing policy on a formula
total price= (cost of parts and labour)*7 or even much higher.
I wish you success and happiness in this life and in the next one.
I've heard plenty of pro sound installations sound bad too. What do we conclude from that?
If 99 systems out of 100 sound bad, is it the EQ? The level they push? Over compressed trailers? Lack of maintenance? I see those issues all the time.
On the other hand, if 1 system out of 100 sounds good, that is proof that the hardware is capable.
Only a person with a real axe to grind would arrive at the conclusion that: "I heard bad sound at the theater today. See, diffraction is evil."
That's not science.
David
Can we all agree that a delayed reflections up to 2-3ms are especially detrimential to fidelity as has been known for decades (from studio research by SynAudCon group) and recently proven by Lee and Geddes that sensitivity of our hearing to those signals increases with playback level?
So why is it that far of a stretch to conclude, that designing devices to have the most compact impluse response (log squared, ETC, whatever) not just in one angle, but throughout the coverage angle, as has been done by some, is A Good Thing?
This thread is about measuring one type of phenomena that, besides other effects, causes the impulse to not be compact (essential to our hearing). Dr Geddes has shown the method how to measure those and Makarski has done it.
That besides HOMs there are other causes of non-compact impulse (mouth diffraction, cabinet edge diffraction, diffraction from speaker stand!, diffraction from nearby television, coffee table, and so on...) doesn't say that we should discard HOMs as a source of audible coloration. On the contary. Excellent results are obtained when we consider ALL effects that cause audible coloration (and due to insufficient research in psychoacoustics all effects that most probably could cause audible coloration).
((What has been shown, is that a Le Cleach horn gives the most compact impulse. But it does so in a rather narrow angle and the respnse changes with angle (beaming). Good for non-environment listening room, not so good where room ambiance or multiple seating is needed. Diffraction horns (especially 1st generation, EV, Mantaray, Bi-Radial, etc) , are not smooth! Evident in the impulse response and rippleiy off-axis plot. Ripples change with angle - thats how to most easly spot internal reflections, HOMs))
So why is it that far of a stretch to conclude, that designing devices to have the most compact impluse response (log squared, ETC, whatever) not just in one angle, but throughout the coverage angle, as has been done by some, is A Good Thing?
This thread is about measuring one type of phenomena that, besides other effects, causes the impulse to not be compact (essential to our hearing). Dr Geddes has shown the method how to measure those and Makarski has done it.
That besides HOMs there are other causes of non-compact impulse (mouth diffraction, cabinet edge diffraction, diffraction from speaker stand!, diffraction from nearby television, coffee table, and so on...) doesn't say that we should discard HOMs as a source of audible coloration. On the contary. Excellent results are obtained when we consider ALL effects that cause audible coloration (and due to insufficient research in psychoacoustics all effects that most probably could cause audible coloration).
((What has been shown, is that a Le Cleach horn gives the most compact impulse. But it does so in a rather narrow angle and the respnse changes with angle (beaming). Good for non-environment listening room, not so good where room ambiance or multiple seating is needed. Diffraction horns (especially 1st generation, EV, Mantaray, Bi-Radial, etc) , are not smooth! Evident in the impulse response and rippleiy off-axis plot. Ripples change with angle - thats how to most easly spot internal reflections, HOMs))
I guess mostly for home cinema surround stuff
I was just now looking at an old pair Decca Ribbon type horns
that has always been a strange upright thing, but hey, maybe they actually did get it right, without knowing
An odd statement. From my observation, many Europeans are quite enthusiastic about horn speakers.
To me it is not entirely about sensitivity, not at all actually. It is about directivity control. There is a one to one relationship between how narrow one can control directivity and the physical size of the source. To get a narrow controlled directivity down to a reasonably low frequency requires a large source. Large sources tend to be very efficient which also has its advantages.
Yes, for my own uses I agree, I'm looking for controlled directivity, not necessarily exceptional sensitivity. I don't even listen particularly loud.
Member
Joined 2003
Earl,
I realize you aren't going to do it but, if you set out to measure HOMs, what process/procedure would you use?
Perhaps measurement could become a community effort.
A couple of clarifications here.
First, I did not start this thread, but being the expert on HOMs I thought it advisable to comment.
I am not an experimentalist, I am a theorist. I do not have a lab or any real way to do measurements. I can clear my living room of furniture and make some measurement in there from time to time, but I need to be in and out fast (I do not live alone!!) I cannot set up to do ongoing research. So there is no way that I am personally going to ever make measurements of HOMs. People who use this fact to claim I am dodging the issue are simply being disingenuous.
Someone asked "Why do we care about HOMs". I CARE because without them we cannot have a detailed mathematical description of how horns work. They exist, and are always present, that much is irrefutable and has been proven by Markarski. I showed how they had to exist mathematically and he showed that they do in fact exist. So let's put that part to rest.
How audible are they? I personally have never made any claims on this either way. I showed how HOM like signals have some very unique characteristics that meld very well with subjective aspects of horns. This is not proof, I have never claimed it was (physics can never prove something is true, it can only prove something is not true. No one has proven that what many hear as poor sound in a horn, and many many do, is NOT HOMs.)
Diffraction in a horn and HOMs are not different things, they are the same thing. When a wave propagates down a horn and reaches a discontinuity two things happen. The first is that some of the wave is reflected from this junction back down the device and impinges on the diaphragm. It many cases we can see this effect in the impedance curve. The second is that the wave going forward is diffracted. Now this diffraction has two components. The first is in the shape of the main mode and we simply call this the transmitted wave. It would not be proper to call this portion of the wave "diffraction". However, the diffraction does generate waves that do not move down the axis of the device but at angles relative to it. These waves are, by definition HOMs. So you can see that diffraction and HOMs are really the same thing - they are not separable in either measurement or theory. The reflected wave is separable but yet it still gets mixed up with the main wave and the diffracted wave after these leave the device.
But this still isn't a complete picture because HOMs exhibit cutoff. This is when the wavenumber of the diffraction is imaginary and the wave decays exponentially rather than propagates. These are called evanescent waves. Depending on the distances involved these waves can completely dissipate by the time the wave exits the device, but that is not guaranteed. Some of these evanescent waves could reach the mouth of a short device.
Diffraction at the mouth is unique in that the diffracted wave is now in free space and as such all wavenumbers are allowed. So mouth diffraction is a kin of HOMs, but being as they are in free space they do not exhibit cutoff. The reflected portion of the mouth propagates back down the device and can sometimes be seen in the impedance curve. A compression driver has to have two peaks, but any more than that is proof that diffraction has occurred in the device somewhere.
If the wavenumber is high enough, then it becomes real and this wave will propagate down the device reaching the mouth and propagate out into space, but it will always follow behind the main wave in time - a tail if you will. But this is also true of the reflected wave that moves back down the device after reflecting off of the diaphragm. The delay times would be quite different however.
It is my opinion that only the main non-diffracted or reflected wave is desirable. All other aspects are aberrations and undesirable. How do we weight the relative audible effects of them? That is completely unknown and would be a massively complicated problem to take on. Hence my approach is and always has been to reduce all of these effects as I simply do not know which is which and how audible any of them are. I do believe, and there is no doubt in my mind about this, that reducing them all yields a far preferable device. Again, this cannot be proven, but no one has proven that it is not true either. All of the evidence says that it is true. None that I know of refutes it.
And all that IS science.
First, I did not start this thread, but being the expert on HOMs I thought it advisable to comment.
I am not an experimentalist, I am a theorist. I do not have a lab or any real way to do measurements. I can clear my living room of furniture and make some measurement in there from time to time, but I need to be in and out fast (I do not live alone!!) I cannot set up to do ongoing research. So there is no way that I am personally going to ever make measurements of HOMs. People who use this fact to claim I am dodging the issue are simply being disingenuous.
Someone asked "Why do we care about HOMs". I CARE because without them we cannot have a detailed mathematical description of how horns work. They exist, and are always present, that much is irrefutable and has been proven by Markarski. I showed how they had to exist mathematically and he showed that they do in fact exist. So let's put that part to rest.
How audible are they? I personally have never made any claims on this either way. I showed how HOM like signals have some very unique characteristics that meld very well with subjective aspects of horns. This is not proof, I have never claimed it was (physics can never prove something is true, it can only prove something is not true. No one has proven that what many hear as poor sound in a horn, and many many do, is NOT HOMs.)
Diffraction in a horn and HOMs are not different things, they are the same thing. When a wave propagates down a horn and reaches a discontinuity two things happen. The first is that some of the wave is reflected from this junction back down the device and impinges on the diaphragm. It many cases we can see this effect in the impedance curve. The second is that the wave going forward is diffracted. Now this diffraction has two components. The first is in the shape of the main mode and we simply call this the transmitted wave. It would not be proper to call this portion of the wave "diffraction". However, the diffraction does generate waves that do not move down the axis of the device but at angles relative to it. These waves are, by definition HOMs. So you can see that diffraction and HOMs are really the same thing - they are not separable in either measurement or theory. The reflected wave is separable but yet it still gets mixed up with the main wave and the diffracted wave after these leave the device.
But this still isn't a complete picture because HOMs exhibit cutoff. This is when the wavenumber of the diffraction is imaginary and the wave decays exponentially rather than propagates. These are called evanescent waves. Depending on the distances involved these waves can completely dissipate by the time the wave exits the device, but that is not guaranteed. Some of these evanescent waves could reach the mouth of a short device.
Diffraction at the mouth is unique in that the diffracted wave is now in free space and as such all wavenumbers are allowed. So mouth diffraction is a kin of HOMs, but being as they are in free space they do not exhibit cutoff. The reflected portion of the mouth propagates back down the device and can sometimes be seen in the impedance curve. A compression driver has to have two peaks, but any more than that is proof that diffraction has occurred in the device somewhere.
If the wavenumber is high enough, then it becomes real and this wave will propagate down the device reaching the mouth and propagate out into space, but it will always follow behind the main wave in time - a tail if you will. But this is also true of the reflected wave that moves back down the device after reflecting off of the diaphragm. The delay times would be quite different however.
It is my opinion that only the main non-diffracted or reflected wave is desirable. All other aspects are aberrations and undesirable. How do we weight the relative audible effects of them? That is completely unknown and would be a massively complicated problem to take on. Hence my approach is and always has been to reduce all of these effects as I simply do not know which is which and how audible any of them are. I do believe, and there is no doubt in my mind about this, that reducing them all yields a far preferable device. Again, this cannot be proven, but no one has proven that it is not true either. All of the evidence says that it is true. None that I know of refutes it.
And all that IS science.
I think that is a very fair statement. Besides frequency response, which must remain the primary consideration, and frequency response off axis (important when audience coverage is needed), then achieving a compact and clean impulse response is a very worthy goal.
The trouble is that the discussion always drifts to HOMs, when measurements, such as the ones by Le Cleac'h that I referenced earlier, show big evidence of mouth reflections, and minor evidence of diffraction bend reflections. Yet HOMs are hard to identify in these measurements.
Why then do we think, beyond being technical curiosities, that HOMs are important?
If directivity is important then the smoothest and most consistent polars to date have been achieved by horns with narrow diffractions slots and modified conical flares. If you want a little more response smoothness and lowered internal reflections you can open and soften the diffraction slots, as JBL has done on their most recent generation of devices.
Tossing it all out to prevent the possibility of HOMs isn't supported by the measurements. Likewise, if we are talking about a linear phenomena (a frequency or time response error, independent of level) then it doesn't make sense that there would be subjective effects from HOMs that came near the much greater, easily measurable effects of conventional reflections.
David
Hi Paul
Makarski notes that the directivity is not strongly affected by the HOMs. Still there has to be an effect. This is where I would first look, but you would need extremely accurate measurements of the polar response because you are looking for a needle in a haystack. None of the polar responses that I have seen are anywhere near accurate enough and even those that I do have not shown to be accurate enough either (and mine are the most accurate that I have seen.)
But lets say that accuracy of the polar response can be solved and I am sure it could by someone with the resources. Then you would need to decompose the measured polar response into its "mode" noting what was due to the main mode and what was due to the HOMs. That requires that you have a coordinate system that allows for this separation. I tried this some years back with a spherical system, but it didn't work out too well and it was difficult to get a giant sphere with a horn embedded in it. This is how I do my polar maps now so it was not a total loss and I can propagate any wave front back to a sphere and look at what is happening but the resolution is not high enough to resolve the details.
Lately I have been working on the modal decomposition of the circular aperture in an infinite baffle. This is doable, albeit not "done". But the potential gotchas are that no baffle is "infinite" - how much will this affect the results? Hard to know. Can I get a baffle "big enough"? I don't know. I was hoping to work with Mr. Kolbrek on this but I think that he got too busy.
If someone built a big baffle, say 8 foot square, put a round waveguide in it and did polar measurements at say every 2.5 degree all the way from the normal to the baffle (90 degrees) at 1 meter, then I would have some data to work with, but no guarantees that anything would come of it. This is hardly a trivial task for not much guarantee - that's what research is all about. No company is going to do this - none that I know of at least. We started this at B&C but the project was never competed. Would have been interesting.
Oh and by the way, the modes that one gets from the flat round aperture are not the same modes as inside of the waveguide. So one has to find a transform from one set to the other. This is common in physics, but numerically how stable it is another unknown. With a spherical baffle the modes in the device are the same as the modes on the sphere so there is no transform required. That's why I thought of the sphere at first. Anybody got an eight foot sphere?
The waveguide and driver would have to be installed flush into this sphere, and the sphere would have to be exactly the correct radius (or a new waveguide built). Then the sphere would have to be somehow supported away from any other surfaces by at least 4 meters (hung from a ceiling would be great), and rotated to make the measurements. SO the minimum room size looks like it would be either outdoors or about 10 m x 10 m x 10 m. No dimension could be smaller but any could be larger. You could move the mic, but this is never as accurate as rotating the source.
You could do all that?
But sure, if you can do all that, I can send you a waveguide - that's the easy part!!
I looked into renting an indoor soccer stadium, but they do not have much downtime, are very expensive and you would have to be in and out in one to two hours.
You could do all that?
But sure, if you can do all that, I can send you a waveguide - that's the easy part!!
I looked into renting an indoor soccer stadium, but they do not have much downtime, are very expensive and you would have to be in and out in one to two hours.
In my above description I forgot to mention that HOMs could originate from the driver itself and according to Makarski this is the major source. I have no trouble believing that for a waveguide in which an attempt has been made to minimize the inherent internal diffraction of the connect horn. But I cannot see that being the case where there is intentional diffraction along the waves path. An easy test is to look at the impedance curve. If you see any sign of more than two peaks then there is internal diffraction and one cannot conclude that the source is the major culprit. But if the system, as mine do, has only two peaks with no signs of any other peaks (< 10 kHz) then it is reasonable to assume that the driver could be the major source because driver generated HOMs would not show any impedance aberrations.
Specifically: "(important when audience coverage is needed)", I'd say that may be understating things. Consider that most environments could be assumed to have progressively (though peaky) increasing attenuation vs frequency (not to mention simple propagation loss), constant directivity(or at least relatively constant) up to 10kHz is key in ensuring that you don't entirely lose tonal balance, both in terms of the contribution of the room response off-axis, but also in it's perceived contribution to tonality on-axis. Above 10k... I'd say it's fair to say all bets are off, with highly lossy environments and less hearing acuity in this range.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.