The difficulty Pano is having is that he has not drank the Kool-Aid.
First, it starts by not calling the horn a horn. Rather we will call it a "waveguide". Next we will invent some new terms. Relections, which can result from a refelection from an undersized mouth or from diffraction around a discontinuity in the flare, will now be called a "HOM".
Now if anyone objects to the "new view" we can ostracize them and not let them play in the sandbox.
Get with it Pano and drink the Kool-Aid.
First, it starts by not calling the horn a horn. Rather we will call it a "waveguide". Next we will invent some new terms. Relections, which can result from a refelection from an undersized mouth or from diffraction around a discontinuity in the flare, will now be called a "HOM".
Now if anyone objects to the "new view" we can ostracize them and not let them play in the sandbox.
Get with it Pano and drink the Kool-Aid.
The problem guys is that all of this has been discussed and I am not in the mood to go over and over it all again. Read what been written.
I am not a fan of the HOM bandwagon myself, especially since almost no one who has waved its flag is even remotely close to understanding what they actually are. I haven't seen anyone here truly understand what they are. Take earlier in this thread, the numerous references to HOMs being standing waves. I don't refute every false statement made around here because that would be a full time job. And I don't refute every hypothesis about the audibility of HOMs because I don't know the answers either - maybe what they are guessing at is right. I don't know, none of you know either.
I am the expert (glad someone recognized that), if the AES Journal gets a paper on horns, I review it. John Vanderkooy himself has stated (quite recently in fact) that I am the most knowledgeable acoustician in this area that he knows of. So I will not tolerate disrespect regarding things that I have researched and published on this subject.
If you want to know how to measure HOMs ask me (respectfully). If you expect me to do the measurements, well then don't hold your breath. I have never been an experimentalist, I am a theorist. I hypothesized HOMs back in the early 90's. They were proven to exist in the early 2000's, I am not going to argue any of that, its documented fact.
How audible are they? I don't know that they are, and none of you knows that they aren't either. But consider this. We look at nonlinear distortion at the 1% level all the time. DO you think that it is easy to find a 1% waveform variation in a radiating sound wave? Well its not. It's like looking for a needle in a haystack. Could an HOM be audible at 1%, well yes, we found it audible at about 20% in our test but that was only at about 80 dB. OSHA would not let us test any higher. Those results showed that the audibility increased with increased SPL. Extrapolating the results implied that they would be audible at about 1% in the range of about 100 dB.
What of this is hard to accept? That it's me saying it? That's what it sounds like sometimes.
I am not a fan of the HOM bandwagon myself, especially since almost no one who has waved its flag is even remotely close to understanding what they actually are. I haven't seen anyone here truly understand what they are. Take earlier in this thread, the numerous references to HOMs being standing waves. I don't refute every false statement made around here because that would be a full time job. And I don't refute every hypothesis about the audibility of HOMs because I don't know the answers either - maybe what they are guessing at is right. I don't know, none of you know either.
I am the expert (glad someone recognized that), if the AES Journal gets a paper on horns, I review it. John Vanderkooy himself has stated (quite recently in fact) that I am the most knowledgeable acoustician in this area that he knows of. So I will not tolerate disrespect regarding things that I have researched and published on this subject.
If you want to know how to measure HOMs ask me (respectfully). If you expect me to do the measurements, well then don't hold your breath. I have never been an experimentalist, I am a theorist. I hypothesized HOMs back in the early 90's. They were proven to exist in the early 2000's, I am not going to argue any of that, its documented fact.
How audible are they? I don't know that they are, and none of you knows that they aren't either. But consider this. We look at nonlinear distortion at the 1% level all the time. DO you think that it is easy to find a 1% waveform variation in a radiating sound wave? Well its not. It's like looking for a needle in a haystack. Could an HOM be audible at 1%, well yes, we found it audible at about 20% in our test but that was only at about 80 dB. OSHA would not let us test any higher. Those results showed that the audibility increased with increased SPL. Extrapolating the results implied that they would be audible at about 1% in the range of about 100 dB.
What of this is hard to accept? That it's me saying it? That's what it sounds like sometimes.
Michael Makarski JAES paper
Hi there P: From Earl's clew, found a reference for Michael Makarski investigates to what extent HOM's at driver mouth contribute to sound radiation... This is found in AES Preprints: AES 117th Convention 2004 and listed as 6188. Maybe a JAES member will be able to obtain a copy and post information for us at Diyaudio. regards, Michael
Hi there P: From Earl's clew, found a reference for Michael Makarski investigates to what extent HOM's at driver mouth contribute to sound radiation... This is found in AES Preprints: AES 117th Convention 2004 and listed as 6188. Maybe a JAES member will be able to obtain a copy and post information for us at Diyaudio. regards, Michael
Last edited:
I have thought about this carefully and I'm not sure you are in an honest search for measurements of HOMs at and beyond waveguide/horn mouth.
You are seemingly engaging in a gigantic troll.
In threads in which the measuring of HOMs and difficulty of doing so has been discussed going back as far as 2007, it's been made clear that due to the cost in TIME and MONEY, that Geddes is not going to do the measurements you apparently want.
Get used to it. There are important things you are missing while you are demanding the undeliverable.
You are seemingly engaging in a gigantic troll.
In threads in which the measuring of HOMs and difficulty of doing so has been discussed going back as far as 2007, it's been made clear that due to the cost in TIME and MONEY, that Geddes is not going to do the measurements you apparently want.
Get used to it. There are important things you are missing while you are demanding the undeliverable.
I have addressed the method of measurement in relation with design improvement process. You can criticize the process if you wish. But if you are looking for a way to quantify in comparative terms which horn will perform better, there is no way you will get a standard way of comparison like distortion. HOMs is a result of reflections which is a general term as I have mentioned before, the cause of reflection has to be determined looking at the design and the data together. Additionally, since these are "modes", this means there is some stored energy which is also evident looking at CSD at each measurement location as I mentioned. I have posted one set of measurements in another thread, but did not go into the discussion of HOMs simply because nobody responded in a way that they had some sort of feeling what the data meant. It would be meaningless for me to talk to myself. Even I taught in graduate course, I got 0 questions, if people are not even interested in getting their money's worth of knowledge, why would someone talk about it for free?
Beautiful. Forget evidence, forget measurement, just drink the kool-aid. If not, you're a rude old troll. 😛
At least Earl has just given a clear explanation of why he can't or won't provide the measurements. And I thank him for that, it's an honest admission. It took a lot of poking on my part (aka, rude trolling) to get there, but he stepped up and gave a good and reasonable explanation. Bravo!
Now that we have established that, might we move on to at least trying to measure HOMs? There were some rather good ideas proposed here lately. Impulse, ETC and other ideas.
Tom's story of the foam on a stick gave me an idea of my own. My Altec horns and drivers have a sharp notch around 8300 Hz. This may well be the result of reflection or HOMs causing cancellation. If stuffing foam into the throat of the horn reduces the depth of the notch more than it affects the response around it, that could be enlightening. Worth looking at, I think.
So talk to us. Show us those measurements and tell us what you see in them. I'm sure what you have to say won't fall on deaf ears. 🙂
Well, I already addressed how HOM is measured, I see no response either positive or negative. Nobody even asked why that way. Nobody responded to previous data I measured from out of a horn to inside a horn. Why would I want to repeat? If anyone is remotely interested in what I said, they would have pulled the data I posted in other horn threads and discussed it here. Most in this thread also participated in the other threads.
In our specific context, or HOM's generally?
(..I'm thinking it's the latter.)
I also didn't specify this in my previous posts, though of course the intention was specifically within our context. 😱
..right or wrong, I've always thought you explained HOM's as right angle diffraction elements, that are generated with diffraction and start their movement from the point of diffraction back to the source. If even somewhat correct, then it makes sense for Makarski to measure at the horn entrance rather than the mouth, where HOM's should be "bunched up" (..or perhaps converted to greater intensity 😕 ) - after being generated from the mouth (and to a lesser extent along any bend in the horn's length). Yeah, another who doesn't understand it either.. 😱 but at least someone that doesn't claim to. 😀
In any event, nice post! I Thank you. 🙂
I have always thought your simulations looked very impressive
but I'm getting confused now
it sounds like HOM data are relatively easy to implement in the design process 😕
Same as he said in 2007 and several times on diyAudio, since then.
Like I said, you're missing the important stuff.
Simulations were mainly to understand what might effect impulse response and directivity. Later test samples revealed the limitations of simulation. The actual process is very tedious. The technical objective is to have as smooth a curve as possible throught the all measurements inside the horn.
I've suggested this paper by JM LeCleac'h before:
http://forums.melaudia.net/attachment.php?aid=1760
Here he looks at 16 horn/compression driver combinations. Each one has a response curve, an impulse response, a 3d wavelet transform of the impulse and a group delay curve.
This give a very in depth look at the time domain performance of a variety of horns. Besides the early response that contains the broad shape of the horn's frequency response, there are all of the other delayed reflections that will come from whatever cause exists. Some reflections are seen as quite distinct echoes. For example most of the horns have a significant echo at about 3 ms. LeCleac'h states that this is the horn mouth reflection. Energy reflects from the mouth back to the throat and then out again. 3ms corresponds to the transit time of two times the horn length. In addition LeCleac'h singles out mouth diffraction as a strong secondary reflection. It is closer to the main impulse as it has a path determined by sound diffracting around the horn contour to the mouth and then coming to the listener via the slightly longer diagonal path.
Horns with discontinuities such as bends in planar sections (Manta Ray) or diffraction slots will also have reflections at the corresponding times. LeCleac'h also identifies HOM related reflections.
The strong reflections are visible in the impulse response and even relatively subtle ones are seen in the color plots. The color plot shows level by color, time on the horizontal axis and frequency on the vertical axis. A reflection of a certain time delay will look like some sort of vertical wedge to the right of the primary impulse.
Now we are told that the horns are operating in the linear part of their range. Curves were taken at 91dB and the efficiency is such that the impulse responses would probably hold true for from 5 to 105dB. As such the impulse response should fully quantify the sound of the horn, at least for a listener on axis in a dead room (and with typical horn directivity, even in a somewhat lively room). That means that the measured impulse could be used for an FIR filter and any musical waveform could be convolved with it, giving the output (the sound) that we would expect from a particular horn.
Thought of that way, we should just think of the horn as merely a device imparting multiply reflected energy to any signal. The audible effect of the device can be solely determined by the standard audibility of reflections. Reflection delay, level and frequency spectrum are all that matter (in this case sound direction is essentially the same as the primary sound).
In this way it hard to think of these reflections as being fundamentally different to reflections from any other cause, such as, say, cabinet reflections from a poorly mounted tweeter, or reflections off of a countertop surface that a system might be sitting on. Reflections may come from the mouth, from internal discontinuities or from HOMs, but there is no reason to believe that the predicating discontinuity is of any particular importance. Whatever the physical (or in the case of our FIR filter, electrical) cause, the sound is determined by delay, level and spectrum of the various reflections.
This being the case shouldn't we look at the levels and delays of the various reflections? Certainly the mouth reflections are the largest by far, and the most visible in the impulse responses. What LeCleac'h suggests are mouth diffraction reflections seem to typically be second strongest. Diffraction slots, bends and HOMs seem to be lower and more diffuse in time. (The measurement LeCleac'h suggests is mostly HOMs is attached.)
Why LeCleac'h can measure these and other can't is something I don't understand. Also, if this evidence is pretty typical of most horn performances, and if we are truly looking at linear effects (frequency response related effects independent of input level), then it would suggest that we really ought to be concentrating on reducing primary and secondary mouth reflections. At least before we get hung up on lesser reflection causes.
Regards,
David
Attachments: 3ms mouth reflection as impulse, and in 3D, comparison of clean and messy impulses, horns with HOMs
http://forums.melaudia.net/attachment.php?aid=1760
Here he looks at 16 horn/compression driver combinations. Each one has a response curve, an impulse response, a 3d wavelet transform of the impulse and a group delay curve.
This give a very in depth look at the time domain performance of a variety of horns. Besides the early response that contains the broad shape of the horn's frequency response, there are all of the other delayed reflections that will come from whatever cause exists. Some reflections are seen as quite distinct echoes. For example most of the horns have a significant echo at about 3 ms. LeCleac'h states that this is the horn mouth reflection. Energy reflects from the mouth back to the throat and then out again. 3ms corresponds to the transit time of two times the horn length. In addition LeCleac'h singles out mouth diffraction as a strong secondary reflection. It is closer to the main impulse as it has a path determined by sound diffracting around the horn contour to the mouth and then coming to the listener via the slightly longer diagonal path.
Horns with discontinuities such as bends in planar sections (Manta Ray) or diffraction slots will also have reflections at the corresponding times. LeCleac'h also identifies HOM related reflections.
The strong reflections are visible in the impulse response and even relatively subtle ones are seen in the color plots. The color plot shows level by color, time on the horizontal axis and frequency on the vertical axis. A reflection of a certain time delay will look like some sort of vertical wedge to the right of the primary impulse.
Now we are told that the horns are operating in the linear part of their range. Curves were taken at 91dB and the efficiency is such that the impulse responses would probably hold true for from 5 to 105dB. As such the impulse response should fully quantify the sound of the horn, at least for a listener on axis in a dead room (and with typical horn directivity, even in a somewhat lively room). That means that the measured impulse could be used for an FIR filter and any musical waveform could be convolved with it, giving the output (the sound) that we would expect from a particular horn.
Thought of that way, we should just think of the horn as merely a device imparting multiply reflected energy to any signal. The audible effect of the device can be solely determined by the standard audibility of reflections. Reflection delay, level and frequency spectrum are all that matter (in this case sound direction is essentially the same as the primary sound).
In this way it hard to think of these reflections as being fundamentally different to reflections from any other cause, such as, say, cabinet reflections from a poorly mounted tweeter, or reflections off of a countertop surface that a system might be sitting on. Reflections may come from the mouth, from internal discontinuities or from HOMs, but there is no reason to believe that the predicating discontinuity is of any particular importance. Whatever the physical (or in the case of our FIR filter, electrical) cause, the sound is determined by delay, level and spectrum of the various reflections.
This being the case shouldn't we look at the levels and delays of the various reflections? Certainly the mouth reflections are the largest by far, and the most visible in the impulse responses. What LeCleac'h suggests are mouth diffraction reflections seem to typically be second strongest. Diffraction slots, bends and HOMs seem to be lower and more diffuse in time. (The measurement LeCleac'h suggests is mostly HOMs is attached.)
Why LeCleac'h can measure these and other can't is something I don't understand. Also, if this evidence is pretty typical of most horn performances, and if we are truly looking at linear effects (frequency response related effects independent of input level), then it would suggest that we really ought to be concentrating on reducing primary and secondary mouth reflections. At least before we get hung up on lesser reflection causes.
Regards,
David
Attachments: 3ms mouth reflection as impulse, and in 3D, comparison of clean and messy impulses, horns with HOMs
Attachments
Others have, although it has been more or less ignored. Zilch measured the Altec 511 here:
AudioKarma.org Home Audio Stereo Discussion Forums - View Single Post - Z19 Crossover
An externally hosted image should be here but it was not working when we last tested it.
I am aware of Mr LeCleach's paper, but I don't accept that what he calls HOMs actually are. Because HOMs have cut-in frequencies, any reflection that goes to lower frequencies cannot be an HOM. There should be a staircase type of effect as each mode "cuts-in". I don't see that where he says there are HOMs. It may be possible to use this kind of analysis to find HOMs, but I don't think that he did. Also, the delay time would be much less than a reflection time.
I do agree that the #1 problem with many horns is mouth reflection/diffraction. I have always said that.
My main concern with horns and waveguides has always been directivity control and not any single axis response. Mr LeCleach has always had the opposite view. So we each developed techniques that optimally show what it is we are looking for. Nothing surprising in that.
I do agree that the #1 problem with many horns is mouth reflection/diffraction. I have always said that.
My main concern with horns and waveguides has always been directivity control and not any single axis response. Mr LeCleach has always had the opposite view. So we each developed techniques that optimally show what it is we are looking for. Nothing surprising in that.
If directivity control is key then good old diffraction slot type horns are the best available in that regard, yet we can't use them because of the HOM issues!
The point I'm making is that if HOMs are well lower in level than other more common issues, and in fact hard to even measure, then why do we care about them?
Why the attitude of some that they must be banished at all costs? (answer, because of the opinion (myth?) that they sound proportionately worse than conventional (old fashioned) aberrations.)
David
- Status
- Not open for further replies.