I don't think Pano doesn't believe they exist just that it hasn't been proven with appropriate direct evidence what causes they can infer with different systems.
And I agree with you Scott about the reduced space loading increasing the efficiency raising EIRP, but not Q
My off comment above about having fun listening to a 6kHz sine wave wasn't a joke. You can hear these standing waves or HOMs if you prefer with this test, walk anywhere and they are present. increase the frequency and they change accordingly tighter and tighter. Decreasing them increases the wavelength and the reverse is true. They are not normally audible through natural sound sources unless high enough in level on a continuous note being played. Introduce diffraction and reflection their effects are felt, poorly time aligned at the crossover point in both the Y and Z axis will introduce IMD from the rarefaction intermingled with baffle diffraction and every other blip on the radar. This energy causes unbalance pressure zones on a compression driver causing an unnatural breakup
One of my favorite tools is a Time Domain Reflectometer. Have used everything from RF low frequency gear to EHF radar and optical. To see down a cable/waveguide/fibre length out the antenna/termination IS very interesting. In this regard I prefer analog crt types, digital do not have the resolution to show the finer details, like graphics cards designed for gamer types and medical use that have increased shading range. Skin effect causes loss, any change in flow is seen as an impedance shift, with spray from the diffraction shifting the phase as it goes. Similar to how low frequency group delay shifts the focus.
Still I think of a waveguide as a transmission line and a horn as an extension of a waveguide / reflector. 1/4 wave stub in a waveguide that flares to a horn is the same 😉
Last edited:
And we will see evidence soon, yes? Acoustical measurements would be nice. Something that show they exist and are possibly significant at audio frequencies. Possible?
It took me 5 min to find Makarski's thesis and read all about how he did it.
"Tools for Professional Development ...."
gosh, and you say that actually makes sense
I have no idea what that paper means
and guess many of us will have to trust you on that
though I did try to find the word HOM, without any luck
but maybe I overlooked it
I thought it would at least be mentioned, just to have pointer
I have no idea what that paper means
and guess many of us will have to trust you on that
though I did try to find the word HOM, without any luck
but maybe I overlooked it
I thought it would at least be mentioned, just to have pointer
Thanks Frank and John. 
Reading it now. Finally, after almost 5 years of a thread about measuring HOMs, we have something to sink our teeth into. Nice.
Some very good, well thought out work there in chapter 3. It does seem that the driver is the major player, but I'll have to read chapters 4 and 5. Am I right in reading that the higher order modes show as roughness in the top end of the frequency response?
Reading it now. Finally, after almost 5 years of a thread about measuring HOMs, we have something to sink our teeth into. Nice.Some very good, well thought out work there in chapter 3. It does seem that the driver is the major player, but I'll have to read chapters 4 and 5. Am I right in reading that the higher order modes show as roughness in the top end of the frequency response?
In my initial scan of the paper, it seems that diaphragm breakup modes are being discussed as the high order modes, and the effect that acoustic loading has on them; i.e. how they are propagated and/or fed back.
Last edited:
Yes, that's what I'm seeing too. Also that only a few of the modes past the fundimental actually make it out of the horn into the far-field.
Diaphragm breakup modes would be a major contributor to HOMs, and in some cases of severe breakup these breakup modes would swamp out any other source of HOMs. At B&C we started to look at the HOM content of their compression drivers, but, as with all things, other priorities got in the way and we did not continue. So it is very difficult for me to say if these modes would be dominate or not. I suspect that above 8-9 kHz breakups of the diaphragm would be dominate. Below that the acoustic sources would likely be the greater aspect since the diaphragm is fairly uni-modal.
But remember that Makarski was only interested in HOMs to the extent that they would alter the frequency response or the polar response, which is not likely to be the case. He never considered audibility at all.
But remember that Makarski was only interested in HOMs to the extent that they would alter the frequency response or the polar response, which is not likely to be the case. He never considered audibility at all.
"Higher order modes" is a very generic term. All the modes at higher frequency than the first breakup mode (in aircraft structure we call it first bending mode) are called higher order modes. In horns, it seems that these cone breakup is only one of the sources of these modes.
So it seems from the paper. Pipe resonances are also calculated and shown.
We don't yet have any strictly horn related measurements that might be audible. Unless I've missed them.
E.G., show a horn that should have significant HOMs vs one that should not - using the same driver. What differences can we measure? Are those differences audible? Can we be sure that we are measuring HOMs and not some other effect?
We don't yet have any strictly horn related measurements that might be audible. Unless I've missed them.
E.G., show a horn that should have significant HOMs vs one that should not - using the same driver. What differences can we measure? Are those differences audible? Can we be sure that we are measuring HOMs and not some other effect?
Hi Pano, Earl
I don’t want to put words in Earls mouth but I wouldn’t have thought of the modes shown in that paper as being HOM’s but rather the normal distribution of an acoustically large source.
But… consider that without the complication of the horn what a piston radiator that becomes large compared to the wavelength being produced changes how It radiates.
http://www.soundandcommunications.com/archive_site/audio/images/2005_pics/sc05_09_audio10.gif
From infinitely small (compared to the wavelength) up to around ¼ wavelength, a piston source radiates in all directions equally. Here you can combine a number of identical sources and they continue to radiate as one source and the sources feel each other’s radiation pressure and so as with subwoofers, become more efficient as well. Anyway, if you were to look at these patterns “on axis”, the larger sized radiators alone also produces that concentric mode pattern
Mode coupling is more common than one might think too, if you make a bass horn that has parallel walls, one finds a notch in the frequency response where the distance between walls is ½ wavelength.
Now, the 2 inch exit driver used in that example is more than 3 wavelengths across at 20Khz. Without a horn attached, the driver might well produce a radiation pattern like is shown for the 3wl example above. As the driver exit is already large enough to confine the radiation pattern, what happens when you put it on a horn? Ans, up high the driver continues to radiate as before and IF there are little side lobes, those may bounce off the walls as opposed to being axial. To me, it has always been those reflections bouncing around which take the longer path within the horn that are HOM’s. Also, by adding the horn, it may generate other energy that reflects similarly within the horn.
Some rules which are ignored too often;
To gain any electroacoustic efficiency from horn loading, one must be operating on the sloped part of the radiation resistance curve and that knee (k=1) is where the radiator circumference is about 1 wL. For one, this means that even with a one inch exit driver, ALL the impedance transformation was finished well before reaching the driver exit. The horn you attach does nothing for acoustic loading up high. Consider what diameter is a horn with a 1WL circumference at 20KHz?
At the opposite end, a horn has a point where sound will radiate away at a fixed angle set by the horn wall angle and dimensions. The thumb rule which seems to be pretty close is Don Keele’s pattern loss Frequency formula, implies pattern flip as one losses pattern control and show why the long round horns could by narrowing their pattern, get away without electrical compensation for the drivers falling response up high..
It is the region above the “sound as a fluid” simple point source and below pattern control that the interior of the horn contour can add more points of radiation in addition to whatever comes from the driver, it’s dimensions. These can also bounce around. Like Baffle diffraction, you do not want rapid changes in the acoustic geography as these can radiate “as if” there was a tiny little source at some frequency.
Not surprisingly, there is what to me appears to be a similar effect in optics which is partly wave behavior. This is what limits the systems “Airy disk” which in a lens governs how small it is possible to make the focal point. In the case of filling out a straight walled horn is related to how small the source has to be at the apex to avoid the modes.
Airy disk - Wikipedia, the free encyclopedia
Earl, I was thinking, in the way old days I was at paper by Laurie Fincham about using an impulse as the test signal. There are reasons now that arriving at the impulse response is better than taking the FFT of an impulse test signal, but if I remember, he used a 20us pulse (like 20V too). I wonder, since the HOM’s travel a longer path, if they might be more visible in a time domain measurement like ETC (which I think is easier to read than the impulse response of the same system)?
Another possibility might be to do normal impulse measurements but band limit your test signal with a filter so that you could look at octave or fractional octave slices alone, it might allow you contrast them more .
My guess is these things would be frequency and geometry dependent but to be honest I don’t see what makes the HOM’s worse (relative to the input signal) with increasing level (but maybe more audible with overall increasing level?).
Best,
Tom
I don’t want to put words in Earls mouth but I wouldn’t have thought of the modes shown in that paper as being HOM’s but rather the normal distribution of an acoustically large source.
But… consider that without the complication of the horn what a piston radiator that becomes large compared to the wavelength being produced changes how It radiates.
http://www.soundandcommunications.com/archive_site/audio/images/2005_pics/sc05_09_audio10.gif
From infinitely small (compared to the wavelength) up to around ¼ wavelength, a piston source radiates in all directions equally. Here you can combine a number of identical sources and they continue to radiate as one source and the sources feel each other’s radiation pressure and so as with subwoofers, become more efficient as well. Anyway, if you were to look at these patterns “on axis”, the larger sized radiators alone also produces that concentric mode pattern
Mode coupling is more common than one might think too, if you make a bass horn that has parallel walls, one finds a notch in the frequency response where the distance between walls is ½ wavelength.
Now, the 2 inch exit driver used in that example is more than 3 wavelengths across at 20Khz. Without a horn attached, the driver might well produce a radiation pattern like is shown for the 3wl example above. As the driver exit is already large enough to confine the radiation pattern, what happens when you put it on a horn? Ans, up high the driver continues to radiate as before and IF there are little side lobes, those may bounce off the walls as opposed to being axial. To me, it has always been those reflections bouncing around which take the longer path within the horn that are HOM’s. Also, by adding the horn, it may generate other energy that reflects similarly within the horn.
Some rules which are ignored too often;
To gain any electroacoustic efficiency from horn loading, one must be operating on the sloped part of the radiation resistance curve and that knee (k=1) is where the radiator circumference is about 1 wL. For one, this means that even with a one inch exit driver, ALL the impedance transformation was finished well before reaching the driver exit. The horn you attach does nothing for acoustic loading up high. Consider what diameter is a horn with a 1WL circumference at 20KHz?
At the opposite end, a horn has a point where sound will radiate away at a fixed angle set by the horn wall angle and dimensions. The thumb rule which seems to be pretty close is Don Keele’s pattern loss Frequency formula, implies pattern flip as one losses pattern control and show why the long round horns could by narrowing their pattern, get away without electrical compensation for the drivers falling response up high..
It is the region above the “sound as a fluid” simple point source and below pattern control that the interior of the horn contour can add more points of radiation in addition to whatever comes from the driver, it’s dimensions. These can also bounce around. Like Baffle diffraction, you do not want rapid changes in the acoustic geography as these can radiate “as if” there was a tiny little source at some frequency.
Not surprisingly, there is what to me appears to be a similar effect in optics which is partly wave behavior. This is what limits the systems “Airy disk” which in a lens governs how small it is possible to make the focal point. In the case of filling out a straight walled horn is related to how small the source has to be at the apex to avoid the modes.
Airy disk - Wikipedia, the free encyclopedia
Earl, I was thinking, in the way old days I was at paper by Laurie Fincham about using an impulse as the test signal. There are reasons now that arriving at the impulse response is better than taking the FFT of an impulse test signal, but if I remember, he used a 20us pulse (like 20V too). I wonder, since the HOM’s travel a longer path, if they might be more visible in a time domain measurement like ETC (which I think is easier to read than the impulse response of the same system)?
Another possibility might be to do normal impulse measurements but band limit your test signal with a filter so that you could look at octave or fractional octave slices alone, it might allow you contrast them more .
My guess is these things would be frequency and geometry dependent but to be honest I don’t see what makes the HOM’s worse (relative to the input signal) with increasing level (but maybe more audible with overall increasing level?).
Best,
Tom
Yes Tom, if the HOMs that Geddes is talking about are caused by reflections and different path lengths, then we might hope that ETC or sonograms would show them clearly. So far, we haven't seen that. If the measurements match the models, we can all be happy.
This is a thread about how to measure Higher Order Modes but there isn't much measuring going on.
This is a thread about how to measure Higher Order Modes but there isn't much measuring going on.
Hi Tom
I did exactly the impulse response measurement that you suggest and showed it at ALMA many years back. I was able to show how the foam did reduce the later arriving - albeit very small - portions of the impulse response. Was this ALL HOM, I don't know. It was a simple test that could be HOMs, but as with ALL simple tests it could also be something else.
Why would HOMs, or anything for that matter, be more audible at higher SPLs? It certainly seems like the exact opposite of what one would expect. Masking in the frequency domain increase with level, which is why we don't hear nonlinearities at higher levels. So WHY would HOMs be more audible at higher SPLs than less?
As with many things the "why" is hard to answer, but the fact that they do is pretty clear. As Moore points out in his papers, our ears become more sensitive to excess group delay at higher SPLs and within 1000 - 4000 Hz. He claims (and he is the expect) that this is well known. HOMs have excess group delay. Is it such a stretch to put two and two together?
On paper, HOMs would not be so difficult to measure. I could certainly write out the math and what needed to be done (I did that for B&C), but "on paper" is cheap. Building the facilities, similar to those Makarski had, is not cheap. I don't have those resources. And all that to prove what? To me there is nothing more to prove. Others seem to need to see blood before they will buy into the idea however.
- Status
- Not open for further replies.