Snickers-is, I'd be interested in hearing more about your struggles with Seos waveguides. I've been relatively happy with the results of a SEOS 12 with DE250's however I'm looking at where to move next to improve on them. I'm considering larger axisymmetric horns including JMLC horns and OS waveguides.
Dan
I think it is very important to be aware of the large spectrum of nuances to a combination like this. For example, Seas 12 and Seas 15 would behave differently from each other simply due to their size. Simulations in BEM has shown that the throat roundover performance also depends on the throat diameter, meaning an 1,4" would benefit from a slightly slower transition than a 1" for the same level of HOM. Then we have the drivers. Drivers can not only create HOM themselves, but one driver can create HOM in the interfacing to one horn, but not in the interfacing to another horn, while another driver could be the exact oposite. It is really hard to look at one type of horn and make a complete judgement, it is very case dependent.
The audibility of HOM is another thing. Some wants to write it off completely, while others regard this as a significant issue. But what we can say for sure is that different frequencies, different time difference between the initial wave and the following HOM, different pattern complexity, different relative levels etc will highlight HOM audibly in different ways.
But without having all the answers, we can still review simulations, and we can still review user cases to help improve our understanding. And there are some clear findings that, even if we can not quantify them subjectively, we can say with certainty what solutions avoid HOM better than others. Some kind of tractrix shaped throat, followed by a slightly curved horn, ending in a roundover to avoid the worst exit diffractions seem to be as good as JMLC on HOM, but without the beaming.
In the driver there are two main sources of HOM. One is any abrupt change in cross section expansion, both negative and positive. The other is what comes out of the phase plug. When modes are generated in the phase plug, it will result in uneven pressurisation after the phase plug at this frequency. These resonances originate in the compression chamber so we can not fully aviod it in concentric phase plugs. However, having more slits (relative to the diaphragm diameter) can move these modes upward in frequency and reduce modes to a less significant level. Also note the difference between different phase plug mathematical models.
As we can see there are differences in the size of the core element, as well as the pacement of the outer ring. 2441/45 is closer to what we see on TAD 4003, and some say they have slightly better bandwidth.
I think the main takeaway is that phase plug geometry matters. But radial phase plugs are not a walk in the park either, as they do not very effectively control the distribution of pressure in the exit as it all comes out in a single wide slot.
Directivity changes the ration of direct sound to diffuse sound - as it would with decreasing the distance to a less direct speaker. But it does not change the decay or sound of the room itself.
I would for sure not go so far and say controlling the room and directivity will sound the same - it will not!
There are a few mechanisms happen all at once, you have to learn them one by one and then put it together to what our ear and brain receives.
The big difference between a microphone and our ear is that our hearing is time dependend. We hear a reflection differently when it arrives 5ms or 15ms after the main signal.
Just roughly - in the first 10ms the ear adds all the impulses to one. 10ms are about 3m path - so your side reflections or ceiling reflections would need to travel 3m longer as your direct signal. That never the case in normal size living rooms. So these reflections add to the main signal and when off axis frequency response is not linear - speaker response is not linear for our ear.
When building a studio room these first 10ms are crucial and the goal is to keep them free of strong reflections - absorption is one of the methods to do so.
There are different types of masking - Wikipedia should be helpful here.
There is "level masking" (a loud sound masks a silent sound) which depeds on the distance between the frequencies (as you describe). There is also time masking - a loud sound AFTER a silent sound can mask the silent sound! (That's how the cochlea and our nerves work - a loud signal get's processed faster, could be danger) All these masking effects are well researched and the reason why compression like mp3 is possible.
The "basic dynamic range" of the ear is about 60dB btw. The auditory ossicle helps to switch "gears" when signals get louder. Therefore it takes a while after a loud impact to hear more silent signals again.
Psychoacoustics ... it's such a great science!
There is a concept named "non environment room" which I built in my listening room. "It exposes the truth" is a good description
https://www.soundonsound.com/techniques/sos-guide-control-room-design
edit: here an interesting graph about time and power of reflections and how we receive them:
I tend to view the term 'HOM', in line with the early work that I think first defined it.
So HOMs to me, has a specific definition as quoted below, that has been around a long time.
I think if we call diffraction a type of HOM, or phase plug anomalies a HOM, we lose the value of the term HOM.
The snips came from an AudioXpress reprint....https://www.grc.com/acoustics/an-introduction-to-horn-theory.pdf
Snickers-is, thanks for this. I hadn't really considered the case of a tractrix profile with a roundover (ala JMLC?) as being much different from a JMLC horn at least from the amount of beaming. Would such a tractrix have wider dispersion than JMLC or less narrowing of dispersion at high frequencies?
Are you familiar with Geddes' patent on compression driver phase plugs? I'm wondering if following his method would result in more even pressurization. Thank you for the link to the Celestion video last week on Compression Driver design. That was quite interesting.
On some of the HOM discussion in this thread, it sounds like there are different definitions of HOM. Some are probably more familiar with the Geddes definition which seems distinct from how you are using. I won't say either is corrent or incorrect, only that some are probably more familiar with the Geddes' definition. Here's an interview with Earl from 2012:
In a previous post, I got the impression you had some particular frustrations in your experiences with SEOS waveguides. Do you attribute these to a fundamental aspect of the profile used for SEOS or maybe rather that the SEOS WG's may be less well suited to some compression drivers than others?
Dan
Hello
Why are you not looking at the termination side, the exit at the horn attachment to the driver ?? Tad uses a 5 slit JBL 4
You have 2 completely different phase plug designs shown. Traditional from the say the 1930"s 2441,2445 and modern from the early 90's Concentric Wave starting with the 2446 and carried through to the 2453. The 476Mg uses the same phase plug.
You cannot compare them based on ring spacing alone or make any valid comparisons. Look at how different they are terminated horn side.
Rob
Attachments
Last edited:
Reply to:
https://www.diyaudio.com/community/...rivers-today-2022.382609/page-98#post-7604896
Well, the thread is not about horn’s or waveguides but rather compression drivers. Yes there is a very substantive correlation but here the focus is on the compression driver.
In respect of that:
https://www.diyaudio.com/community/threads/waveguides-and-compression-driver-exit-angles.200340/
https://www.diyaudio.com/community/...rivers-today-2022.382609/page-98#post-7604896
Well, the thread is not about horn’s or waveguides but rather compression drivers. Yes there is a very substantive correlation but here the focus is on the compression driver.
In respect of that:
https://www.diyaudio.com/community/threads/waveguides-and-compression-driver-exit-angles.200340/
Last edited:
I strongly disagree on that. It is actually 100% in line with the description you quoted.
As this is a quote from Mr Bjørn Kolbrek, we may as well put it in context with his own words from his blog:
Especially this:
"If the wave front in the horn is not plane, cylindrical or spherical", meaning any sound field where the shape of the wave front deviates from these fundamental simple shapes.
And this:
"sound fields in general", which makes no requirement as for what is the source or sources of sound, and what causes the deviation from the wave front.
Please note that what I was describing was not Tractrix all the way out to the roundover. The main part should not be Tractrix, nor JMLC, if you want to avoid significant beaming. I think this is illustrated well numerous times in the ATH thread.
Not sure if you are pointing towards the actual phase plug patents or the foam plug. However, I do not find any of these patents useful. It looks to me like trying to fix a problem without actually fixing it. When you pressurize a horn via a dome, through an annular phase plug with multiple slits, and both the dome and the exit diameter are larger than a quarter og a wavelength, there is no way we can fully remove HOM just by altering the phase plug. The foam plug only act as a brake. Having a taller profile in the centre, it flattens the wave front slightly, but it does not actually resolve the issue with pressure differences as the brake is isotropic.
I think there is actually just one valid definition. In this context, the "mode" is one simple wave front in a space, originating from one single source geometrically. Once we change this shape we do no longer have one single wave fron in space originating from a single source geometrically, even if the sound technically originates from a single source. Thiss adds more simple wave fronts in the same space, aka modes, combining with the original wave front. This will be valid if we have a simple or complex gradual change in the wave front, if we have a diffraction, if we have a complex radiating surface, if we have multiple sources, if we have a reflection, or any other thing that alters the original wave in one way or another.
Two extreme examples may be a spark in an anechoic chamber, it is regarded as a reference transient, but it is not a perfect sphere, therefore it is actually not a simple wave front. The presence of HOM in this example is probably not audible as it is very small and at very high frequencies.
A loudspeaker in a reflective room will also form an altered wave front. This is normally not described as HOM since the human ear has no problem filtering HOM from the original signal.
But where HOM starts to get interesting is when we use line array pro audio systems to reach large audiences. A line array is not an ideal line, so we will have combined wave front shapes within the signal. If we made a "line array headphone", this would not be audible. If we listen to a 10 meter line array from a distance of 200 meters, it is not very audible either. But somewhere in between there, we have a transition zone. If the people rigging the gear makes a mistake in the relationship between curvature and throwing distance, some of the audioence will end up within the transition zone, and that does not sound pleasant at all.
So the definition is simply all about a distorted wave front shape. If the distance from any point on a wave front to the virtual originating point is not equal, the wave front is distorted, and we can say that HOM is present.
I have struggled a bit with SEOS. It is known that OSWG typically have some issues with HOM. Past that, my detailed knowledge of SEOS is not sufficient, so I can not extrapolate this into a conclusion.
But yes, matching drivers to horn is important, and most driver manufacturers give us more or less nothing in terms of details to help us determine this.
I think if folks (including experts) want to define any deviation in shape of a wavefront as being able to be described by a set of HOMs, then fine, so be it.
Seems to me that is more of a theoretical mathematical definition than one of any pragmatic value...
I mean, What's the point in a term that is now so broad it defines nothing...because it is a descriptor of anything/everything .....throat mismatches, phase plug anomalies, horn geometry issues, mouth reflections back into throat, mouth diffractions..... and on...
Are they all HOMs? What's the point? Other than saying maybe someday we may find the math to describe all those phenomenon individually?
I still like the idea of HOMs as first presented years ago with conical horns...at least that is something i can visualize as separate, from the set of wavefront disrupters just listed. Just my 2c, who am I.....
What do you mean? I mentioned both the horn side of the phase plug and the interface between the horn and driver, and I mentioned the importance of the number of slits. What exactly are you missing?
You have 2 completely different phase plug designs shown. Traditional from the say the 1930"s 2441,2445 and modern from the early 90's Concentric Wave starting with the 2446 and carried through to the 2453. The 476Mg uses the same phase plug.
You cannot compare them based on ring spacing alone or make any valid comparisons. Look at how different they are terminated horn side.
Rob
Sure, but Bob Smith's work on this (I believe this is the correct link: https://www.diyaudio.com/community/attachments/asaj025-0305-pdf.571477/) identified that the distance between the slits were crucial for the frequency response through the phase plug, independently of the slit lengths and widths. That does not mean these slits widths and lengths does not matter, but they are individual parameters that should be considered individually.
Both Vente & Thuras, who worked on the low pass function of the compression chamber and phase plug slits, and Bob Smiths investigation of the slit spacing lead to a lot of the things we see in most modern compression drivers with annular phase plug design. However, they did work on these same things in Japan, in parallell with the Americans, and they did arrive at similar conclusions. But they prioritized differently, so if we look at the later JBL designs and compare them to both TAD, JVC, Sony and Yamaha compression drivers, we see that there may be a trend where the Japanese used bigger core element in the phase plug. The slit spacing used by most US and Europeapn drivers are, as far as I know, based on the work of Bob Smith.
I believe this is in perfect harmony with reality.
In my compression driver design I made sure to keep everything at horn theory both in the compression chamber, in the actual phase plug and in the exit of the driver. In a horn, we can not expect that to be possible due to the dimensions of the horn, and the fact that a horn has to be finite. But we can strive to keep it low.
I still do not think there is a difference, between the two. However, maybe there is a subtle intuitive understanding missing somewhere in here?
To have HOMs in conical horns, you must first drive them with non-matching wavefronts (conical horns have no HOMs when driven with spherical wavefronts). So these are no longer HOMs of the horn itself, but due to this mismatch. Similar to a flat disc in a baffle, for example. These are all the same principles, just can have different names. It all leads to a non-perfect wavefront in the far field - different from any fundamental mode. But typically, this has a whole bunch of origins and it's not really possible to separate them...
Does not that depend on to what degree the problems we experience inside the horn originates in the compression driver? If something goes wrong inside the phase plugg due to its design, we can not regard this a matching issue, even if this issue may be less exposed in some horn designs.
Thank you Mabat, that is something I didn't grasp. I thought the classic conical HOM research was showing HOMs when driven with spherical wavefronts.
So i thought a HOM was totally a function of the horn itself.
Thank you Snicker-is, for hanging in politely with me. I see from Mabat's post, the distinction I was drawing does not exist.