bear said:
Bear,
I'm afraid You didn't get the clue. HOM is not - not! - defined within other than OS and conical waveguides. It can not be taken against non-OS, resp. non-conical devices. Due to it's simple non-existance. I think that point was made very clear. We have the funny situation that - by their definition - HOM can only exist in waveguides that avoid them by first principle. I know it's funny, but it is what the whole mess is about.
You may re-read the post of the OEM on that and my more elaborate reply.
bear said:
From what? EV uses two split fins in all 2" HP horns, as far as I know. The HP940 avoids a lot of trouble that all the 2" JBL biradials make regarding directivity and ampl.response. You should have a pair of EVs for evaluation, they are cheep second hand.
so long
soongsc said:
What CARA actually does is simulating pressure distribution over time within closed boundaries.
It does this based upon ray tracing method taking into account several conditions and variables *especially* useful to model room acoustics – but not necessarily limited to that, as we are solely talking about "linear acoustics" here.
If, for example, you define a certain boundary to be 100% absorptive you can "open" that closed contour like for a horn / wave guide / directivity control device at its mouth.
The main restrictions are that it is tediously to model complex contours – even in 2D – but there is no chance to even thinking about precisely modelling axisymmetric or non-axisymmetric 3D horn contours in (current version) CARA.
One other restriction is that you can't model tiny structures (compared to rooms).
Thus you have to scale up your horn / wave guide / directivity control device and get a down shift in frequency behaviour compared to the original size in reward (hence I never inserted the CARA frequency numbers in my "bunch of pretty pictures").
All in all – your far filed / near field concerns do not apply .
CARA comes in English and German version – isn't exorbitant expensive and features a responsive service line
http://www.cara.de/
Michael
wxa666 said:
If you allow me to put it different - its probably better *not* necessarily to focus (with blinkers) on the subset (HOM) only, but rather to see HOM in the general frame of the implications of the subsequent actions of diffraction > reflection > delay > interference
Not to say that HOM *as a subset* may not be interesting to investigate any further as it *may* stand for a specific perception as assumed by Earl.
Michael
mige0 said:
If you allow me to put it different - its probably better *not* necessarily to focus (with blinkers) on the subset (HOM) only, but rather to see HOM in the general frame of the implications of the subsequent actions of diffraction > reflection > delay > interference
Not to say that HOM *as a subset* may not be interesting to investigate any further as it *may* stand for a specific perception as assumed by Earl.
Michael
Your eagerness on HOM is out of focus. I don't see any relation to the question on how HOM may be measured in Your many posts here! You obviously want to design a waveguide by Your own. You show some results with standard measurements as amplitude over frequency, but that's OFF TOPIC - here. Why don't You start an own thread on it, or keep it in the thread it is already discussed in?
You may come back to here, when You gain some results, that are related to HOM which are - despite their absence - well defined in themselfs. Your attempt to literally and intentionally missuse the term just to justify Your writing is quite annoying.
Thank You
wxa666 said:
Just to get clarification.
You are saying that HOM does not exist in non-OS or conical-waveguides. Meaning all other CD horns are immune to it? They do have their "horn" sound but HOM isnt something we should worry about?
That point has never been made clear...maybe to you but there are many different threads on many different sites figuring out how to minimize/measure HOMs on non-OS, non-conical waveguides.
I found his questions/opinions to be very useful, Ie...Im learning from him. Is there a need to post saying he is being "annoying"?
Your posting that "someone didnt get a clue" is rather rude but you call others "annoying"....just discuss the topic.
Do people actually think they are above others here? I read it Daily, its amazing that we can not just have discussions with that type of post. Some people are going to be wrong (like me!) and some are going to be right, we are going to have disagreements but it would be nice for everyone to stay respectful.
Hello,
Measuring HOM by capturing the wavefront is undoubtedly quite complicated (however it is possible).
But there are (may be) other methods to measure HOM.
May I propose a method:
From the another thread we can learn that higher order modes are delayed compared the the basic mode. Then maybe we can think that HOM is another wavefront delayed and attenuated from the main wavefront. To capture that we can measure the frequency response in a temporary steadystate condition, because main wavefront and HOM are sinusoidal signals in this case, and their sum will show the total effect in steady state condition. You just have to make sure the steady state condition is temporary to avoid room reflections.
I think the best excitation signal for such a condition is a stepped sinusoidal burst. Select the burst length the maximum room allows without reflections. During the burst, main wavefront and HOM will be summed. Calculate the envelope of the burst (while in the steady state). Plot the frequency response using the burst envelopes. Examine the freq response to see any bumps or notches. Select a bump/notch and switch to time domain at that frequency to see if there is any ringing (this can be done in waterfall plot).
Then conclusion:
- If no bumbs/notches -> no HOM
- If a bumb/notch and ringing -> resonance
- If a bumb/notch and no ringing -> potential HOM
Notice that this method does not need assumptions of the shape of the basic mode
- Elias
Measuring HOM by capturing the wavefront is undoubtedly quite complicated (however it is possible).
But there are (may be) other methods to measure HOM.
May I propose a method:
From the another thread we can learn that higher order modes are delayed compared the the basic mode. Then maybe we can think that HOM is another wavefront delayed and attenuated from the main wavefront. To capture that we can measure the frequency response in a temporary steadystate condition, because main wavefront and HOM are sinusoidal signals in this case, and their sum will show the total effect in steady state condition. You just have to make sure the steady state condition is temporary to avoid room reflections.
I think the best excitation signal for such a condition is a stepped sinusoidal burst. Select the burst length the maximum room allows without reflections. During the burst, main wavefront and HOM will be summed. Calculate the envelope of the burst (while in the steady state). Plot the frequency response using the burst envelopes. Examine the freq response to see any bumps or notches. Select a bump/notch and switch to time domain at that frequency to see if there is any ringing (this can be done in waterfall plot).
Then conclusion:
- If no bumbs/notches -> no HOM
- If a bumb/notch and ringing -> resonance
- If a bumb/notch and no ringing -> potential HOM
Notice that this method does not need assumptions of the shape of the basic mode
- Elias
doug20 said:
Sorry,
HOM is a term that is known in microwave technology and - in audio since the OS-/conical-type of waveguide has been - one must say: revealed. The latter led to an analytical investigation of HOM that wasn't possible before, because it is possible in OS/conical waveguides only. In any other HOM is not defined because the basic mode is not defined to which HOM has to be related to.
That is a clue to get. Some people stated that even a singer singing would generate HOM. It wasn't meant as a joke.
My intend was to keep the thread clear from all that unrelated stuff. HOM is discussed so widely spread all over the DIY department, it is ubiquitous. The more a minimum of discipline is needed here, to keep things straight. This thread is by its original question about the measurement of what is called HOM. It is a matter of respect to not stretch it to the possible annoyance af a singer by his HOM. But the damage is done - some have occupied, lets say STOLEN this thread.
wxa666 said:
Not precisely true as you forgot the pipe, which has a plane wave for the fundamental mode *and* is the easiest contour to investigate.
Also Earl stretches (at least his definition of) *HOM* as to be produced by any horn - don't think its out of the blue.
*If* you are familiar with the definition and impacts of HOM in micro wave apps - has this been investigated earlier and whats been its outcome?
You always can ask but I don't think he would confirm you point of view.
Michael
wxa666 said:
Well, I think that this thread is now quite secondary to the main Geddes On Waveguides... so I think I'll turn my attention there. Try to catch up on the reading.
As far as I am concerned HOMs likely exist.
Their audibility is open to discussion.
What Dr. Geddes hears without foam vs. with foam may or may not be "HOM".
The paper cited earlier said that HOMs only appear at >10kHz. That is interesting.
I'll stick to the foam vs. no foam measurement idea for now as a practical point to start with.
EV horns are less than optimal in my view.
If you like them, I am ok with it.
Probably "fins" = diffraction or reflection effects, so not my first choice. Might work fine, dunno.
I'm no expert on any of this. I'm probably wrong most of the time.
_-_-bear
I just checked what the HP940 is... my suggestion and advice is to purchase an equivalent horn (for the freq range you want) from Dr. Geddes. The HP940 has sharp discontinuities in the horn construction (Mantaray style) as well as a horrific 90 degree bend to the baffle... diffraction slot at the throat as well.
In short, everything that the Geddes design is intended to avoid and "fix".
You should have no trouble finding HOMs with this horn!
Sorry to be so blunt on this...
_-_-bear
In short, everything that the Geddes design is intended to avoid and "fix".
You should have no trouble finding HOMs with this horn!
Sorry to be so blunt on this...
_-_-bear
soongsc said:
I do too in many cases. If the DUT were perfectly linear, it wouldn't matter what signal was being used. The more nonlinear it is, the more the signal type used has an effect. One signal may "trip up" the device in one way, another signal makes it act a little differently. It's mostly due to energy distribution.
Here's a good discussion about that on AudioRoundTable.com:
- Measurement signal types, by Keith Larson
bear said:I just checked what the HP940 is... my suggestion and advice is to purchase an equivalent horn (for the freq range you want) from Dr. Geddes. The HP940 has sharp discontinuities in the horn construction (Mantaray style) as well as a horrific 90 degree bend to the baffle... diffraction slot at the throat as well.
In short, everything that the Geddes design is intended to avoid and "fix".
You should have no trouble finding HOMs with this horn!
Sorry to be so blunt on this...
_-_-bear
I never found those, though. All Your advice is pretty clear a perpetuation of the advertizing of the OS-inventor and soley OEM and vendor. Alas, there seems little beef in it, as we still don't know how o measure that mysterious evil that comes with non-OS-horns. Of which the HP940 is a top notch spectacular one, what You found in a sudden.
What You might have overlooked is, that these fins are against non axial modes within the throat, what I mentioned originally. And it works. So, the fact that I dont see "HOM" with that terrific HP940 might be due to an attenuation by the right cure. Both, OS and HP940 may avoid "HOM" by construction. Would You mind to long for a pair of both to investigate the matter by Yourself?
Your perception of what the concept of "HOM" is, may not be right that, what a scientist has. The inventor Prof. Earl Geddes does not to much to straighten it out. The kernel of it is o/k. But the use in advertizing is k/o. I put on the stamp "snakeoil", and I'm trough.
Have fun
wxa666 said:
Fine - why not leave the place in silence then instead of prolonging your aggression – probably finding better companionship in hifi-forum or the like?
*If* on the other hand you want to share some of your enlightenment why not post constructively – laying out your arguments and weight it against those from others ?
Michael
(having fun !)
Or this: http://www.acoustics.hut.fi/teaching/S-89.3430/download/Response_measurements.pdfWayne Parham said:
Wayne Parham said:
The ARTA manual discusses different signals for acquiring the impulse response.
MLS -- he's not a big fan as the signal has several problems with PC sound cards and doesn't handle distortion well. Sort of an old-school signal meant to be used with dedicated hardware.
Swept sine -- works well and handles distortion okay but needs a low-noise environment.
Periodic noise -- his favorite for most tasks, reasonably good at rejecting both noise and distortion. ARTA is the only package I'm aware of that has that signal available for generating the impulse.
noah katz said:
Noah,
As promised, here are instructions on how to "clone" an XT1086. It also shows how to fix the kink in the throat.
Here's the URL:
http://www.diymobileaudio.com/forum...creating-perfect-soundstage-3.html#post761163
Here's a few pics:
Here's an aluminum foil plug, made from the XT1086 bell. You can see the kink in the throat quite readily.
Here I've added bondo, to fill out the kink.
Spackle works really nicely too.
Here are polar measurements, with a BMS 4540ND compression driver. These are a bit rough, but that's because this is going in my car
So these measurements are actually performed in the car, with the waveguides sitting on my dash! Even with a gated measurement, it's impossible to avoid serious reflections.
So it's challenging to get good results, due to all the reflective surfaces. There's an obvious reflection at 3khz and 6khz. Since the two reflections are mirror images, it likely indicates that there's a resonance off something thats 1.125" long. Maybe the throat itself? A little bondo should fix this.
Anyhow, each curve is taken at approximately 12 degree increments. Starting with the on-axis curve, it goes orange, red, purple, blue, green.
To be honest, the OS waveguides that I built had better looking polar plots. But again, this is going in my car, and I'm prioritizing size over perfect response. I'll build a foam plug and add some bondo to smooth things out, then post new measurements.
Here's a pic of me working in my kitchen. Has anyone tried making waveguides out of mashed potatoes?
I know this sounds crazy, but ever since yesterday on the road, I've been seeing this shape. Shaving cream, pillows... Dammit! I know this. I know what this is! This means something. This is important.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.