Ed LaFontaine said:
Turns out that the higher the frequency the greater the angle at which it can intersect the waveguide walls. This means that higher frequencies can have more reflections than lower frequencies. At some point the frequency is low enough that wall reflections are not possible. This is called the "HOM cutoff". Below this frequency only planar or one dimensional waves can propagate. Only below this frequency is Websters Eq. valid as only "Webster like" waves can exist.
But what is counter intuitive is that as the wall angles increase, the cutoff frequency moves lower - not higher as intuition might suggest. Clearly this is a diffraction issue where the sharper throat flare diffracts more than the more gradual one.
If the CD provided a true planar wavefront then the HOM would be a sort of minimum, but they would still exist. There is actually a non-flat wavefront which will cancel the first HOM at some frequency, but there is no frequency independent wavefront curvature which will not generate any HOM. This was all discussed in my second paper on Waveguides and in my book.
Re: More comments
Exactly- the reason I brought this up was that there really aren't many speakers in that class. I'd be really interested in hearing a comparison between your speaker and the JBL K2, another speaker that seems to meet the requirements but that (at least it seems like) was designed without HOM considerations.
I see your point. I'm going to have to look at exactly what HOM are before I continue. I was operating under the assumption that the effects "seen" from the outside world due to the HOM were somewhat direction independent, which must not be true.
Will look closer; sorry for the sloppiness. It does seem quite surprising that distortion at those levels isn't distinguishable.
gedlee said:
Exactly- the reason I brought this up was that there really aren't many speakers in that class. I'd be really interested in hearing a comparison between your speaker and the JBL K2, another speaker that seems to meet the requirements but that (at least it seems like) was designed without HOM considerations.
gedlee said:
I see your point. I'm going to have to look at exactly what HOM are before I continue. I was operating under the assumption that the effects "seen" from the outside world due to the HOM were somewhat direction independent, which must not be true.
gedlee said:
Will look closer; sorry for the sloppiness. It does seem quite surprising that distortion at those levels isn't distinguishable.
Re: Re: More comments
And they cost five times as much.
They are definately direction dependent, but this fact is independent of their being non-minimum phase.
Believe me, it surprised us all!! Changed my entire way of looking at sound quality of horns and compression drivers. This result is what made me look into the audibility of HOM. There was something that I was hearing. If it wasn't nonlinear distortion then what was it? Our most recent paper put this last question to rest - its HOMs.
Rybaudio said:
And they cost five times as much.
They are definately direction dependent, but this fact is independent of their being non-minimum phase.
Believe me, it surprised us all!! Changed my entire way of looking at sound quality of horns and compression drivers. This result is what made me look into the audibility of HOM. There was something that I was hearing. If it wasn't nonlinear distortion then what was it? Our most recent paper put this last question to rest - its HOMs.
Is it this one: "Audibility of Linear Distortion with Variations in Sound Pressure Level and Group Delay" ? AES preprint 6888.
I have also read your paper on audibility of distortion in compression drivers with great interest. You had quite short times of exposure, only 15 seconds for each sample. How easy is it to recongnize distortion in that time span?
Best regards,
B
Re: Re: Re: More comments
I'm curious, is the inaudibility of nonlinear distortion in this case predicted by your nonlinear distortion perception metric? It seems that the majority of it being low order distortion combined with masking could account for this. I'll be sure to check out your newest paper.
Patrick Bateman,
I "get" the picture you are talking about and have all along; partial reflection and transmission at boundaries is a general behavior of waves and I see how the reflected wave would be damped more so than the original using the foam; that is not the issue. I have been fooled enough times by simple pictures and intuition that know I need to approach the situation from all possible angles. What I really need to do is read the original papers and "listen to the equations." BTW, I am very interested in hearing your setup- I'll send an email sometime soon.
gedlee said:
I'm curious, is the inaudibility of nonlinear distortion in this case predicted by your nonlinear distortion perception metric? It seems that the majority of it being low order distortion combined with masking could account for this. I'll be sure to check out your newest paper.
Patrick Bateman,
I "get" the picture you are talking about and have all along; partial reflection and transmission at boundaries is a general behavior of waves and I see how the reflected wave would be damped more so than the original using the foam; that is not the issue. I have been fooled enough times by simple pictures and intuition that know I need to approach the situation from all possible angles. What I really need to do is read the original papers and "listen to the equations." BTW, I am very interested in hearing your setup- I'll send an email sometime soon.
Links to Geddes' papers
http://baseportal.de/cgi-bin/baseportal.pl?htx=/Data/exdreamaudio/download_ex
4 are grouped together, scroll down to near the bottom of the page.
http://baseportal.de/cgi-bin/baseportal.pl?htx=/Data/exdreamaudio/download_ex
4 are grouped together, scroll down to near the bottom of the page.
One thing we haven't discussed here -
What's more obnoxious - distortion or HOMs?
Based on my listening experience with the HOM-reducing plug, HOMs are *far* more obnoxious than distortion.
First, my apologies to Dr Geddes, because I know he's not keen on subjective descriptions. But here goes...
Every one of us have listened to a cheap P.A. system that isn't audibly distorting, but the character of the response is like a set of fingernails on a chalkboard. To a lesser extent, even "high end" horns do this. When I listen to Avant Garde horns at high levels, I can hear that "raspiness" clearly. The most frustrating aspect of this "sound" is that it cannot be fixed with EQ (I've tried, believe me.) When I've played around with response curves, the speaker invariably sounds "darker", but that "grating" sound is still there.
Now put the foam in the waveguide, re-EQ, and it's all but gone.
The difference is nothing less than remarkable - it literally makes a waveguide sound like a conventional dome tweeter, but with astound dynamics and headroom. So if the scientific explanations are a bit much, just go and listen to one.
I may have mentioned this before, but I flew 1500 miles to do just that.
Now on to the question :
HARMONIC DISTORTION VS HOMs
With a bit of work, I can definitely "notice" the sound of distortion. But the "sound" of distortion isn't nearly as obnoxious as the "sound" of HOMs. I have a hunch this has a lot to do with the time component; IMHO this has to do with the fact that the ear is more sensitive to problems in the time domain than the frequency domain. This also explains why tube amps, loaded with harmonic distortion, are still very listenable.
What's more obnoxious - distortion or HOMs?
Based on my listening experience with the HOM-reducing plug, HOMs are *far* more obnoxious than distortion.
First, my apologies to Dr Geddes, because I know he's not keen on subjective descriptions. But here goes...
Every one of us have listened to a cheap P.A. system that isn't audibly distorting, but the character of the response is like a set of fingernails on a chalkboard. To a lesser extent, even "high end" horns do this. When I listen to Avant Garde horns at high levels, I can hear that "raspiness" clearly. The most frustrating aspect of this "sound" is that it cannot be fixed with EQ (I've tried, believe me.) When I've played around with response curves, the speaker invariably sounds "darker", but that "grating" sound is still there.
Now put the foam in the waveguide, re-EQ, and it's all but gone.
The difference is nothing less than remarkable - it literally makes a waveguide sound like a conventional dome tweeter, but with astound dynamics and headroom. So if the scientific explanations are a bit much, just go and listen to one.
I may have mentioned this before, but I flew 1500 miles to do just that.
Now on to the question :
HARMONIC DISTORTION VS HOMs
With a bit of work, I can definitely "notice" the sound of distortion. But the "sound" of distortion isn't nearly as obnoxious as the "sound" of HOMs. I have a hunch this has a lot to do with the time component; IMHO this has to do with the fact that the ear is more sensitive to problems in the time domain than the frequency domain. This also explains why tube amps, loaded with harmonic distortion, are still very listenable.
gedlee said:
I love it when Dr Geddes throws out one of these "nuggets" that has huge implications on the sound quality of our speakers. Earlier in the thread he mentioned that HOMs are level-dependent, which also has monumental implications on sound quality.
Think about the aforementioned post for a second - if narrow angle waveguides have a higher "HOM cutoff", does that mean that they also have lower HOMs?
This might explain John Sheerin's preference for a narrow angle conical waveguide on his Unity clones:
http://ldsg.snippets.org/HORNS/unity2.html
Or it could just be room interaction. Or both. Something to consider, no doubt.
If I understand correctly, the sharp transition angle at the throat is a major, if not the major factor in generating HOM's, It would make sense then, that the wider the angle, the more HOM's would be generated near the throat. Whether or not this is correct, my experience is that the foam in the throat is the most important. In my case, I put a foam plug all the way to the phase plug of a TAD 2001, and a little more than a third of the way out of a Unity (conical) horn. Without the plug, some harshness and slight "honk" is perceptible. With the short plug, it's gone. I don't notice a difference between that and the horn filled out to the front edge. More sensitive ears than mine may differ.
Sheldon
Edit: I should add that the extra foam didn't seem to degrade the sound, other than some additional attenuation.
Sheldon
Edit: I should add that the extra foam didn't seem to degrade the sound, other than some additional attenuation.
I'm working in Santa Monica for the next month or so. If you'd ever like to show off your Unities, I'd love to hear them!
It's ironic that I've built speakers "inspired" by The Unity, but I've only heard them once. And that one time was far from an ideal environment - it was at Club Ra inside of the Luxor hotel in Las Vegas. (or was that Turbosound?)
It's ironic that I've built speakers "inspired" by The Unity, but I've only heard them once. And that one time was far from an ideal environment - it was at Club Ra inside of the Luxor hotel in Las Vegas. (or was that Turbosound?)
Sheldon,
"Without the plug, some harshness and slight "honk" is perceptible. With the short plug, it's gone. I don't notice a difference between that and the horn filled out to the front edge."
It makes sense, and it takes a lot less foam, and the shallower angle of the horn wall makes it eaiser to cut the foam to match.
"Without the plug, some harshness and slight "honk" is perceptible. With the short plug, it's gone. I don't notice a difference between that and the horn filled out to the front edge."
It makes sense, and it takes a lot less foam, and the shallower angle of the horn wall makes it eaiser to cut the foam to match.
Kolbrek said:
First, the passage was 15 seconds, but listeners could hear it as many times as they wished.
When you can A/B a distorted signal with an undistorted one in real time, its almost trivial to detect differences when the are audible. It seldom takes me more than 3 sec or so to detect a true difference.
Interestingly enough we found a high correlation between uncertainty in the data and the time it takes a subject to perform the test. Basuically this says that if you listen long enough, you will convince yourself that you hear a difference even when one does not exist.
Quite profound, right!?
Re: Re: Re: Re: More comments
The results of all four of our papers on the audibility of distortion are consistant. They all point to the same thing - in loudspeakers nonlinear distortion is not an issue at "normal" listening levels. Of course I COULD design a speaker where this was not true, but the point is that I can also design one where it is true. I prefer to do the later (although I sometimes suspect that many designers prefer to do the former).
Rybaudio said:
The results of all four of our papers on the audibility of distortion are consistant. They all point to the same thing - in loudspeakers nonlinear distortion is not an issue at "normal" listening levels. Of course I COULD design a speaker where this was not true, but the point is that I can also design one where it is true. I prefer to do the later (although I sometimes suspect that many designers prefer to do the former).
Patrick Bateman said:
Since nonlinear distortion in a compression driver is virtually inaudible, but the HOM are clearly audible the HOM would have to be the more obnoxiuos, right?
Your last para is right on the money. Turns out the ear masks very well in the frequency domain, but not in the time domain. So if the nonlinear distortion is not time delayed it is highly masked by the signal. BUT, if there is a time delay in the distortion (not necessarily nonlinear) - like HOM or diffraction - then this effect is only weakly masked and quite audible. The important thing is that the masking in the time domain is SPL dependent just like the masking in the frequency domain - except that in the time domain the masking is less at higher levels (the opposite of masking in the frequency domain).
By the HOM, the PERCEPTION of the HOM is level dependent. The generation of the HOM is not.
Member
Joined 2003