This is not an apples to apples comparison. Yours are ground plane measurements and mine are free field - they are not equivalent. And since your claims and responses here are anything but honest and candid, I am not going to argue the point anymore with you. But anyway you look at it, your claim of a "hole at 7.5 degrees" was a incorrect - typical.
Earl, look. I think your catenary flare idea is great. I applaud you for your research, and for promoting the OS flare. Hat's off to you for that. I think once you're promoting elliptical waveguides, they'll really be something.
But round horns present a real problem. You can work with them, but they're hard to get the verticals right, or even acceptable, in my opinion. Is it a worthwhile tradeoff, maybe, maybe not. Your speakers sound pretty good and people are happy with them, so that's cool. But again, I think the challenge presented to you using a round horn is a real handicap,and I prefer an asymmetrical horn.
What I will not stand for is your assertion that my "claims and responses here are not honest and candid". This is unacceptable, Earl. I don't want to argue with you either, and times like these always make me very uncomfortable because at the end of the day, I want to respect you. But when you come off like that, it just really is unprofessional, beneath your dignity and your standing in the field.
The 7.5 degree hole is from your own measurements. This was a case where your forward lobe is shifted upwards by a few degrees, enough to lift the lower null up almost to the baffle normal. The shift wasn't huge, but with null angles so shallow you don't have much room. That's my whole beef with round horns.
But round horns present a real problem. You can work with them, but they're hard to get the verticals right, or even acceptable, in my opinion. Is it a worthwhile tradeoff, maybe, maybe not. Your speakers sound pretty good and people are happy with them, so that's cool. But again, I think the challenge presented to you using a round horn is a real handicap,and I prefer an asymmetrical horn.
What I will not stand for is your assertion that my "claims and responses here are not honest and candid". This is unacceptable, Earl. I don't want to argue with you either, and times like these always make me very uncomfortable because at the end of the day, I want to respect you. But when you come off like that, it just really is unprofessional, beneath your dignity and your standing in the field.
gedlee said:
The 7.5 degree hole is from your own measurements. This was a case where your forward lobe is shifted upwards by a few degrees, enough to lift the lower null up almost to the baffle normal. The shift wasn't huge, but with null angles so shallow you don't have much room. That's my whole beef with round horns.
Re: Re: Re: Re: Re: Where to get the QSC?
Pretty smooth how you have co-opted the oppositions argument for yourself and use it against them. I believe it was the "round horn" camp that first said "nice idea Wayne, but what about the woofer?" Why does this have to be a "them versus us" thing? I don't put myself in any camp, and when you do the argument is no longer around facts but simply picking sides. And that is exactly what's happening now.
I understand where you are coming, and I like the idea IF it didn't cause other problems, for example power response. Now I'm not saying it certainly does, but it's a legit concern and I haven't seen it answered to my satisfaction. What I needed was a sim that could take into account a horn device, so I could plot what was in my mind's eye. But it seems you have just such a sim and your plots are exactly what I feared. From your sims, the vertical response (what angle?) of a 2 way with a round 90 degree CD horn:
Yep a dip at the crossover, hard to do anything about that in any speaker. Here is the same sim but with a rectangular 40 degree horn:
Still have the dip, but now the high frequency energy is much lower in relation to the woofer. No gentle slope. Just null at the crossover and nothing after that. Now tell me how this is smoother in power response? I'd rather have smooth power response with a narrow band dip, than the same narrow band dip and unsmooth power response. I just don't see a compelling argument for your method. You bring up lobing, but that is implementation specific and you can use the nulls to your advantage for reducing ceiling/floor bounce. Sometimes narrower may be better if you can do this. Give me a good reason to consider some other factors, because other wise I just don't see the advantage.
If you want to reduce HF reflection ABOVE the crossover, than maybe that is a good issue to consider in the trade offs, but that has nothing to do with lobing, crossover regions, DI matching to woofers, etc.
BTW what sim are using here?
Wayne Parham said:
Pretty smooth how you have co-opted the oppositions argument for yourself and use it against them. I believe it was the "round horn" camp that first said "nice idea Wayne, but what about the woofer?" Why does this have to be a "them versus us" thing? I don't put myself in any camp, and when you do the argument is no longer around facts but simply picking sides. And that is exactly what's happening now.
I understand where you are coming, and I like the idea IF it didn't cause other problems, for example power response. Now I'm not saying it certainly does, but it's a legit concern and I haven't seen it answered to my satisfaction. What I needed was a sim that could take into account a horn device, so I could plot what was in my mind's eye. But it seems you have just such a sim and your plots are exactly what I feared. From your sims, the vertical response (what angle?) of a 2 way with a round 90 degree CD horn:
Yep a dip at the crossover, hard to do anything about that in any speaker. Here is the same sim but with a rectangular 40 degree horn:
Still have the dip, but now the high frequency energy is much lower in relation to the woofer. No gentle slope. Just null at the crossover and nothing after that. Now tell me how this is smoother in power response? I'd rather have smooth power response with a narrow band dip, than the same narrow band dip and unsmooth power response. I just don't see a compelling argument for your method. You bring up lobing, but that is implementation specific and you can use the nulls to your advantage for reducing ceiling/floor bounce. Sometimes narrower may be better if you can do this. Give me a good reason to consider some other factors, because other wise I just don't see the advantage.
If you want to reduce HF reflection ABOVE the crossover, than maybe that is a good issue to consider in the trade offs, but that has nothing to do with lobing, crossover regions, DI matching to woofers, etc.
BTW what sim are using here?
Wayne Parham said:
What about the other plot of teh Yamaha? I like how in that one you didn't highlight and label the nulls in the 1.5-2.5 khz area like you did with Earl's. Just the one 45 degree angle where response had smoothed out. Might give the casual observer the idea that one was much smoother than the other. I see two speakers that both have nulls at low frequencies, but one has horribly ragged response through the entire HF region, while the other is incredibly smooth.
Brandon
I would mostly agree with you perspective. To me the jury is still out regarding any real advantage to non-axisymmetric patterns. I can see arguements both ways. I remain unconvinced, but certainly open minded enough to actually build and test an elliptical waveguide. What do I expect to happen? Actually nothing. I don't expect a significant or even audible difference either way - that's my guess. The reasons for this belief are well know I think. I object to this fore-drawn conclusion about the significance of the vertical polars and the witch hunt to find flaws in my designs that don't exist.
I would also like to reiterate my objection to "ground plane" measurements used in these kinds of comparisons. This technique has never been proven to be valid and I can see where it would ne very easy to achive much better performance from this technique. A full free field makes no assumptions and has to be considered as the standard. Until it can be proven that the two techniques yield the same data, I won't trust ground plane to be accurate.
I would mostly agree with you perspective. To me the jury is still out regarding any real advantage to non-axisymmetric patterns. I can see arguements both ways. I remain unconvinced, but certainly open minded enough to actually build and test an elliptical waveguide. What do I expect to happen? Actually nothing. I don't expect a significant or even audible difference either way - that's my guess. The reasons for this belief are well know I think. I object to this fore-drawn conclusion about the significance of the vertical polars and the witch hunt to find flaws in my designs that don't exist.
I would also like to reiterate my objection to "ground plane" measurements used in these kinds of comparisons. This technique has never been proven to be valid and I can see where it would ne very easy to achive much better performance from this technique. A full free field makes no assumptions and has to be considered as the standard. Until it can be proven that the two techniques yield the same data, I won't trust ground plane to be accurate.
About your "them verses us" post, I'm sorry, augerpro, but I have no idea what you're talking about. I've always viewed the loudspeaker as a system, not looking at just the horn. If I just looked at the horn alone in freespace, a round horn wouldn't look so bad. It's when combined with other drivers on a baffle, and when used in a room with relatively low (8 foot) ceilings that the asymmetrical flare looks better to me. Seems like you've read a few of the ideas on both sides of the asymmetrical / axisymmetrical argument, so I'm pretty sure you get the jist of the debate.
On the Geddes verses Yamaha measurements, I'm not comparing quality there. I'm only talking about the position of the forward lobe and the vertical nulls. I would also suggest that ceiling slap is reduced at HF when using an asymmetrical flare, but that wasn't the point in this recent exchange. We were talking about the null at 7.5 degrees, as shown in the measurements. That's the danger of having sound sources spaced too far apart vertically.
On the Geddes verses Yamaha measurements, I'm not comparing quality there. I'm only talking about the position of the forward lobe and the vertical nulls. I would also suggest that ceiling slap is reduced at HF when using an asymmetrical flare, but that wasn't the point in this recent exchange. We were talking about the null at 7.5 degrees, as shown in the measurements. That's the danger of having sound sources spaced too far apart vertically.
That the vertical nulls occur has never been questioned. That your speakers have less of an audible problem because they use a non-round horn is. Your blatant statement that your speakers sound better than mine because of this, and that I have been trying to mislead people about its significance is the issue. I have never mistated the situation and I have never attempted to mislead anyone. I'm not sure that you can say the same thing.
Wayne Parham said:
To me that is one legitimate issue where your method has an answer. But as it's above the crossover point, all of this talk about lobing and what not is an unrelated issue. I like the idea, my only problem is the narrower you go the higher the frequency where response flares at the horn's cutoff. This is implmentation specific of course, but I have a feeling most typical designs would run into conflicts with the narrower (vertically) horns versus the ability for the CD to play low or the woofer to play high. It is something I'm looking into now as I have a cheap design for a friend I'm doing that will most likely use a rectangular horn.
Wayne Parham said:
But so does the Yamaha Wayne! You just chose not to label those plots, but they are there nontheless. We can argue whether one is worse than other but to pretend that one has a problem while the other doesn't is just not the truth.
When two sound sources are stacked vertically, the height of the forward lobe is set by the distance between sound sources. Lots of other things come into play, but no matter what, the further apart the sound sources are, the narrower the forward lobe must be. There's no debating that.
You can optimize a speaker with distant sound sources, or you can screw up one that has closer spacing. But if you design them both with attention to getting the forward lobe right, the one with narrow spacing will yield the larger, cleaner forward lobe. Too far apart and there's virtually nothing you can do to make the forward lobe right.
I'm not putting words in your mouth, not saying you're misleading anyone, or anything of the sort. I'm talking about facts, using measurements to make my case. I sometimes illustrate the issues as I understand them, other times describe with words, but I've been using actual measurements to show real-world data.
I've always had a problem with round horns, and I've been consistent about the reasons why, since long before I ever knew you, Earl. It isn't a personal attack on you. I was making similar arguments against round tractrix horns 10 years ago.
I have a big problem with this kind of dismissiveness and rudeness, Earl. That has caused problems in the past, and I ask you not to continue it here again, now. It's not fair to characterize my comments the way you have.
About "significance" - the measurements are key here. I'd say a 10dB notch 7.5 degrees away from the baffle normal is significant. If you don't think it is, that's OK, but please stop saying that is misleading. It's a 10dB notch, Earl. It's straight in front of the speaker, and the narrow angle means it doesn't fall to the bottom edge of of the speaker until you're at least 10 feet back. I don't know about you, but I can hear that.
I don't think your measurements are free field, are they? You are using some sort of pseudo-anechoic measurement, as I recall. Seems like you said you do them indoors, with gating and smoothing. I also think you're using a hand-rolled measurement system, right?
I measure with a variety of methods and signal types. I regularly do ground plane, free space (suspended in air) and pseudo-anechoic (gated, using real-time signals such as MLS, impulse, noise or chirp).
In this case, I was using LMS, which is a pretty solid platform. It uses stepped sines, which is a proven and reliable method. I also measured outdoors, which removes boundary influence and doesn't require gating.
There are a handful of things that change with each signal type or measurement environment. Pseudo-anechoic can't reveal bass response, for example. Ungated measurements can show a notch from ground bounce, if the sound source is high enough from the ground. But one thing I don't see is movement of the position of the vertical nulls. As long as there isn't a boundary reflection that interferes, they pretty much look the same in any of these kinds of measurements.
You can optimize a speaker with distant sound sources, or you can screw up one that has closer spacing. But if you design them both with attention to getting the forward lobe right, the one with narrow spacing will yield the larger, cleaner forward lobe. Too far apart and there's virtually nothing you can do to make the forward lobe right.
gedlee said:
I'm not putting words in your mouth, not saying you're misleading anyone, or anything of the sort. I'm talking about facts, using measurements to make my case. I sometimes illustrate the issues as I understand them, other times describe with words, but I've been using actual measurements to show real-world data.
I've always had a problem with round horns, and I've been consistent about the reasons why, since long before I ever knew you, Earl. It isn't a personal attack on you. I was making similar arguments against round tractrix horns 10 years ago.
I have a big problem with this kind of dismissiveness and rudeness, Earl. That has caused problems in the past, and I ask you not to continue it here again, now. It's not fair to characterize my comments the way you have.
About "significance" - the measurements are key here. I'd say a 10dB notch 7.5 degrees away from the baffle normal is significant. If you don't think it is, that's OK, but please stop saying that is misleading. It's a 10dB notch, Earl. It's straight in front of the speaker, and the narrow angle means it doesn't fall to the bottom edge of of the speaker until you're at least 10 feet back. I don't know about you, but I can hear that.
gedlee said:
I don't think your measurements are free field, are they? You are using some sort of pseudo-anechoic measurement, as I recall. Seems like you said you do them indoors, with gating and smoothing. I also think you're using a hand-rolled measurement system, right?
I measure with a variety of methods and signal types. I regularly do ground plane, free space (suspended in air) and pseudo-anechoic (gated, using real-time signals such as MLS, impulse, noise or chirp).
In this case, I was using LMS, which is a pretty solid platform. It uses stepped sines, which is a proven and reliable method. I also measured outdoors, which removes boundary influence and doesn't require gating.
There are a handful of things that change with each signal type or measurement environment. Pseudo-anechoic can't reveal bass response, for example. Ungated measurements can show a notch from ground bounce, if the sound source is high enough from the ground. But one thing I don't see is movement of the position of the vertical nulls. As long as there isn't a boundary reflection that interferes, they pretty much look the same in any of these kinds of measurements.
Wayne,
" I'd say a 10dB notch 7.5 degrees away from the baffle normal is significant... It's straight in front of the speaker, and the narrow angle means it doesn't fall to the bottom edge of of the speaker until you're at least 10 feet back."
Both you and Earl recommend toeing in the speakers so that their axes cross in front of the listener; does the horizontal width of the notch extend that far?
" I'd say a 10dB notch 7.5 degrees away from the baffle normal is significant... It's straight in front of the speaker, and the narrow angle means it doesn't fall to the bottom edge of of the speaker until you're at least 10 feet back."
Both you and Earl recommend toeing in the speakers so that their axes cross in front of the listener; does the horizontal width of the notch extend that far?
gedlee said:
Post your verticals, Earl, which you have refused to do for over a year now.
I am particularly interested in how well your contention that these issues are mitigated when the listening axis is 22.5° may be substantiated by actual measurements.
gedlee said:
I have suggested the reason -- your relentlessly self-promoting posture that everything else sucks by comparison to your offerings, which assertion has not stood up well under scrutiny in this thread....
noah katz said:
This issue doesn't happen with movement along the horizontal. It happens when you move up or down, or angle the speakers up or down.
Take a loudspeaker using 90° round horn that has a null at 7.5° down. Put a book under the front edge of that speaker, tilting it back. It doesn't take much to lean it back 7.5°. Take a second speaker just like it and don't lean it back. Do an A/B test, you'll definitely hear the difference. The one that's tilted back has a pronounced hole in the crossover region.
Now take a speaker with an elliptical or radial horn, 90°x40° to 90°x60° or so. If properly designed, it shouldn't have a null until you've tilted it back (or forward) much further. Do the same A/B test. You'll be out 20° or further before you can hear the difference between the tilted one and the straightforward-facing one. By this angle, you'll not so much hear a hole in the crossover region as you'll hear the whole high end start to fall away, because you're nearing the edge of the pattern.
Now, you might ask, what's the point? Why do I care about how the speaker acts when pointed the wrong direction? That's not the way they're supposed to be used. The thing is, one of the main benefits of uniform directivity is good coverage, maintaining spectral balance and good sound quality from seat-to-seat. Part of this, in my opinion, is providing a good vertical pattern that has a reasonably tall forward lobe. People stand and sit, and they sit at different heights. Sometimes, they even like to do it on the floor.
I think most of us agree that a 90° arc of horizontal coverage is attractive for home hifi. Where we differ is in what's the best coverage for the vertical. I personally think that 40° to 50° is a good goal, mostly because it is attainable. This isn't a hard fixed value though, I could be happy with 30° if it were in the form of 15° above and 15° below. This would put it outside the baffle height a few feet back from the speaker, which I think is reasonable. The problem to me is when you have a null angle so close to the baffle normal (straightforward) that the null is basically right in front of the speaker through a good part of the listening rom, like for ten feet back or something. That's a problem if you asked me.
Some may not agree, in fact, I know not everyone reading agrees. But for those of you, I challenge you to do this experiment and A/B your speakers with a book under the edge. Listen to one speaker on a null axis and compare to another on a forward axis. If you can't hear the difference that 10dB notch causes, get your hearing checked because it is definitely audible. This isn't an academic excercise, interference nulls are quite audible.
There's good news too, a way to reduce the problem. If you have speakers with lower null at 7.5° down and upper null at 22.5° up, the best thing to do is to have stands built that angle the speaker down 7.5°. That splits the difference, properly aligning the forward axis. This orientation is most important if the speakers are placed on stands a foot or foot and a half up. The angled stands won't solve ceiling slap, but they will move the lower null down away from the listeners.
"This issue doesn't happen with movement along the horizontal."
So the notch isn't an issue if the speakers are toed in as you recommend?
So the notch isn't an issue if the speakers are toed in as you recommend?
IMO you spin things a lot more than Earl.
He hasn't "refused" to post them; as he said, he doesn't have the data, and is not at your beck and call to do generate it.
He hasn't "refused" to post them; as he said, he doesn't have the data, and is not at your beck and call to do generate it.
ZilchLab said:
Post your verticals, Earl, which you have refused to do for over a year now.
I am particularly interested in how well your contention that these issues are mitigated when the listening axis is 22.5° may be substantiated by actual measurements.
I have suggested the reason -- your relentlessly self-promoting posture that everything else sucks by comparison to your offerings, which assertion has not stood up well under scrutiny in this thread....
regarding solutions.....
There's the option of tilting the whole speaker to the top or bottom, naturally. One might ALSO move the waveguide , if, say, they were using a traditional "Woofer cab plus horn" arrangment.
Off the top of my head, it seems that one could move the waveguides towards the inside (center) edge of the speaker, and by adjusting mounting height, maintain the CTC. We're assuming a mounting system which would allow the front edge of the horn to overhang the front baffle. This would push the 'vertical' null off axis to the outside of the speakers, reducing any influence upon the axial listening area. That combined with a 'straight firing' arrangement (no toe-in) would push this dip outside of the main listening area, and keep the on-axis 8kHz dip suppressed for 'sweet spot' locations.
There's the option of tilting the whole speaker to the top or bottom, naturally. One might ALSO move the waveguide , if, say, they were using a traditional "Woofer cab plus horn" arrangment.
Off the top of my head, it seems that one could move the waveguides towards the inside (center) edge of the speaker, and by adjusting mounting height, maintain the CTC. We're assuming a mounting system which would allow the front edge of the horn to overhang the front baffle. This would push the 'vertical' null off axis to the outside of the speakers, reducing any influence upon the axial listening area. That combined with a 'straight firing' arrangement (no toe-in) would push this dip outside of the main listening area, and keep the on-axis 8kHz dip suppressed for 'sweet spot' locations.
noah katz said:
The notch is an issue, whether toe-in is employed or not.
What I suggested is tilt. If the vertical nulls are shifted up slightly, angle the speaker down slightly to compensate.
This should be done in addition to toe-in.
Toe-in serves a different purpose than tilt. It is done to improve imaging over a wide listening area. Crossing the forward axis slightly in front of the listening area makes the "sweet spot" be wider. What it does is to maintain relative amplitude balance between two stereo speakers, even when the listener moves from side to side. Toe-in is done to provide stereo balance over a fairly wide listening area, improving imaging.
Hopefully, the speakers have spectral balance - good response - over a wide horizontal range. The toe-in imaging trick requires that to work. The tilt I proposed is to make sure sprectral balance is also good over a range of vertical positions, coving most of the room vertically, so various seated heights will also have good spectral balance.
badman said:
That's true. If the woofer and tweeter are mounted on a baffle, you have to tilt the whole speaker to fine tune the position of the forward lobe. But if the waveguide is set on top of the cabinet - as is usually the case with wood horns for aesthetic reasons - then you can move the tweeter front to back slightly to adjust the position of the forward lobe. That was proposed by Altec many, many years ago. The thing is, you probably shouldn't move the horn too far back, or you'll get reflections off the top of the cabinet. I also suggest using an attractive horn cradle that is pinned to the cabinet to prevent movement. Once you've found the right position, you want it to remain fixed.
If I were to place Summas on a 16" stand, If I am doing my calcs correctly, at 12', i'd have a window of approximately 20-72" vertically to avoid nulls. This means that when sitting on the floor or standing up, I'm still OK. if I get closer to the speakers, lay on the floor or grow a foot taller, I'd be in quite a bit of trouble. I don't see myself listening on the floor, jumping up in the air repeatedly or listening from a step ladder. If I stay within this vertical window, what are they problems I should expect to hear?
dnewma04 said:
It's good you can sit 12 feet back from your speakers. I think many people need to be a little closer than that.
- Status
- Not open for further replies.