I tried 60° flares and found them too directional for my tastes. They narrow the sweet spot too much, and make the people sitting off-axis hear sound that lacks spectral balance. Outside the pattern, the tonal characteristics change sharply. So too narrow a beam makes more of the room have this imbalance, and it just sounds unnatural to me.
I don't even like the sound when speakers are pulled out in the middle of the room so far there is a large portion of the room that gets rear radiation, which is also unbalanced. Doing that helps reduce anomalies from the nearest walls, but it still seems to sound unnatural. Ironically, when you pull the speakers too far out, it seems like you need to sit closer to the speakers to get more direct sound with spectral balance. It's sort of like how the sound shifts from live to dead when you exit the room and go down the hall.
So for me, I find the 70° to 90° horizontal beamwidth is most natural. Not sure if it is causal or coincidental that the DI of a trihedral 90° corner is equivalent to a 70° quadrahedral pyramid, but I do find that 70° to 90° horizontal beamwidth gives the most realistic sounding presentation for home hifi and home theater.
Last edited:
I also find a sweet spot there for a speaker at listening height. A much narrower horn may be longer, not fill the room's reverberant field as evenly and may result in a closed-in character to the sound of the horn. I'm not certain what causes this last one but I've noticed it even with one wall out of four being beyond a certain length at a narrow angle. The wider angle will be easier to design for.
P.S. I'm less sure with this in the vertical aspect and I seem to prefer a narrower pattern.
P.S. I'm less sure with this in the vertical aspect and I seem to prefer a narrower pattern.
Last edited:
I think you're right about rooms with awkward aspect ratios, especially small rooms. If one dimension is much longer than the others, it can be tough to get a good setup that works right. Another one that's potentially problematic is angled or gabled ceilings. You can get 'em sounding good, but it's harder. Speaker directivity becomes even more important in rooms like those.
As an aside, it seems like we sometimes overlook vertical directivity and talk mostly about horizontals, but I think verticals are important too. In fact, in some rooms, I think the vertical reflections (especially the ceiling) have more of an adverse affect than anything else. The best thing to do is to deal with the room, but sometimes that is impossible or impractical to optimize. But I think loudspeaker directivity can go a long way towards improving the sound.
Another aside, I often wonder if the reason some people prefer large horns is their vertical beamwidth control. I tend to avoid physically large tweeter horns because the center-to-center penalty, and the narrow forward lobe it creates. But a midhorn can be used to provide directivity control to a lower frequency. When used with appropriate crossover points, the vertical nulls can still be widely spaced, allowing for a nice clean forward lobe. I use a large midhorn in my constant directivity cornerhorns, and I think it helps both because there is no self-interference from the front wall or ipsilateral side wall, and also because the midhorn provides pattern control to a lower frequency.
As an aside, it seems like we sometimes overlook vertical directivity and talk mostly about horizontals, but I think verticals are important too. In fact, in some rooms, I think the vertical reflections (especially the ceiling) have more of an adverse affect than anything else. The best thing to do is to deal with the room, but sometimes that is impossible or impractical to optimize. But I think loudspeaker directivity can go a long way towards improving the sound.
Another aside, I often wonder if the reason some people prefer large horns is their vertical beamwidth control. I tend to avoid physically large tweeter horns because the center-to-center penalty, and the narrow forward lobe it creates. But a midhorn can be used to provide directivity control to a lower frequency. When used with appropriate crossover points, the vertical nulls can still be widely spaced, allowing for a nice clean forward lobe. I use a large midhorn in my constant directivity cornerhorns, and I think it helps both because there is no self-interference from the front wall or ipsilateral side wall, and also because the midhorn provides pattern control to a lower frequency.
Last edited:
I remember that Linkwitz once wrote something similar, but apparently his opinion has changed somewhat: Sound_field_control_for_stereo
I am somewhat surprised though that he still does not seem to consider the front wall reflection as being problematic under most circumstances. Perhaps he assumes that a sufficiently large distance between speaker and front wall is (or should be) maintained, 4-5 ft or so?
Please remember that he advocates for 6ms reflection delay as a minimum and that the front wall is diffusive. That's quite different from 4ft distance and a plain surface, which obviously does not work.
Obviuosly Linkwitz doesn't care about an accurate and precise image. I sure hope studio people don't buy into it. That's not the place to introduce a diffusive frontwall contribution and dipoles. It will have a detrimental effect on accuracy and areas like clarity, intelligbility, tonality and localzation will suffer. For home listening one can of course do whatever what one wants.
Is the front wall in Linkwitz's living room realy diffusive? Did he ever present in-room measurements?
Why 6ms? How did he arrive at that number?
rest assured that many people including rather educated ones have tried and approved it..what's funny is that you can go for a hike for weeks, come back here and still find the same people camping on the very same position and still not open to the slightest amount of curiosity.
We are not opening another endless and useless war like the one for the Behringer beating the hell out of the Orions are we?
We are not opening another endless and useless war like the one for the Behringer beating the hell out of the Orions are we?
Last edited:
I also wonder.. because 6ms does not seem obvious at all on the reflection threshold curves, but so does 10 or 15 ms too.. Actually the image shift treshold gets lower after 10ms and the detection thresold after 20ms.
No in room measurements.
He probably came to the conclusion about 6 ms because it fits a distance that works for many and it's a little less detrimental then what comes before. It has no real value in psycoacoustics.
His recommendation of a diffusice frontwall is also something more related to personal preferences rather then a goal of accuracy. That's fine though as long as people know what they are getting. Personally I would recommmend broadband absorption for frontwall and use diffusion behind the listening position with sufficient distance. Using diffusers alone at frontwall with dipoles will not only worsen the image, but it will also function as a filter in most cases because of the frequency area typical diffusers work in. This is much less of a problem at the rear wall because of the greater distance the sound needs to travel. But then, there are many things that are not optimal that can work ok.
His recommendation of a diffusice frontwall is also something more related to personal preferences rather then a goal of accuracy. That's fine though as long as people know what they are getting. Personally I would recommmend broadband absorption for frontwall and use diffusion behind the listening position with sufficient distance. Using diffusers alone at frontwall with dipoles will not only worsen the image, but it will also function as a filter in most cases because of the frequency area typical diffusers work in. This is much less of a problem at the rear wall because of the greater distance the sound needs to travel. But then, there are many things that are not optimal that can work ok.
Tonal neutrality is something Linkwitz is definitely interested in, so one would expect him to consider the front wall reflection more seriously.
Although, one can argue that a 2-channel representation is never 'accurate' compared to the live event and that many recordings may benefit from some spatial enhancement for casual listening. For studios one should stay away from this, obviously.
Hey - have you looked how he sets the speakers - they are like 2 meters from the front wall. He also tells that this is important! It is definitely the only real solution to this. We can see that his living room is far from ideal and symmetrical and has hardly any acoustical treatment done. My current sound and video setup
Coming to this discussion late, I am not aware of all that has been said. I, of course, have my position on this topic as well and in the first few pages I did not see it expressed.
It is not just CD that is important it is CD with a high DI (directivity index). CD alone is not a goal for me, but High DI CD is an absolute necessity. This gets to the room and reflections which seems to be lightly touched upon in the last few pages. But High DI CD has to be done without diffraction - read "smoothness" along any axis. Otherwise one has "thrown the baby out with the bathwater" as Wayne says.
I am also unfamiliar with a "war" between SEOS and Wayne, other than to say that I know both sides very well and they both can get very unprofessional so I am not surprised to find them at odds. I am at odds with both camps.
One thing that I did read early on which I absolutely have to comment on is the ranking of importance in a speaker that had "distortion" in second place. I am not sure, but I believe that by this "nonlinear" distortion is implied. I would like to point out that there is not a single shred of evidence to support such a claim, while there is a great deal that refutes it. Nonlinear distortion in a loudspeaker system can easily be made inaudible at any reasonable listening level. No so with diffraction (which becomes more audible at higher SPLs) or non-smooth DI (which cannot be corrected with any amount of EQ).
There is so much to say on "Uniform Directivity - How important is it", but I am just not the typist that many here are, otherwise I would write my lecture. It seems to me that my time is better spent on white papers which explain the situation and are available to all rather than in a thread where only a handful of people will read it.
For example, it is shocking to me how few people here at DIY are aware of all the work (readily available I might add) that has been done on the audibility of nonlinear distortion and how many continue to believe things that simply are not true. Better to discuss the important things like "Why a high DI is desirable" and why it needs to be smooth (though not necessarily flat). Those are aspects that matter - THD does not matter.
It is not just CD that is important it is CD with a high DI (directivity index). CD alone is not a goal for me, but High DI CD is an absolute necessity. This gets to the room and reflections which seems to be lightly touched upon in the last few pages. But High DI CD has to be done without diffraction - read "smoothness" along any axis. Otherwise one has "thrown the baby out with the bathwater" as Wayne says.
I am also unfamiliar with a "war" between SEOS and Wayne, other than to say that I know both sides very well and they both can get very unprofessional so I am not surprised to find them at odds. I am at odds with both camps.
One thing that I did read early on which I absolutely have to comment on is the ranking of importance in a speaker that had "distortion" in second place. I am not sure, but I believe that by this "nonlinear" distortion is implied. I would like to point out that there is not a single shred of evidence to support such a claim, while there is a great deal that refutes it. Nonlinear distortion in a loudspeaker system can easily be made inaudible at any reasonable listening level. No so with diffraction (which becomes more audible at higher SPLs) or non-smooth DI (which cannot be corrected with any amount of EQ).
There is so much to say on "Uniform Directivity - How important is it", but I am just not the typist that many here are, otherwise I would write my lecture. It seems to me that my time is better spent on white papers which explain the situation and are available to all rather than in a thread where only a handful of people will read it.
For example, it is shocking to me how few people here at DIY are aware of all the work (readily available I might add) that has been done on the audibility of nonlinear distortion and how many continue to believe things that simply are not true. Better to discuss the important things like "Why a high DI is desirable" and why it needs to be smooth (though not necessarily flat). Those are aspects that matter - THD does not matter.
Last edited:
I was just re-reading Toole's chapter on comb filtering earlier today. And it appeared that diffraction, from baffle edges, does not seem to an audible problem, i.e., it is not perceived the way it shows up in measurements. Dr. Geddes, can you point to some research that you or others have done that shows diffraction from baffle edges is audible?
It's a very different approach then CD speakers. This will give a spacious soundfield, but not an accurate one. Personally I don't like high gain specular reflections and I know this setup wouldn't impress me in the long run. Sure, 2 m from frontwall helps some, but you have to remember that dipoles send as much energy behind as they do forward. The result is often a quite poor frequency response besides specular reflections.Hey - have you looked how he sets the speakers - they are like 2 meters from the front wall. He also tells that this is important! It is definitely the only real solution to this. We can see that his living room is far from ideal and symmetrical and has hardly any acoustical treatment done. My current sound and video setup
Take a look at this setup with Orion speakers. The response is awful. 10 dB difference above 1 kHz! Believe me, you would not see that with a well designed speaker. Even if the room may have some major problems.
Attachments
I don't recall any discussion in Toole about diffraction - reflections yes, but not diffraction. The time delays for the two things are quite different.
My work in diffraction was within a waveguide and it was shown that a simulated diffraction was audible, but most importantly that its audibility increased with increasing SPL. This means that at some SPL the diffraction will always be audible no matter how low it is. This also means that diffraction will "sound like" nonlinear distortion in that it sounds worse and worse as the signal level goes up.
Admittedly the evidence for diffraction audibility is scant, but the evidence that nonlinear distortion is NOT audible is extensive. This is why I started looking at diffraction - I knew that there was something that made systems sound bad as the SPL increased, but I had learned that it was not nonlinear distortion. It has to be something and all available evidence, as scant as it is, points to diffraction.
You have to remember that all of Toole's work is done at a single SPL level. It just may be that this level is too low for the edge diffraction to be audible. One seldom finds something in the data that they are not looking for.
I think he address reflections that are 1 ms delayed from the main signal. That's about 1 ft. Not incomparable to a baffle for a 15" woofer and the delay that might result from re-radiated sound. His argument is that there is spectral smoothing that occurs when the brain gets successive "looks" at the same signal. And the spectral smoothing takes care of the notches that typically appear in a measurement of comb filtering. I'm summarizing, but you all should read chapter 9.
Regarding distortion, I have read your studies and tend to agree. On-axis response smoothness and flatness is king, followed by off-axis uniformity.
Regarding distortion, I have read your studies and tend to agree. On-axis response smoothness and flatness is king, followed by off-axis uniformity.
I know that he says that we can't hear the notches from comb filtering, but if that is his explanation then its just a guess. One that doesn't sound all that plausible to me. When summing localization is going on < about 2.5 ms our ear acts very differently than it does > 2.5 ms. And my recollection of Toole's book is that he did not have much data < 2.5 ms.
I think that this is well within that area where Toole and Griesinger and I are not all in agreement about things.
At any rate, it is the group delay from diffraction that I would worry about not the comb filter notches.
I've got some cabinets around 28"W that can clearly demonstrate the audibility of... something to do with edges. 
At the outside edge, they have a shallow recess, and then an edge that protrudes a couple mm past the baffle. I suppose there is both diffraction and reflection happening. Just listening to one speaker, the sound seems to be coming from inside the cabinet. If you put strips of 2" insulation down the edges, different sounds suddenly have different perceived depth along a line through the cabinet. It's not subtle. The difference in response measurement is pretty subtle.
It could just be something "special" about my non-ideal cabinets, though. It was my first time messing with something so wide, and the first time I noticed such an obvious change from edge differences.
At the outside edge, they have a shallow recess, and then an edge that protrudes a couple mm past the baffle. I suppose there is both diffraction and reflection happening. Just listening to one speaker, the sound seems to be coming from inside the cabinet. If you put strips of 2" insulation down the edges, different sounds suddenly have different perceived depth along a line through the cabinet. It's not subtle. The difference in response measurement is pretty subtle.It could just be something "special" about my non-ideal cabinets, though. It was my first time messing with something so wide, and the first time I noticed such an obvious change from edge differences.
This actually has been discussed throughout the thread in some detail. I agree with you in this point, and basically take the position that 70° to 90° horizontal beamwith is optimal for home hifi and home theater. Where possible, I like it to be constant from the Schroeder frequency upwards. But the matched-directivity two-way approach is the next best thing, where beamwidth narrows up to the crossover point, and then becomes constant above that.
Those are my preferences, others have stated theirs too.
No point in even bringing that up except to sling mud. I don't think we need you to act as an arbiter of professionality. Cool?
But if you must know, the hobbyists that plotted the SEOS flare profile and made the drawings for it are friends of mine. There is no bad blood from them, they're guys that frequented my Pi Speakers forum and the Econowave threads long before DIYsoundgroup was around.
The problem seems to have started when I said I wouldn't chose that flare profile in a private emails, and selected a different profile instead. Someone told 'em I wasn't going to use the SEOS horn, and this started a barrage of defensive comments on the messageboard where they're sold. The guy selling those horns took offense, and started commenting on the two designs - mine and his - which I felt I needed to counter with my own rebuttals.
Some of that is in this thread, but mostly we're talking about general design approaches, not specific brand names.
To be honest, Erich is non-technical, and it is really weird to even have an argument with him. For one thing, he isn't well versed with design details so his attacks are sort of half-baked. And really, the speakers he is selling are fundamentally similar to mine, just using a different horn. So some of the arguments he would use are the same things I would have said. It's like he was an echo.
But I do believe I have made better design choices all the way around, from waveguide shape to implementation and system design. My choices are driven from three decades of work on this exact loudspeaker paradigm, so I'd like to think my design approaches have matured over the years. I think there are more similarities than differences, but where there are differences, I think my choices make a more refined loudspeaker.
Beyond that, I don't want to see this thread descend into a series of ad-hominem attacks. I think it's bad form to keep bringing it up here. So can we keep the discussion on principles rather than personalities, please?
I have never had the resources to do controlled studies with groups of people. I don't have blind tests with lots of samples. As a hobbyist, I have always had to rely on empirical evidence, and on electrical and acoustic models and measurements.
But I have also never been swayed by audiophile psycho-babble. And I would agree that some things that can be measured cannot be heard. In fact, I believe that measurements are way more sensitive than our ears, in most cases. So we can measure stuff we would never be able to hear.
I'm saying all that to tell you that I have always generally agreed with you when you said THD isn't hugely important. I also agree that higher harmonics are more objectionable than lower harmonics, and that odd ones are worse than even ones. We're good with all that.
Further, I can believe that I might be swayed by the fact that I know a driver has a shorting ring, so I would have a preconceived notion that it would sound clearer. Maybe I just think the JBL 2226 sounds better than other similar drivers because it has a shorting ring. It does sound clearer to me, but that might be because I think it should.
But there is one metric I know I consistently and reliably hear as a difference, and that is listener fatigue. When I go to a trade show where I listen to music all day long for several days, if I take a loudspeaker model using a driver without a shorting ring, I'll go home at the end of the weekend with temporary tinnitus. I won't want to even listen to a thing for a day or two. But if I take the exact same model using a driver that does have a shorting ring, I am still refreshed at the end of the show. Same thing if I crank up the two pairs of speakers to concert levels. The one without a shorting ring makes my ears ring in a few hours, the one with a shorting ring doesn't.
To me, this is very good evidence, even if subjective and empirical. It is consistent and repeatable. And while I would call it subjective, other people have said they have the same experience, so it seems to be consistent across the board. The same things that reduce distortion seem to reduce listener fatigue. It seems to be most noticeable at high volume levels, or at lower volume levels that are played for an extended period of time.
I realize this is slightly off-topic, to talk about distortion in a thread about directivity. But I did want to respond to your comments about it, nonetheless.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.