May I ask why a wide dispersion speaker is preferred?

[Duke]

OT

Paging anyone selling a pair of Geddes speakers.

I just e-mailed a link to your post to someone who has a pair of Abbeys for sale (or did as of a week or so ago).

 

[FrantzM]

Thanks Duke.
I did find a pair twice. Logistics were an issue as the sellers wanted local pickup .
Frantz

 

[Chromatischism]

"The idea of an “ideal” directivity will be examined and it will be shown that CD alone is not enough, in a small room one needs a narrow
directivity that is also CD." - Geddes

So maybe we should be matching directivity to room size. Maybe there is an ideal range of directivity to room size ratios in which we would all receive roughly similar amounts of reflected wall energy.

 

[suttondesign]

Based on my experience building multiple Linkwitz designs, which are designed to have delayed reflections off the surfaces and ordinary things in a room, I believe Siegfried was probably right that the brain uses the delayed reflections and reduced SPL from reflected sounds as spatial cues to, in effect, process the direct energy in a manner that gives it priority or extra weighting. Relatedly, I found that the Dutch & Dutch 8C, while tremendously smooth in response, did not have the same quality of soundstaging, which I attribute to it having less room interaction than the Linkwitz designs, which are either dipole figure-8 or else cardioid patterns.

I also believe, based on those experiences and my time on this site, that the reflected energy should have a similar profile to the direct energy. I think the Linkwitz are pretty good in this respect, but my ears tell me they can't hold a candle, in response smoothness top-to-bottom, as the Dutch 8C (for example). My dream speaker would be a Linkwitz LX521 with a thoroughly-tested and tweaked response curve.

 

[Duke]

So maybe we should be matching directivity to room size. Maybe there is an ideal range of directivity to room size ratios in which we would all receive roughly similar amounts of reflected wall energy.

By "small room", I think Earl means "the size rooms we normally see in home audio." This is to distinguish from a "medium-sized" room, like a college lecture hall; or a "large room", like a concert hall. At least, that's the way I've heard him use the term.

I do agree with your idea of matching the speaker's radiation pattern to the room. I would look at path lengths and the direct-to-reverberant energy ratios. To a crude first approximation, ime the larger the room (and therefore the longer the reflection path lengths), the lower the direct-to-reverberant sound ratio can be, BUT for the sake of clarity we still want the ear to be able to distinguish the direct sound from the reflections.

Based on my experience building multiple Linkwitz designs, which are designed to have delayed reflections off the surfaces and ordinary things in a room, I believe Siegfried was probably right that the brain uses the delayed reflections and reduced SPL from reflected sounds as spatial cues to, in effect, process the direct energy in a manner that gives it priority or extra weighting. Relatedly, I found that the Dutch & Dutch 8C, while tremendously smooth in response, did not have the same quality of soundstaging, which I attribute to it having less room interaction than the Linkwitz designs, which are either dipole figure-8 or else cardioid patterns.

Agreed.

One way of looking at a dipole's figure-8 radiation pattern is this: It's like you take the radiation pattern of a wide-pattern speaker, chop it in half, and fire half of it backwards. So now you have the benefits of a somewhat narrow pattern as far as minimizing early reflections, plus the additional relatively late-onset reflection energy from the backwave, assuming the speaker is far enough out into the room. Linkwitz recommends a minimum of 3 feet as I recall, and applying Earl Geddes' suggestion to dipoles, we come up with about 5 feet.

I think the additional relatively late-onset reverberant energy - the backwave - does a good job of presenting the spatial cues on a good recording, without the "small room signature" cues inherent in early reflections. To my ears, the improvement in soundstage depth (with a good recording) from having dipole speakers five feet out into the room, as opposed to merely two feet out into the room, is far greater than one would expect from those additional few milliseconds of delay. I think that with good dipoles well set-up we literally hear "more of the recording venue" and "less of the playback room."

 

[LeftCoastTim]

I believe all natural sound sources are omnidirectional. So having a speaker that is closer to omnidirectional would sound more “natural”.

However, most speakers are not designed to control directivity, so room interactions would change their in room frequency response away from the ideal. Of course people noticed this and some go as far as to eliminate all room reflections.

But recorded music is far away from a natural sound, and many speakers don’t have a flat frequency response, direct or indirect.

Only way to verify any of this scientifically is lots of double blind tests using various room configurations and speakers.

I’m glad someone has already done this work.

 

[BradleyPNW]

Small room means all residential rooms. In contrast to movie theaters and auditoriums.

In general, the conflict is soundstage/spaciousness vs imaging. Wide = spaciousness. Narrow = imaging. Wide directivity leads to strongly correlated early reflections that confuse imaging. Narrow directivity improves imaging at the expense of spaciousness.

There was a "shootout" between a wide directivity Revel Salon 2 and a narrow directivity JBL M2. Both speakers were constant directivity with regard to frequency response. The people who participated preferred the Revel Salon 2. However, Earl Geddes said he treats his room a certain way. That means it's possible the room tilted the balance in favor of the Revel Salon 2. IOW, the room chose the winner. Maybe the JBL M2 would have won in a room treated the way Geddes suggests. (edit: I think I'm wrong about this. I assumed Salon 2 was wide directivity but it might be narrow like the M2.)

I tried to make speakers to get around the narrow vs wide problem. I combined a CBT array with panel dipoles. My theory was the CBT would supply even dB over distance + wide sound field directivity while the vibrating panels would create weakly -- rather than strongly -- correlated early reflections. I don't know how well they work because I don't have a golden ear. They sound good to me, a non-audiophile. They supply heaps of spaciousness because they're tall dipole arrays but I also hear fairly clear imaging. I'd have to sit in Earl Geddes' room to see what the difference is. OTOH, my speakers were extremely easy to make. Nowhere near the effort required to build Geddes New Summas & treat a room. My panel CBTs show lots of vibration above 1kHz and increase vibration as they increase in frequency. I think that results in weakly correlated early reflections that don't confuse imaging. The CBT portion also controls the vertical sound field. Sitting or standing, the arrays sound like they are at ear height, which helps with imaging. Oh, and the CBTs prevent floor and ceiling reflections. Yet the panels are home made so they require a lot of DSP equalization to flatten frequency response.

So that's the narrow vs wide directivity problem as I understand it.

 

[vavan]

I assumed Salon 2 was wide directivity but it might be narrow

if Salon2 is narrow what is wide enough then?

 

[Chromatischism]

It's considered wide compared to most speakers. That is a typical feature of a 3-way speaker that is using a small midrange driver.

 

[KaiserSoze]

... One way of looking at a dipole's figure-8 radiation pattern is this: It's like you take the radiation pattern of a wide-pattern speaker, chop it in half, and fire half of it backwards. So now you have the benefits of a somewhat narrow pattern as far as minimizing early reflections, plus the additional relatively late-onset reflection energy from the backwave, assuming the speaker is far enough out into the room.

Possibly worth mentioning is the means by which that figure-8 pattern is achieved. For very higher frequencies it is primarily due to ordinary directivity of the driver, assisted by the baffle; for midrange frequencies it is these same effects plus some interference/cancellation where the front and back wave meet at the sides; for low frequencies it is purely the interference/cancellation effect at the sides. To keep these individual effects in the necessary balance, such that the figure-8 has roughly the same shape for all frequency, likely requires exceptional skill.

 

[BradleyPNW]

if Salon2 is narrow what is wide enough then?

JBL M2's 120 degrees is wide. I should have described them as wide.

 

[Chromatischism]

By "small room", I think Earl means "the size rooms we normally see in home audio." This is to distinguish from a "medium-sized" room, like a college lecture hall; or a "large room", like a concert hall. At least, that's the way I've heard him use the term.

I do agree with your idea of matching the speaker's radiation pattern to the room. I would look at path lengths and the direct-to-reverberant energy ratios. To a crude first approximation, ime the larger the room (and therefore the longer the reflection path lengths), the lower the direct-to-reverberant sound ratio can be, BUT for the sake of clarity we still want the ear to be able to distinguish the direct sound from the reflections.

There's definitely a sweet spot. It's pretty obvious that (too) late reflections sound bad. All I have to do to test that is move away from my speakers and notice the increasing amount of echo, or try to listen to music in my hard-floor living room. It's uncomfortable. It's one step away from a gymnasium.

Out of curiosity I just tried various angles of toe-in on the front speakers, some of which I would call extreme. This is hard to A/B because I don't have an assistant to help me turn them, but with this crude test I can only conclude that while the time-intensity trade does indeed happen and quiets down the speaker nearest you when you move to the side (note: this only works with very consistent, but quieter, off-axis sound, like with the Buchardt S400 I have here), I wasn't crazy about the sound overall in the center. It's hard to explain but the system seemed a little too center-focused. Squaring the speakers with the room wasn't great, either. It seems I am still preferring crossing them just behind me. I will keep at it until I'm pleased with the results...

 

[Duke]

... I can only conclude that while the time-intensity trade does indeed happen and quiets down the speaker nearest you when you move to the side (note: this only works with very consistent, but quieter, off-axis sound, like with the Buchardt S400 I have here), I wasn't crazy about the sound overall in the center. It's hard to explain but the system seemed a little too center-focused.

Spreading the speakers somewhat further apart than normal may address the "too center-focused" issue. You can test the geometry without having to move the speakers by scooting your chair forward a bit.

 

[Chromatischism]

Spreading the speakers somewhat further apart than normal may address the "too center-focused" issue. You can test the geometry without having to move the speakers by scooting your chair forward a bit.

Can't move them further apart. They're 10 inches from the side walls already. The room is very narrow but I'm going to try another approach.

 

[yourmando]

Possibly worth mentioning is the means by which that figure-8 pattern is achieved. For very higher frequencies it is primarily due to ordinary directivity of the driver, assisted by the baffle; for midrange frequencies it is these same effects plus some interference/cancellation where the front and back wave meet at the sides; for low frequencies it is purely the interference/cancellation effect at the sides. To keep these individual effects in the necessary balance, such that the figure-8 has roughly the same shape for all frequency, likely requires exceptional skill.

I've been dreaming of building a mini 2-way dipole to go above my Linkwitz Orion dipole mains as "height" speakers for auro 3d & atmos.

But then I talk myself out of it because making the directivity perfectly smooth would probably be something I could only achieve by pure luck!

 

[Chromatischism]

Can't move them further apart. They're 10 inches from the side walls already. The room is very narrow but I'm going to try another approach.

I have the speakers further into the room now, still crossing behind me. There's more SBIR, but I think the staging is better and the sound is more pleasing. I think this room just isn't big enough to do the wide + extreme toe config.

 

[Neddy]

Wow...so, lots of Science here. But, like RayDunzi, I've been fortunate enough to compare my 'old tech' beamy system to M2s, so I'll throw in my observations.
In the listening position (also different rooms, mine is much larger than the M2s were in), there wasn't that much difference (yes, the M2s were spectacular wrt HF).
But what really 'got me' was that in the M2 setup, I could move around on the listening couch, or indeed around the room, and 'image' remained largely the same - amazing!!
Definitely a big move up from the 'head in the vice & it's GREAT' experience with my system.
So. Why don't I have M2s now? Cost.
If I were starting from new, and had the budget, I'd definitely go for a well designed, smoothly imaging system, especially if you share a couch with others, and move around a bit.
Experts - carry on.

 

[Bjorn]

But how many of you have actually listened to speakers that are truly uniform horizontally from approximately 500 Hz and up to at least 6 KHz (this covers the most important freq. area) with both wide and narrow dispersion and in the same room? Are you comparing apples to apples?

 

[Chromatischism]

Wow...so, lots of Science here. But, like RayDunzi, I've been fortunate enough to compare my 'old tech' beamy system to M2s, so I'll throw in my observations.
In the listening position (also different rooms, mine is much larger than the M2s were in), there wasn't that much difference (yes, the M2s were spectacular wrt HF).
But what really 'got me' was that in the M2 setup, I could move around on the listening couch, or indeed around the room, and 'image' remained largely the same - amazing!!
Definitely a big move up from the 'head in the vice & it's GREAT' experience with my system.
So. Why don't I have M2s now? Cost.
If I were starting from new, and had the budget, I'd definitely go for a well designed, smoothly imaging system, especially if you share a couch with others, and move around a bit.
Experts - carry on.

For that, look for constant directivity designs, ones that have largish waveguides, like the JBLs or Buchardts, or larger horns to control directivity even lower, like JTR speakers.

 

[MattHooper]

In general, the conflict is soundstage/spaciousness vs imaging. Wide = spaciousness. Narrow = imaging. Wide directivity leads to strongly correlated early reflections that confuse imaging. Narrow directivity improves imaging at the expense of spaciousness.

In my experience, the best I've ever heard of "combining" those two qualities came from my Thiel 3.7 speakers in my room.
The soundstage could be utterly massive and deep, just "melt the entire back wall away" with suitable recordings. And yet images were not diffuse; just the opposite, they were more focused, specified and dense than the majority of speakers I've heard. Most other speakers sounded a bit diffuse and unfocused in comparison. No idea how much of this was due to things like their particular coaxial drivers, time/phase alignment or whatever, but it was intoxicating. (I now have the slightly smaller version, the Thiel 2.7s which have similar characteristics).