Discussion: How important is smooth vertical directivity?

[napilopez]

Vertical directivity seems to be something of a contentious topic, and discussions have popped up in a few different threads on how to deal with uneven vertical directivity and the vertical reflections they cause. It's perhaps complicated because relatively few publicly available measurements provide a full frontal plane of vertical directivity, making it harder to discuss the performance of specific speakers. So I thought it might be worth having a specific thread to discuss the topic.

A few discussion topics:

-How audible is uneven vertical directivity (outside the listening window)?
-How much should speaker designers prioritize even vertical response?
-Barring a coaxial design, is wide horizontal, narrow vertical the way to go?
-How detrimental are ceiling and floor reflections to playback?

It seems like if smooth vertical directivity were essential to good sound, coaxial speakers would be the holy grail of audio. But while many coaxial speakers are very good, they're far from the final word. Conversely, many well-regarded speakers don't seem to have particularly smooth vertical directivity, at least once you go beyond the listening window of ±15 degrees or so. I assume uneven vertical directivity will affect the power response in-room, but it seems that as long as it 'balances' out, it isn't too audible.

 

[aarons915]

I've been interested in this topic for awhile after comparing my LS50 to some Revel M105 blind, I'm not sure how the LS50 could have won but they did and I've been looking for possible reasons why ever since. One of the only studies I've found on this subject is : "The Effect of a Vertical Reflection on the Relationship between Preference and Perceived Change in Timbre and Spatial attributes" in the AES journals. 59 of 78 of the ratings were positive regarding the vertical reflections but the interesting thing to me is that the reflection was a timbrally distorted reflection, similar to most vertically aligned speakers. Here's the graph to show what I mean:

I theorize that given a perfect reflection, like in coaxials or full range drivers, that preference would be just as strong as a horizontal reflection. More research needs to be done in this area in my opinion.

 

[napilopez]

I've been interested in this topic for awhile after comparing my LS50 to some Revel M105 blind, I'm not sure how the LS50 could have won but they did and I've been looking for possible reasons why ever since. One of the only studies I've found on this subject is : "The Effect of a Vertical Reflection on the Relationship between Preference and Perceived Change in Timbre and Spatial attributes" in the AES journals. 59 of 78 of the ratings were positive regarding the vertical reflections but the interesting thing to me is that the reflection was a timbrally distorted reflection, similar to most vertically aligned speakers. Here's the graph to show what I mean:

View attachment 38072

I theorize that given a perfect reflection, like in coaxials or full range drivers, that preference would be just as strong as a horizontal reflection. More research needs to be done in this area in my opinion.

Thanks for pointing that paper out - I hadn't seen it. Do you mean to say the simulated reflection artificially incorporated that ~2.7k dip to imitate a typical vertically aligned speaker?

It's a curious thing that we seem to prefer having vertical reflections around despite the noted timbral change. It would be interesting to see this experiment repeated without the dip.

 

[aarons915]

Thanks for pointing that paper out - I hadn't seen it. Do you mean to say the simulated reflection artificially incorporated that ~2.7k dip to imitate a typical vertically aligned speaker?

It's a curious thing that we seem to prefer having vertical reflections around despite the noted timbral change. It would be interesting to see this experiment repeated without the dip.

Yes the vertical reflection was measured and then simulated with another speaker. I agree, I think repeating the experiment without the dip would show even more preference, similar to how we perceive horizontal reflections.

 

[DonH56]

Most if not all multiway speakers exhibit a dip as you move off-axis vertically due to phase alignment of the drivers falling off. They are aligned for on-axis reproduction and as you move above or below the speaker the drivers' path lengths are enough to cause interference effects. It is rarely an issue because most of us sit in relatively the same vertical plane to listen. Standing up changes the sound but rarely are we (or maybe just "am I") listening critically when wandering around the room.

 

[MZKM]

Most if not all multiway speakers exhibit a dip as you move off-axis vertically due to phase alignment of the drivers falling off. They are aligned for on-axis reproduction and as you move above or below the speaker the drivers' path lengths are enough to cause interference effects. It is rarely an issue because most of us sit in relatively the same vertical plane to listen. Standing up changes the sound but rarely are we (or maybe just "am I") listening critically when wandering around the room.

The ceiling/floor reflections will be noticeably different than the direct sound, making them more of an issue than if they were as good as the horizontal performance. Most of us have carpet/rugs and ceiling fans though, so that helps lessen the severity.

 

[napilopez]

Most if not all multiway speakers exhibit a dip as you move off-axis vertically due to phase alignment of the drivers falling off. They are aligned for on-axis reproduction and as you move above or below the speaker the drivers' path lengths are enough to cause interference effects. It is rarely an issue because most of us sit in relatively the same vertical plane to listen. Standing up changes the sound but rarely are we (or maybe just "am I") listening critically when wandering around the room.

Sure, but if the listening-window sound were all that matered, we wouldn't be talking about directivity and the like all the time. I'd be curious to see blind test results between a top-level 'traditional' multiway vs a top-level coaxial.

Anyone have a KEF Blade and Revel Salon2 lying around their home?

 

[DonH56]

The ceiling/floor reflections will be noticeably different than the direct sound, making them more of an issue than if they were as good as the horizontal performance. Most of us have carpet/rugs and ceiling fans though, so that helps lessen the severity.

Not floor-ceiling reflections, though they matter as you said, but the delay from different drivers in the enclosure because they are not in the same horizontal plane and thus wave interference effects occur. From straight ahead the woofer and tweeter may arrive in phase at the MLP, but if you move above then the tweeter arrives before the woofer by a small amount. That causes ripples in the frequency response. If yo uperfectly time-align on-axis then it will still be slightly out of phase off-axis.

That said I completely agree that floor/ceiling bounce is (or can be) a big issue.

Sure, but if the listening-window sound were all that matered, we wouldn't be talking about directivity and the like all the time. I'd be curious to see blind test results between a top-level 'traditional' multiway vs a top-level coaxial.

Anyone have a KEF Blade and Revel Salon2 lying around their home?

How much off-axis response matters depends greatly upon your listening room. The "listening-window-sound" is pretty much all that matters to me as my room is fairly heavily treated. I also believe Dr. Toole's research showed we are much less sensitive to vertical directivity but I have not looked it up so may be wrong.

I have plenty of Salon2's but no Blades around; the only KEF's I have are much older models. And I don't care enough (or am just too lazy, take your pick) to haul them down two floors and hook them up in place of the Salon2's, sorry.

 

[oivavoi]

I think this is an important issue. Great OP, @napilopez!

My own subjective impressions: For listening in the near-field (1 m or so), with the speakers at a reasonable distance to walls and other surfaces, I find TBH that the off-axis response is not super important. I've been listening like this a lot, and find that speakers with so-so off-axis response can still sound really good this way. For me, the direct sound then seems to dominate and mask the off-axis response. The issue with off-axis response gets more and more important the more room sound one allows - i.e. the farther away one sits from the loudspeakers when listening. Horizontal vs vertical directivity? To me, both are important. The main reason being that I like to move around a bit when I listen to music. Being forced to sit exactly in the sweet spot feels unnatural to me, very different from real acoustic events. I want to have relatively good sound to the sides, and also relatively good sound when standing up. If I stand up, and the sound immediately gets very different, it breaks the illusion for me.

I therefore have no doubt that an ideal loudspeaker does have even vertical response. The question becomes how one achieves that, and at what price... A coaxial loudspeaker will typically have more narrow horizontal dispersion, for example, compared to some flat baffle speakers, and sometimes also some forms of distortion. Is that a trade-off which is worth it? Difficult to say.

 

[napilopez]

Not floor-ceiling reflections, though they matter as you said, but the delay from different drivers in the enclosure because they are not in the same horizontal plane and thus wave interference effects occur. From straight ahead the woofer and tweeter may arrive in phase at the MLP, but if you move above then the tweeter arrives before the woofer by a small amount. That causes ripples in the frequency response. If yo uperfectly time-align on-axis then it will still be slightly out of phase off-axis.

That said I completely agree that floor/ceiling bounce is (or can be) a big issue.



How much off-axis response matters depends greatly upon your listening room. The "listening-window-sound" is pretty much all that matters to me as my room is fairly heavily treated. I also believe Dr. Toole's research showed we are much less sensitive to vertical directivity but I have not looked it up so may be wrong.

I have plenty of Salon2's but no Blades around; the only KEF's I have are much older models. And I don't care enough (or am just too lazy, take your pick) to haul them down two floors and hook them up in place of the Salon2's, sorry.

How dare you not do what a stranger on the internet tells you to!

My impression on vertical directivity is that the research is far less extensive than on horizontal reflections and reflections in general. I'm also pretty sure I've seen Toole say somewhere that the first thing he'd treat in a room is the ceiling reflection, though I'm having trouble finding where (I believe it was either here or on AVS). In general, I'm concerned with untreated rooms, especially for speakers meant for home use, as that's how the vast majority of people listen (not to negate your experiences, just that's the focus behind the questions I posed at the outset of this thread)

I think this is an important issue. Great OP, @napilopez!

My own subjective impressions: For listening in the near-field (1 m or so), with the speakers at a reasonable distance to walls and other surfaces, I find TBH that the off-axis response is not super important. I've been listening like this a lot, and find that speakers with so-so off-axis response can still sound really good this way. For me, the direct sound then seems to dominate and mask the off-axis response. The issue with off-axis response gets more and more important the more room sound one allows - i.e. the farther away one sits from the loudspeakers when listening. Horizontal vs vertical directivity? To me, both are important. The main reason being that I like to move around a bit when I listen to music. Being forced to sit exactly in the sweet spot feels unnatural to me, very different from real acoustic events. I want to have relatively good sound to the sides, and also relatively good sound when standing up. If I stand up, and the sound immediately gets very different, it breaks the illusion for me.

I therefore have no doubt that an ideal loudspeaker does have even vertical response. The question becomes how one achieves that, and at what price... A coaxial loudspeaker will typically have more narrow horizontal dispersion, for example, compared to some flat baffle speakers, and sometimes also some forms of distortion. Is that a trade-off which is worth it? Difficult to say.

Definitely agree with you on near-field listening. This has been my impression as well, and it makes sense that the direct sound would overwhelm the reflected sound.

I'm different on my main listening setup though. I'm always in the sweetspot (at least when I care to listen intently), but it's far enough that there's still plenty of reflected sound.

 

[North_Sky]

A short educational article that touches on speaker's vertical directivity (and overall polar sound propagation) | • http://audiojudgement.com/speaker-lobing-polar-response/

There are many more great articles on the same subject (vertical response); I simply selected this one for its simple scientific easy to grasp to everyone, including us understanding.

 

[Hipper]

I think Floyd Toole suggests that we are immune to floor reflections as we are used to them in everyday life.

My speakers have ribbon high and mid range drivers and the designer said that their vertical dispersion was around thirty degrees. I listen around 1.6 metres from the speaker and my room height is 2.4m. When I temporarily placed some absorbent material on the ceiling it made no obvious difference to the sound, which perhaps is not surprising.

I've only ever read that ceiling reflections should be controlled.

 

[DonH56]

My ceiling is treated (covered with panels) and my floor has a heavy carpet and foam over carpet so "OK". IME ceiling bounce is often overlooked and larger than floor bounce because most lay a rug or two down. That said floor bounce in a room with a solid wood or tile floor can be very obnoxious and in those situations I find it worse than ceiling bounce due to the shorter path length. I have a big coffee table in my media room, and miss it for placing magazines and stuff (e.g. drinks, food), but it introduced a really nasty reflection that corrupted the midrange so I got rid of it.

 

[napilopez]

I think Floyd Toole suggests that we are immune to floor reflections as we are used to them in everyday life.

My speakers have ribbon high and mid range drivers and the designer said that their vertical dispersion was around thirty degrees. I listen around 1.6 metres from the speaker and my room height is 2.4m. When I temporarily placed some absorbent material on the ceiling it made no obvious difference to the sound, which perhaps is not surprising.

I've only ever read that ceiling reflections should be controlled.

Yea, floor reflections are another contentious issue. In his book, Toole makes a few points about them:

• When he tried to improve the floor bounce response on speakers, measurements looked better, but made seemingly no difference(he also says he wasn't 'exhaustive' on this matter though).
• Linkwitz seems to agree with him. He says something to the effect that the effect of the floor reflection is readily seen in measurements but "is not necessarily audible" on program material."
• A study (Silzle et al, 2009) showed that removing the floor reflection with absorbers leads to a negative listener response.
• He theorizes humans may have adapted to hearing a reflected floor sound and basically ignore it, although this doesn't necessarily mean that a floor bounce that is very timbrally different from the on-axis sound will be ignored. Just that in general, we're used to hearing a floor bounce.

But this is only briefly touched upon in the book overall. Notably, the section does not mention Søren Bech's 1995 study that simulates the effect of floor reflections. One of the participants in that study said "the single most disturbing reflection in the room is the floor reflection. That is what makes the speaker sound like a radio and not like the actual event.... The floor reflection absolutely must be handled, followed by the ceiling reflection, either by absorption or diffusion. "

Anecdotally, I've not noticed that treating the floor with a thick carpet makes all that much of a difference, but this isn't something I've tried extensively. Intuitively, I'd like to agree with Toole's thoughts on the matter, but my bias is definitely towards liking the idea that speakers can/should sound great without room treatment.

 

[jhaider]

One thing to consider is the general lack of vertical room symmetry. Most of us seek some degree of horizontal symmetry, so reflections tend to be similar. I don’t know about the rest of you, but our living room has different surfaces and fixtures on the floor and ceiling. This asymmetry means that reflections are altered differently anyway.

Anecdotally, in the last year I’ve had speakers with largely symmetrical radiation (Pioneer/TAD coaxes used from about 400Hz up), compact 3 ways with tight midrange-tweeter spacing (NHT C3), and two way waveguide speakers with relatively wide spacing (Amphion Argon 3S, JBL 705i, JBL 708i). On my desk I’ve had DIY speakers with 8” coaxes (KEF and DIY Sound Group) and waveguide speakers with comparatively wide spacing (JBL 705i, briefly and hopefully more in the future Neumann KH80). I can’t really attribute any differences I’ve heard between those speakers to vertical pattern. By contrast, I’ve heard things I think can be reasonably attributed to horizontal issues, such as upper midrange coloration from 2-way speakers with admirable on axis flatness but a large dispersion disruption from a largish midwoofer crossing to a tweeter loaded by a flat waveguide.

 

[DonH56]

How dare you not do what a stranger on the internet tells you to!

My impression on vertical directivity is that the research is far less extensive than on horizontal reflections and reflections in general. I'm also pretty sure I've seen Toole say somewhere that the first thing he'd treat in a room is the ceiling reflection, though I'm having trouble finding where (I believe it was either here or on AVS). In general, I'm concerned with untreated rooms, especially for speakers meant for home use, as that's how the vast majority of people listen (not to negate your experiences, just that's the focus behind the questions I posed at the outset of this thread)

No worries. My main point was that conventional speakers will not generally have smooth vertical response off-axis because the drivers themselves interact due to their physical positions.

 

[SIY]

No worries. My main point was that conventional speakers will not generally have smooth vertical response off-axis because the drivers themselves interact due to their physical positions.

Admittedly only 15 degrees in each direction, but still....

 

[DonH56]

Admittedly only 15 degrees in each direction, but still....

View attachment 38233

Pretty durn good! I don't think 15 degrees is enough for the drivers to get that much out of phase but honestly this is not something I care or know all that much about... At least not enough to grunge through the math. Especially when others here are far more qualified than I -- speaker design is not something I claim any expertise at.

 

[SIY]

Pretty durn good! I don't think 15 degrees is enough for the drivers to get that much out of phase...

I have seen much much worse, and rarely seen anything this good.

 

[617]

I have seen much much worse, and rarely seen anything this good.

Impressive preservation of top octave, but 15 degrees isn't much. +/- 30-40 is where I tend to see the big nulls.