Constant Beamwidth Transducer (CBT) Speakers

[Rick Sykora]

Um, proper CBT design is not exactly novel for either microphone arrays or speakers (same math). Don did some good work, but there are a few other contributors like Jim Flanagan, Jim West, Dave Berkeley, just for starters.

But as usual, Bell Labs designs were not commercialized.

Thanks for sharing the history.

I edited the opening post to note the additional attributions.

 

[Newman]

Well, if it is, something has gone wrong. You may get a specular reflection from the wall, but the beam is still the same, you've just acquired more specular reflections.

I guess I was referring to Keele and Button, Ground-Plane Constant Beamwidth Transducer (CBT) Loudspeaker Circular-Arc Line Arrays, 2005

"Conventional CBT circular-arc line arrays are designed to be used free standing and operate best when not near any large reflecting surfaces. This paper describes a design variation of the CBT loudspeaker line array called a ground-plane CBT array that is specifically designed to operate when placed against a planar reflective boundary. The variation consists of a modified array that is intended to operate near or very close to a single acoustic reflecting surface, such as a floor, wall, or ceiling."

cheers

 

[Bjorn]

First of all; the "CBT effect" is not destroyed when placed close to surfaces. The constant vertical directivity, avoidance of floor reflections, minimization of vertical reflections, and the horizontal directivity are unchanged. However, you have a higher level of specular reflections with close proximity to surfaces as well reinforcement in the lows. Just like any other speakers.

The horizontal directivity in a CBT speaker will vary depending on how the design is. It's possible to have equally a true uniform horizontal directivity as the vertical with the right design. The CBT36 has a generally a much more uniform directivity than most commercial speakers, but it also exhibits some side lobes. The horizontal directivity of the CBT36 is as wide as 180°.

At a certain frequency, baffle step will kick in and the speaker will radiate the sound backwards too. Just like other speakers do. This seems to be starting in the area of 200-300 Hz with the CBT36. In other words, the baffle step is really very low considering the narrow baffle of the speaker! Thus, the speaker does not operate like one single driver in regards to baffle step. Baffle step with one driver in a similar wide enclosure would be at a much higher frequency and therfore start to radiate a lot more energy backward earlier in frequency.

Here's a frequency response measurement in the listening position with the CBT36 placed up against the wall (green graph) vs 1.6m distance from the wall. 1/12 Oct. smoothing if I remember correctly.

And ETCs of both placements, showing the specular energy.

Close to the front wall:

1.6m distance from front wall:

 

[youngho]

Can you please explain why that is the case? Why do the Jbl CBT models seem to work mounted to Walls? My Epique CBT24 speakers seem to work quite well close to the wall behind them. Why is that the case?

I'm sorry, but @Newman says "the "CBT effect" is destroyed by being close to walls."

Well, if it is, something has gone wrong. You may get a specular reflection from the wall, but the beam is still the same, you've just acquired more specular reflections.

Yes, thank you. I find my Epique CBT24 speakers to work quite well close to the wall behind them, and I have seen more than a handful of installations with JBL CBT speakers mounted to walls. The second quoted text was meant to be a joke--I had just mentioned in the preceding post my own experience with a CBT design close to the wall behind them.

Young-Ho

 

[youngho]

youngho said:
Many users of controlled directivity loudspeakers find time-intensity trading to widen the sweet spot when set up so axes cross in front of the listening position and sometimes obviate the need for a discrete center channel speaker (compare with designs from Earl Geddes, Brian Waslo, Duke Lejeune, etc).

Perhaps you should just go back to 1933 and read Stienburg and Snow's paper? Or ask people who produce in stereo why the put emphasis on center-panned signals (voice or otherwise). The far-ear interference is well known and established, and happens regardless of the directional pattern of the loudspeaker. Some level shading can help, to some extent, yes, but you can't eliminate either the issue of conflicting HRTF's or conflicting ITD's vs ILD's.

I'm not sure why you're being so hostile here to other posters, they are, after all, referring to documented acoustic issues.

Of course, improper mixing for 3 front channels can create a problem, as well, but that problem is in production, not in 3-channel playback.

The mentioned names all make controlled directivity designs. I can't edit the original post that you're quoting, but I'd rewrite it as "Many users of controlled directivity loudspeakers find time-intensity trading seems to widen the sweet spot when set up so axes cross in front of the listening position, so much so that some don't seem to feel a need for a discrete center channel speaker (see descriptions of designs from Earl Geddes, Brian Waslo, Duke Lejuene, etc)."

Geddes (through attribution): https://www.audiosciencereview.com/forum/index.php?threads/cross-firing-speakers.25117/#post-857636
Brian Waslo: http://libinst.com/PublicArticles/Setup of WG Speakers.pdf
@Duke : https://www.audiosciencereview.com/...hantom-center-ust-use-case.26617/#post-912403

Young-Ho

 

[j_j]

I guess I was referring to Keele and Button, Ground-Plane Constant Beamwidth Transducer (CBT) Loudspeaker Circular-Arc Line Arrays, 2005

"Conventional CBT circular-arc line arrays are designed to be used free standing and operate best when not near any large reflecting surfaces. This paper describes a design variation of the CBT loudspeaker line array called a ground-plane CBT array that is specifically designed to operate when placed against a planar reflective boundary. The variation consists of a modified array that is intended to operate near or very close to a single acoustic reflecting surface, such as a floor, wall, or ceiling."

cheers

That's absolutely correct. If you design a speaker to work with a particular specular reflection, there y'go, and why not?

 

[j_j]

The mentioned names all make controlled directivity designs. I can't edit the original post that you're quoting, but I'd rewrite it as "Many users of controlled directivity loudspeakers find time-intensity trading seems to widen the sweet spot when set up so axes cross in front of the listening position, so much so that some don't seem to feel a need for a discrete center channel speaker (see descriptions of designs from Earl Geddes, Brian Waslo, Duke Lejuene, etc)."

Nobody (or at least not me) is arguing about the ability to use controlled directivity left/right (making speakers fixed to either left or right, too, if one is also notching out the wall reflection on the other side) to make things somewhat better, but until you use 3 channels of live data, there's still a very real problem. There is, even, with 3, a smaller problem, but the most major problems go away.

Now, Steinburg and Snow had no idea WHY this mattered in 1933, but now we do.

 

[youngho]

Nobody (or at least not me) is arguing about the ability to use controlled directivity left/right (making speakers fixed to either left or right, too, if one is also notching out the wall reflection on the other side) to make things somewhat better, but until you use 3 channels of live data, there's still a very real problem. There is, even, with 3, a smaller problem, but the most major problems go away.

Now, Steinburg and Snow had no idea WHY this mattered in 1933, but now we do.

@j_j , could you possibly comment on your own preferred upmixing algorithm or system among commercially available products in terms of reproducing stereo recordings?

Assuming use of an actual center channel, that is...

Young-Ho

 

[j_j]

@j_j , could you possibly comment on your own preferred upmixing algorithm or system among commercially available products in terms of reproducing stereo recordings?

Assuming use of an actual center channel, that is...

Young-Ho

It is very slightly too early to answer that, but yes, I have a preferred one.

 

[markus]

youngho said:
Many users of controlled directivity loudspeakers find time-intensity trading to widen the sweet spot when set up so axes cross in front of the listening position and sometimes obviate the need for a discrete center channel speaker (compare with designs from Earl Geddes, Brian Waslo, Duke Lejeune, etc).

Time-intensity trading with 2 speakers in a stereo setup unfortunately never worked for me (even have a pair of Geddes speakers). Guess my brain becomes too confused when time and level cues don't match.

 

[youngho]

Time-intensity trading with 2 speakers in a stereo setup unfortunately never worked for me (even have a pair of Geddes speakers). Guess my brain becomes too confused when time and level cues don't match.

Me neither (at least with Gradient Revolutions, Helsinki 1.5s, or 1.4s, also JBL Array 1400), but I've seen it reported multiple times. The CBT24s do the best job that I've experienced in terms of maintaining the relatively central image at non-midline listening positions. I use them with subwoofers, which reduces some of the boundary gain from near-wall speaker positioning but doesn't address SBIR effects above 80 Hz like @Bjorn shows above, but they're really fantastic family room speakers.

 

[j_j]

Time-intensity trading with 2 speakers in a stereo setup unfortunately never worked for me (even have a pair of Geddes speakers). Guess my brain becomes too confused when time and level cues don't match.

Likewise. My brain objects vigorously.

 

[Newman]

It is very slightly too early to answer that, but yes, I have a preferred one.

I’m also very keen to hear your thoughts JJ. Cheers

 

[youngho]

The mentioned names all make controlled directivity designs. I can't edit the original post that you're quoting, but I'd rewrite it as "Many users of controlled directivity loudspeakers find time-intensity trading seems to widen the sweet spot when set up so axes cross in front of the listening position, so much so that some don't seem to feel a need for a discrete center channel speaker (see descriptions of designs from Earl Geddes, Brian Waslo, Duke Lejuene, etc)."

Geddes (through attribution): https://www.audiosciencereview.com/forum/index.php?threads/cross-firing-speakers.25117/#post-857636
Brian Waslo: http://libinst.com/PublicArticles/Setup of WG Speakers.pdf
@Duke : https://www.audiosciencereview.com/...hantom-center-ust-use-case.26617/#post-912403

Geddes at http://www.gedlee.com/Papers/directivity.pdf: "Now, if somehow we could cause the nearer speakers level to fall while the farther speaker's level actually increased then we could partially offset the subjective effect of the image collapsing to the nearer speaker. (The farther speakers time delay will increase, but its level will increase as well.) Nothing can be done to correct for the time delay differences, but if we can make the level differences great enough then we might be able to achieve the effect that we want." and "Recall now our discussion of image shift in a stereo situation where we hypothesized that if the farther speaker could get louder as one moved laterally while the closer speakers level decreased, that we might be able to offset the time delay differences and maintain a fairly stable image with listening position. A glance at the drawing on the next page will show that if the central listening position is at about 22° off-axis of the main left and right speakers, i.e. toe-in, then the situation that we are looking for has been achieved. Namely, at position A, the sound from the two speakers is equal and the image is as required. At point B the closer speaker's level falls while the speaker farther away has an increase in its level. This will tend to offset the normal situation found where the image collapses to the nearer speaker whenever the listening position is not dead center. "

 

[Bjorn]

Geddes at http://www.gedlee.com/Papers/directivity.pdf: "Now, if somehow we could cause the nearer speakers level to fall while the farther speaker's level actually increased then we could partially offset the subjective effect of the image collapsing to the nearer speaker. (The farther speakers time delay will increase, but its level will increase as well.) Nothing can be done to correct for the time delay differences, but if we can make the level differences great enough then we might be able to achieve the effect that we want." and "Recall now our discussion of image shift in a stereo situation where we hypothesized that if the farther speaker could get louder as one moved laterally while the closer speakers level decreased, that we might be able to offset the time delay differences and maintain a fairly stable image with listening position. A glance at the drawing on the next page will show that if the central listening position is at about 22° off-axis of the main left and right speakers, i.e. toe-in, then the situation that we are looking for has been achieved. Namely, at position A, the sound from the two speakers is equal and the image is as required. At point B the closer speaker's level falls while the speaker farther away has an increase in its level. This will tend to offset the normal situation found where the image collapses to the nearer speaker whenever the listening p

I used to have Abbey speakers from Geddes. Actually three units and also a pair of Harpes (surrounds with 8" woofers).

The toe in in front of the listening never worked well in my ears. It sounded off and weird, so I directed them straight to the ears instead which worked much better. Geddes talked about avoiding the nearest reflections with strong toe-in. But his speakers lost directivity control fairly high in frequency, thus you didn't avoid audible reflections with strong toe-in. You simply avoid them in mainly the treble and upper midrange area, and leaving the rest of the spectral content reflective. This colors the sound considerably. Geddes point was the specular reflections below approximately 1000-800 Hz were not very audible, but he is wrong on this. It's still very much audible and the frequency response below this frequency also becomes very uneven when the directivity collapses.

This was very evident to both me and friends who listened to both the Abbeys and CBT36s in a narrow room with only 3.4m (11.15') width with no side wall treatment. Even though the CBT36 both has a very wide dispersion and some side lobes, the general even polar over a wide frequency area lead this speaker to sound far more correct and natural than the Abbey. The Abbey speaker were much more dependent on side wall treatment to sound good and get closer to the CBTs. And frequency response in different rooms were far more even with the CBT36s vs the Abbeys. I should point out that I also did the comparison with ceiling absorption and added a thick carpet for the Abbeys in some of the comparisons. None of mye friends liked the Abbeys much, but they all liked the CBTs.

Horns (which waveguides are) can be great and I love them, but they need to far bigger than both Abbey and Summa in order to have directivity control sufficiently low in frequency. Below is an indoor polar of a horn design I have with no gating. And when it's placed on top of a midbass horn, it extents the directivity control even lower. It both measures and sounds incredible but it's also very large.



If you want a stable image over a wide listening area, you need a speaker with a uniform directivity low in frequency, ideally down to the Schroeder frequency. And if you desire great imaging as well for either one or several listening positions you have to treat the side wall reflections. Unless you want a lateral contribution alá Toole. In the latter case one can make the wall diffusive as long as these reflections arrive sufficiently late.

 

[phoenixdogfan]

The idea seems pretty cool and all...

But am I the only one cringing at the acronym?

Perhaps if you had some CBT you could embrace the desire in yourself for CBTs.

 

[Newman]

Geddes at http://www.gedlee.com/Papers/directivity.pdf: "Now, if somehow we could cause the nearer speakers level to fall while the farther speaker's level actually increased then we could partially offset the subjective effect of the image collapsing to the nearer speaker. (The farther speakers time delay will increase, but its level will increase as well.) Nothing can be done to correct for the time delay differences, but if we can make the level differences great enough then we might be able to achieve the effect that we want." and "Recall now our discussion of image shift in a stereo situation where we hypothesized that if the farther speaker could get louder as one moved laterally while the closer speakers level decreased, that we might be able to offset the time delay differences and maintain a fairly stable image with listening position. A glance at the drawing on the next page will show that if the central listening position is at about 22° off-axis of the main left and right speakers, i.e. toe-in, then the situation that we are looking for has been achieved. Namely, at position A, the sound from the two speakers is equal and the image is as required. At point B the closer speaker's level falls while the speaker farther away has an increase in its level. This will tend to offset the normal situation found where the image collapses to the nearer speaker whenever the listening position is not dead center. "

I have always had an issue with that statment of Earl’s because it contradicts the precedence effect. He says, “if we can make the level differences great enough”…but we can’t. One seat across from dead centre and the delay is 2-3 msec, so the compensation in level needs to be 6-8 dB (Toole, 2008, Fig 6.3). That’s never going to happen. More like 1-2 dB.

So, no. Sorry Earl (I’m a big fan).

 

[markus]

Likewise. My brain objects vigorously.

Did you also try with headphones? It's been a while since I did that test myself but if memory serves well I could reliably steer the image within my head from left to right. So maybe the issue isn't necessarily our brains but the sound field in the room, i.e. either the speakers don't create the necessary ITD/ILD pairs or it's the room reflections.

 

[j_j]

. either the speakers don't create the necessary ITD/ILD pairs or it's the room reflections.

Yes, that is part of the problem. Done with both ITD and ILD matching, or close to matching, things work out substantially, and you can even get the front "out of your head".

 

[dannut]

Above about 2kHz ILD ques dominate the apparent source direction. Yeah, with broadband signals, conflicting ITD ques messes this up.
I also don't get the illusion, but with 'regular' speakers, it is much-much worse. Complete imaging collapse to near speaker, complete collapse with head turning etc. Toed-in-constant-directivity atleast maintains some sense of spaciousness in off-centre listening. In essence - 2 channel stereophony is... stupid. We have known this at-least from Steinberg and Snow.

There have been silent? progress from many fronts. Essentially 2 ways to approach it (only applies to existing 2 channel music recordings):

• make the 'triangle' infinite. That way the whole room is a sweet spot. WFS techniques can accomplish this by placing a virtual speaker far beyond the room boundary. Kinda what also a CBT array does in limited amount, that's why it sounds 'better' with 2 channels. Still problems remain due to head shadowing (2kHz notch problem) and head turning (bad(ly)/changing HTRF with frontal virtual source)
• decompose the 2 channels in to many using somewhat Ambisonic principles, that could be attributed to Michael Gerzon. I would consider @David Griesinger 'Logic 7' and Jim Fosgate-s 'PL2 Music' - due to manual adjustability, the best of the theme. As I have seen, atleast @Floyd Toole is also proponent of 2-many channel upmixing. Probably @j_j had psychoacoustically most elegant way doing it with 'THX Neural'. There are newer 'Dolby Surround' and 'DTS Neural X', but we like some adjustability, 'season to taste' if I may.

I had a desire to open a new thread with these topics, mainly to get up to speed what advancements have been made in this century.