lobing of drive units

[Professor smith]

Phase, Time and Distortion in Loudspeakers

As soon as the diameter of the radiating surface (the cone) of a driver becomes "significant" with respect to wavelength, you will have problems with lobing.

I want to ask about this quote which is taken from the link above.

Firstly, what is lobing in a way i can understand? Does this mean first order crossover speakers should be limited to very small 4 inch or less cones? Any larger and you get lobing which means you would have to decrease the crossover frequency which means increasing demands on the tweeter.

 

[gedlee]

When the wavelength of sound is larger than the source, then all sources act like monopoles (or dipoles) radiating sound in all directions regardless of how the source acts or its shape, etc. As the wavelength becomes comparable to or shorter than the size of the source, the sound can and will radiate differently ihn different directions - it is no longer a "simple source". "Lobing" refers to the characteristics of a typical source which sends a beam (or several) of ever narrower width out in what is called a "lobe".

"Lobing" or "beaming" is not necessarily a bad thing, I do it quite deliberately, but this requires that the source directivity be well controlled so that this "lobing" is consistant in space and in frequency. No simple source or piston source can do this, which is why people use waveguides, because they can.

 

[Professor smith]

When the wavelength of sound is larger than the source, then all sources act like monopoles (or dipoles) radiating sound in all directions regardless of how the source acts or its shape, etc. As the wavelength becomes comparable to or shorter than the size of the source, the sound can and will radiate differently ihn different directions - it is no longer a "simple source". "Lobing" refers to the characteristics of a typical source which sends a beam (or several) of ever narrower width out in what is called a "lobe".

ok but why does it do all of that?

 

[Ron E]

Lobing is essentially like a diffraction pattern. Perhaps the easiest way to understand is to look at "planar" waves hitting a perforated barrier in a ripple tank. If the wavelength is longer than the perforation, the perforation acts as a simple source and converts the planar wave into wave which radiates as a semicircle. If the perforation is wider than the wavelength, this same thing happens at the edges, but the stuff in the middle just goes straight through (beaming). Play with Falstad's ripple tank sim (hours of fun) Ripple Tank Simulation

 

[john k...]

ok but why does it do all of that?

Any source of finite dimension can be though of as being composed of an infinite number of point sources distributed over the surface of the source. When the wave length is large compared to the source size, all these point sources sum in phase everywhere. Think of it like constructive interference. As the wave length becomes smaller the phase of each point source becomes dependent on the position in space where the the sum is observed with some point sources in phase and other out of phase. So there is both constructive and destructive interference between the sources. The results in a sound pressure that varies in intensity with spatial position. Lobing is a term used to discribe this variation in intensity with position. In two dimensions severe lobing looks like the petals of a flower. Each pedal being a lobe.

 

[DrDyna]

Play with Falstad's ripple tank sim (hours of fun) Ripple Tank Simulation

do NOT do this after you've had a few Tanqueray and cranberry, you'll be at it for an hour.

 

[AllenB]

Does this mean first order crossover speakers should be limited to very small 4 inch or less cones?

Your crossover slope will affect the width of the region of overlap, but otherwise not the frequency at which it happens, like driver size will.

Crossover slope will be more of a concern when you deal with the other lobing in this article, which is a matter of joining the sound from two separate drivers, and matching them in phase.

 

[454Casull]

You won't get lobing, you'll get beaming.

You only get lobing from interference with another source.

 

[weltersys]

You won't get lobing, you'll get beaming.

You only get lobing from interference with another source.

A large speaker, say 15", is not a single source at high frequencies. Run it full range, and it will both "beam" and "lobe".

 

[Professor smith]

so will any bookshelf speaker with a small drive unit beam if the crossover is set too high?
how does that tend to sound?

 

[pos]

Yes, the frequency at which lobs will appear is directly proportional to the cone diameter (and shape, at a lesser degree).
If you listen to pink noise with a fullrange driver and move your head you will clearly hear de lobs. The effect is similar to comb filtering, or phasing.

 

[454Casull]

A large speaker, say 15", is not a single source at high frequencies. Run it full range, and it will both "beam" and "lobe".

Yes, but diaphragm break-up is a separate issue from what OP is describing.

 

[meko]

LOBING - always worth preventing?

Sorry to "hijack" this thread, but I just felt like it would fit in here and make more sense than opening a new one.

I want to know what you think about lobing. Specifially, lobing in larger sound systems producing so much SPL that multiple instances of the same driver radiating the same band are becoming necessary.

When I look at the different approaches of either trying to create point-sources with multiple drivers or creating line-sources radiating quasi-cylindrical wavefronts and the complex time-correcting waveguiding often involved in these systems, I sometimes wonder if it's really worth it.

My question is: since efficiency is not as important as is directivity and wavefront control these days, what do I do if I am after the absolute maximum in efficiency and still need multiple drivers? All the approaches to create point or line sources seem to be an extreme distortion of the "ideal horn". I know we have a strong community on here that is in love with the synergy concept or paraline-alike fold systems, so don't get me wrong. I like the real waveguiding science involved very much and find it quite intriguing. However, what if I want round, straight horns that are aimed at minimum distortion and maximum efficiency and linearity?
What if I then want more SPL than, say, a typical 2" compression driver can deliver without overloading the throat with too much pressure?
What do I do? I guess the answer is live with the lobing- but what does that mean? How big do you consider the impact it has on overall experience?
I have heard lobing countless times, but unfortunately the system was horrible to begin with, so I really can't fully estimate the effect. And the only systems that really satisfied me were not using (needing) more than one driver per band and horn, so I guess I am unfortunately lacking the practical experience.
I guess that above a certain frequency, a multi-driver HF unit should be contstructed that is free of lobing and line sources 10 wavelengths long can be created relatively easily at these higher frequencies, so that makes sense.
But above what frequency do you think lobing becomes a no-go?
Do you think lobing is worse than anything else?
Even worse than the sound of inefficient line-arrays?

I would like to know your views on this, thanks in advance for your thoughts and ideas. kind regards!

 

[weltersys]

1)My question is: since efficiency is not as important as is directivity and wavefront control these days, what do I do if I am after the absolute maximum in efficiency and still need multiple drivers?
2) However, what if I want round, straight horns that are aimed at minimum distortion and maximum efficiency and linearity?
3) What if I then want more SPL than, say, a typical 2" compression driver can deliver without overloading the throat with too much pressure?
4)But above what frequency do you think lobing becomes a no-go?
5)Do you think lobing is worse than anything else?
6)Even worse than the sound of inefficient line-arrays?

1) Use a system that can combine multiple drivers to the horn providing the efficiency and directivity of your choice.
2) Horns "aimed" at maximum efficiency are neither minimum distortion or constant directivity.
3)A single 4" diaphragm 2" throat easily can achieve levels exceeding air linearity. See answer #1.
4) The lower the frequency, the wider the lobes are, and the more audible their effects become. At very high frequencies, lobing actually becomes pretty hard to detect. Since we seldom listen to one single (amplified) source point, lobing is a part of every sound system.
5) When lobing causes large deviations in frequency response, it may sound bad, but compared to being killed, no big deal- there are many things "worse" than lobing, for instance a lousy band or mix.
6) Inefficient line arrays don't have to sound bad, and line arrays don't have to be inefficient.

Art

 

[meko]

1) Use a system that can combine multiple drivers to the horn providing the efficiency and directivity of your choice.

Agreed. But for compression drivers, are you speaking of a manifold entrance like back in the old days? Since overloading the throat and creating too much distortion happens already with a single 2" driver, I guess that it would not be a solution without building a horn with a larger throat?!
However increasing the throat area of a horn where 2 or 4 drivers enter would then limit HF extension because of the large diameter wouldn't it?

2) Horns "aimed" at maximum efficiency are neither minimum distortion or constant directivity.

I was referring to something like a theoretically "optimal" flare, a circular, straight, correctly flaring horn that allows a true spherical wavefront to form and provide true loading to the driver and that is not undersized, covers no more than 2-3 octaves and uses no extreme compression ratios. CD is not the goal.

4) The lower the frequency, the wider the lobes are, and the more audible their effects become. At very high frequencies, lobing actually becomes pretty hard to detect.

That makes sense. I guess one could say that it should in many cases be possible to achieve 1/3rd or even 1/4th wavelength spacing between radiators in the range below 200Hz or so. So you are saying lobing is actually most apparent in the ranges above the bass? I always hear people talk about the importance of avoiding multiple HF drivers without phase-flattening solutions as found in line arrays. Could you elaborate on that point? I know that in your line array design you switched from a conical HF horn to a phase-correcting "paraline" line source horn radiating cylindrical wavefronts. Was there no significant difference in lobing? Did you find the lobing of the conical stack of HF horns to be acceptable?

5) When lobing causes large deviations in frequency response, it may sound bad, but compared to being killed, no big deal- there are many things "worse" than lobing, for instance a lousy band or mix.

6) Inefficient line arrays don't have to sound bad, and line arrays don't have to be inefficient.

I'm almost certain yours sounds very good, but the sad truth is they way too often don't.

Thanks Art and all the best!

 

[weltersys]

1)But for compression drivers, are you speaking of a manifold entrance like back in the old days? Since overloading the throat and creating too much distortion happens already with a single 2" driver, I guess that it would not be a solution without building a horn with a larger throat?!
However increasing the throat area of a horn where 2 or 4 drivers enter would then limit HF extension because of the large diameter wouldn't it?
2) So you are saying lobing is actually most apparent in the ranges above the bass?
3) I always hear people talk about the importance of avoiding multiple HF drivers without phase-flattening solutions as found in line arrays. Could you elaborate on that point? I know that in your line array design you switched from a conical HF horn to a phase-correcting "paraline" line source horn radiating cylindrical wavefronts. Was there no significant difference in lobing? Did you find the lobing of the conical stack of HF horns to be acceptable?

1) The "way old" days of RCA and JBL dual drivers used curved plumbing to manifold 2" exit drivers, which eliminated the lobing from multiple horns, but caused some serious dips and peaks in the response. Gunness mid 1980's manifolds reduced the response dips and peaks, but the higher powered drivers used revealed the new problem of 10 pounds of SPL shite in a 5 pound throat.
And yes, throat width must be limited to less than a wavelength to avoid dispersion and bandpass problems.
2) No, that's not what I said.
Lobing can be most apparent at low frequencies, but is most annoying in the speech range, where anyone can tell if vocals "sound funny", and our hearing is most sensitive.
3) I had previously used multiple side by side very narrow dispersion conical horns. Had I used enough (around 7 horns for a nominal 90 degree horizontal coverage) with high frequency drivers physically small enough to describe a 90 degree segment of an arc, with horn mouths adjacent, the lobing would have been reduced to an acceptable level, but still would not be as "seamless" as a single source.
Next, I used vertically stacked 90 x 30 horns, eliminating the horizontal HF lobing problem, but the overlapping patterns still made for nasty vertical lobes, and the flanking mid drivers (which continued the 90 degree HF horn walls) created horizontal lobes in the mid range due to the wide center to center spacings.

After converting the PA to using Paraline HF and offset horn loaded mids, all the driver's output radiation are now within 1/4 wavelength horizontally, virtually eliminating horizontal lobing.
With the array describing a vertical arc, vertical lobing is also for the most part eliminated.
When set up more "flat front", there is still some lobing apparent in the near field, the destructive off axis interference pattern reducing level, which helps keep SPL consistent from front to back.

Using the Paraline, like any device radiating "cylindrical wavefronts" (over some frequency range) is not without some (for the most part, correctable) frequency response problems, but the very even coverage and SPL density using lightweight modular pieces it affords is a great improvement over the previous deployments.

Art

 

[meko]

Thanks Art, it is so much appreciated. I understand the point you made about sensible tradeoffs.

I guess up until 500...600 Hz I can fit multiple 8" or 6" drivers into a shared horn if I do it right.
My Problem begins with the 2" compression drivers. I was going to go a rather "vintage" route with non-metallic diaphragms this time, thinking about jbl 2482 or fostex phenolic drivers. But I don't know if I should be putting something like a paraline or multipath-delay-line-array horn in front of them in order to get something like line sources.
Where do you cross over to your DH1AMT? I quite like the sound of the DH1A drivers, did you feel the paraline did something weird to the quality? Probably not I guess. You said the response is slightly altered- could you elaborate on that? Thanks again and all the best.

 

[weltersys]

1)My Problem begins with the 2" compression drivers. I was going to go a rather "vintage" route with non-metallic diaphragms this time, thinking about jbl 2482 or fostex phenolic drivers.
2)Where do you cross over to your DH1AMT?
3)I quite like the sound of the DH1A drivers, did you feel the paraline did something weird to the quality? Probably not I guess. You said the response is slightly altered- could you elaborate on that?

1) Your problems will be made much greater using separate mid and HF drivers and horns in a line array. Though I have fond memories of the 2482, I would not consider it (or any other driver not going up to around 16 kHz) for line array use.
2) The 8" mid drivers cross over to the DH1AMT/Paraline acoustically around 1000 Hz, the electrical crossover for the mids is lower.
3)The Paraline makes the frequency response far rougher than it would be on a conical horn. I don't recall saying "slightly altered", though the response can be corrected with EQ, the deviations are fairly major as can be seen in the comparison of a DH1AMT on a conical horn, on a Paraline, and 5 Paralines with mids equalized. Both non-smoothed and smoothed (as seen in most charts) are shown.
With EQ correction, nothing "weird" about the sound quality, and the pattern control is excellent, so everyone gets to hear a very similar presentation even though multiple drivers are used.

 

[meko]

Thanks Art! The graphs are quite interesting. How do you think these deviations actually occur? (Damn, I haven't gotten through the whole "square pegs" thread yet ) Are they simply cancellations/peaks caused by the dimensions and shape of the paraline path?
I know I will be having quite some issues with a multi-way, separate-horn array using multiple drivers in each passband. I think I could either go the "grateful dead" route, stack like there's not tomorrow and then deal with the lobing, or actually shape the wavefronts. I don't like either of these too much and what's worse is I don't know which has the worst overall effects on the final product, considering not only sound, but aesthetics, weight, total efficiency and the "aura" of the system, which is kind of important to me. It's really hard for me to see it clearly at the moment.
If I consider wavefront-shaping, I think I would go the rather "low-tech" route of the classic micro-capillary pathlength-difference systems like this rather nicely engineered example by turbosound, since total horn depth is not that much of an issue and the creation of the parts would be manageable using cnc routing machines, which I fortunately have access to.

An externally hosted image should be here but it was not working when we last tested it.

I would imagine it should be possible to shape the wavefronts of pretty much any compression driver using these time-delay systems. I guess it is a sensible approach and probably has less impact on experience quality than lobing, especially in the "vocal band", as you pointed out.
Of course another problem is the crossover area from, say, a line-source array of massive phenolic 2" drivers to a line-source array of 1" drivers in the area of 5...7k. Maybe I could get away with filters a little steeper, but maybe not.

Another issue is the horn shape. I would have loved circular horns, but it becomes very much impossible if I want them to radiate quasi-cylindrical wavefronts... well well... decisions.

 

[Bill poster]

I've recently noticed a trend for 3 inline/line array tweeters, one design using a kind of waveguide- which seem to work well; is this because of the central tweeter somehow lessening the effect of lobing and adding a more spacial sound?