A new approach to point source?

[TNT]

Andrew Jones is the creator of the three MoFi "SourcePoint

... and he is insightful enough to not call it a point source but rather play on the concept... source point instead of point source...

The way how I read the polar maps you shared, is that no matter where you stand encircling the speaker and no matter what frequency, the SPL is always constant. It doesn't seem like a point source.

But it is - that is what a real point source would look like. It may not "seem" to you, but pure technically, it is.

controlled directivity point source

well if the directivity is controlled - what is left of "point source"? A real point source would be 360 degrees omni...

A (real!) point source is identified by its dispersion and not its design. It is possible to create a point source that is not a very small item etc. Look at the old Bose design:

2 Mid Century Vintage Bose Corner Speaker 2201 Walnut | #152309443

This is for a pair of very early bose 2201 corner speakers. They are done in a walnut case and the speaker is covered in an awesome nubby textured fabric. On the front at the base of the speaker, also

www.worthpoint.com

if you put one of these in each corner of a room, you will have a point source behaviour thanks to reflections of the nearby surfaces.

But as usual - concepts are misused and mangled into confusion... that's all I wanted to say.

//

 

[TNT]

those waterfall plots

What you show is not waterfall presentations.

//

 

[Heinrich]

But as usual - concepts are misused and mangled into confusion... that's all I wanted to say.

As usual, exactly. What is a point, embedded in how may dimensions? Confusion.

One wants to, at least, mitigate the probs/ with the vertical suckout due to x/over and center to center separation of drivers. That was it, confusion not.

A ring of midrange drivers around the tweeter reinstalls proper symmetry, but at the cost of undisirable side effects. In parts the results are hilarious, because the attempt is ill advised to begin with. It is not the symmetry …

Thing is, the regular consumer only knows what the advertising says, rarely combines info from different sources due to lack of training in the field (which is very much o/k, granted) We are better, we read the plots and find, the wonder still has to take place.

Besides of idealistic desires, the infamous ‚wet dream’, what is actually needed? If we do not have a clue, let‘s devlop an idea on how to determine the needs, before we try to solve an only imaginated problem. Is such called engineering?

 

[thewas]

As it has been said most of such commercial home audio applications are rather failures but there are exist also few positive exceptions like for example the Perlisten models which use two more tweeter sized "mid drivers" lower in frequency in combination with an assymetrical waveguide to narrow the vertical directivity compared to horizontal one:

Directivity Pattern Control
DPC (Patent Pending) – We teamed up with companies in the USA and Sweden to bring modern materials technology to hi-fi, utilizing advanced Comsol acoustic modeling. The DPC array is the result of 18 months of painstaking simulations. We just love it and think you will, too.

Beryllium, covering the widest bandwidth up to ultrasonic frequencies, our finest 28mm Beryllium dome takes center stage. Beryllium is one of the strongest materials known to man and does not suffer from the typical harsh break-up modes of other metals due to its outstanding damping characteristics. Accurate, undistorted, and revealing.

TPCD Carbon Fiber – The Beryllium tweeter is flanked by our dual 28mm TPCD ultra-lightweight Carbon fiber midrange domes. Allowing the crossover point to extend more than an octave lower than typical domes, fast, lightweight and without beaming caused by larger diaphragms.

Waveguide lens – all of this technology wouldn’t be worth it without the waveguide orchestrating the directional characteristics of the array. Presenting a coherent wavefront, while optimizing early reflections and sound power, to bring vocal reproduction to an altogether new level.

 

[Arindal]

A ring of midrange drivers around the tweeter reinstalls proper symmetry, but at the cost of undisirable side effects. In parts the results are hilarious, because the attempt is ill advised to begin with.

What is the inherent and inevitable side effect of a midrange array around a tweeter? I refer to really unsolvable problems, not to a compromise of the practical design or the parts used. An 1" midrange is full of compromises and that's understandable.

for example the Perlisten models which use two more tweeter sized "mid drivers" lower in frequency

I understood the concept so that it is in fact two more similar tweeters being part of this waveguide, even identical ones in the R series. They are ensuring a mild transition in vertical directivity between the waveguide itself and the midwoofer array. I would rather call this a filler driver array than two midrange drivers, but the result is surely very smooth. If I recall it correctly, in the floorstander models all three domes forming the array kick in only at around 1.4K with the upper and lower dome being faded out beginning less than an octave higher.

So it is rather a 1.5-way tweeter array avoiding the crucial disadvantages of a conventional tweeter array by placing all three domes very close to each other and fading out the outer ones before ugly cancellation and audible phasing effects might appear. At 4K there are out of the game.

The downsides of this concept are a slight step up in directivity index from tweeter array to waveguide operation, as the latter has to be showing a narrower dispersing in the horizontal domain. The transition between array and woofers are not as low as necessary to eliminate vertical interference, but surprisingly they managed to retain a smooth transition in this band.

In practice the difference in directivity index between omnidirectional and line array behavior is down to a frequency band which is not critical and can surely be EQ´ed, around 400Hz. On the other hand I would see some EQ to be mandatory in most of rooms.

 

[Audionaut]

. and he is insightful enough to not call it a point source but rather play on the concept... source point instead of point source...

This goes more into semantics and is too complicated for me in the context of this discussion, because all major companies identify their coaxial designs as point source.

What you are describing is called omnidirectional in Germany.

Duevel offers interesting concepts

Duevel high-end lautsprecher

duevel omnidirectional high end loudspeakers - einzigartige highend lautsprecher mit einem perfekten rundstrahlverhalten - Music like life in concert. Highend Rundstrahler.

www.duevel.com

MBL anyway

High-End Radialstrahler-Lautsprecher von MBL

Endlich Musik erleben statt nur zu hören - mit den Radialstrahler Lautsprechern von MBL. Welche Technik führt zu diesem beeindruckenden Musikerlebnis? Lesen Sie hier alles zu den innovativen Lautsprechern von MBL.

www.mbl.de

As a practitioner myself, I currently use these coaxial drivers with very wide dispersion in the treble range with active DSP equalization and an 18-inch subwoofer.

B&C 8FCX51

8FCX51 8Ω - Coaxials - B&C Speakers

www.bcspeakers.com

There are many horn designs for very broadband compression drivers from the PA sector, but these are usually somewhat oversized for the average living room. The horn largely determines the dispersion characteristics and the woofers can be integrated into the coherent radiation quite easily with the fully utilized bandwidth (xover 200 to 400 Hz).

Overview

www.bmsspeakers.com

Overview

www.bmsspeakers.com

Celestion Axi2050 - AxiPeriodic Compression Driver

Celestion Axi2050 - High power AxiPeriodic compression driver with a neodymium magnet, titanium diaphragm and large surround area for minimum distortion

celestion.com

Coaxial HF - B&C Speakers

www.bcspeakers.com

 

[Arindal]

What you are describing is called omnidirectional in Germany.

Duevel offers interesting concepts

MBL anyway

To be precise, some people use the term ´omnidirectional´ solely for the aforementioned ideal acoustic pointsource which as a result is producing a perfectly spherical wavefront in all dimensions.

Duevel, MBL, German Physics and similar concepts are rather ´radial sources´ as their omnidirectional pattern is effective in only two dimensions, the horizontal plane. Vertically they might show more or less of a narrowed dispersion pattern (not necessarily a cylindrical one like one would expect it from a perfect radial source in theory).

 

[Heinrich]

This goes more into semantics and is too complicated for me in the context of this discussion,

Sloppy wording is part of the problem.

What is the inherent and inevitable side effect of a midrange array around a tweeter? I refer to really unsolvable problems, not to a compromise of the practical design or the parts used. An 1" midrange is full of compromises and that's understandable.

Isn‘t too easy to tell, so without irony?
Problem: suckout in the vertical under angle due to center center distance, I explicated on that above, you didn‘t pick it up
Solution: ring of drivers around tweeter

The joke: ring has center to center distance between drivers, but now extends the problem to the horizontal, and keeps it in the vertical

LoL

Now, please go ahead with the discussion. Think the original question is answered.

View attachment 453245

The ‚eye‘, is it upwards or downwards? Anyway, it is there. Problem not solved. Question, how bad is it. Answer: no clue, right?

 

[Arindal]

Problem: suckout in the vertical under angle due to center center distance,

Do not really see why this problem should be more critical here compared to any other vertical inline placement of tweeter and midrange, particularly d´Apollito style arrangements.

To keep this further away from a potential listening window, a lower x-over freq might help, or a partial overlap of drivers. The latest Perlisten models called A-series seemingly rely on a single tweeter in a waveguide overlapping both midwoofers.

The ‚eye‘, is it upwards or downwards?

The isobaric graph is hinting that it becomes noticeable at +15deg higher than tweeter axis, in a pretty narrow band. Not great, but given the fact it is something like an array speaker with cylindrical waveform and its tweeter waveguide showing a similar listening window, also not terrible.

 

[Heinrich]

Do not really see why this problem should be more critical here compared to any other vertical inline placement of tweeter and midrange, particularly d´Apollito style arrangements.

To keep this further away from a potential listening window, a lower x-over freq might help, or a partial overlap of drivers. The latest Perlisten models called A-series seemingly rely on a single tweeter in a waveguide overlapping both midwoofers.

The isobaric graph is hinting that it becomes noticeable at +15deg higher than tweeter axis, in a pretty narrow band. Not great, but given the fact it is something like an array speaker with cylindrical waveform and its tweeter waveguide showing a similar listening window, also not terrible.

As I said, two times for the time being, where is the problem? You talk down the vertical lobing affair, which to my knowledge is the motivation for designing what is called, mistakenly, a ‚point source‘. Problem statement: there is lobing in the vertical direction. My suggestion: don‘t fall for ‚solutions‘ that only make it worse. Estimate the hearing‘s tolerance, look how far the problem with lobing actually needs to be mitigated. Only when knowing the real task, start constructing.

I know too well, that such isn‘t the ideal DIY path. That‘s why DIYers are not automatically engineers.

To happy consumers: the only successful solutions I know of is coaxial, KEF, Genelec, or Perlisten. With Perlisten a question remains: is the ‚eye‘ in the plot above directed upwards or downwards? Up: huh, too bad, sorry. Down, congrats, well done!

 

[thewas]

The ‚eye‘, is it upwards or downwards? Anyway, it is there. Problem not solved. Question, how bad is it. Answer: no clue, right?

From my point of view the problem is quite well solved, as the downward "eye" is not very intense and better than probably 95% designs out there which try to combine wide horizontal with narrow vertical one. YMMV of course, in that case feel free also to show measurements of commercial implementations you find quite better.

 

[thewas]

I understood the concept so that it is in fact two more similar tweeters being part of this waveguide, even identical ones in the R series. They are ensuring a mild transition in vertical directivity between the waveguide itself and the midwoofer array. I would rather call this a filler driver array than two midrange drivers, but the result is surely very smooth.

That's why I had put "mid drivers" in brackets as this is the manufacturers text, for me they are also tweeters with the ability to be crossed low (around 1 kHz), also with the advantage of the 6 dB boost of having a pair.

 

[Heinrich]

From my point of view the problem is quite well solved, as the downward "eye" is not very intense and better than probably 95% designs out there which try to combine wide horizontal with narrow vertical one. YMMV of course, in that case feel free also to show measurements of commercial implementations you find quite better.

It is directing downward, you confirm? Well done in case.

For commercial speakers it could become a criteria whether or not the x/over steers it this or the other way. Not all users are willing, for the very best reasons in the world, to stick to ‚critical stereo listening’, remaining seated all the time. When standing up, the ‚eye‘ is better not directed up towards the listener‘s ears, clearly.

Didn‘t I already tell, KEF solves the problem entirely?

If we want to contribute to better products, we could in estimating by how much the presence of the ‚eye‘ disturbs the experience. There is nothing yet, afaik. To solve a problem w/o understanding it calls for making it worse, as is documented to its best by the so called mid array.

 

[Audionaut]

So is this design a modern day alternative to point source?

Sloppy wording is part of the problem.

There is no problem.
There is a question on this thread and there are various answers.

Concepts such as the Raduno, which was one of the first to describe this quasi-coax in the form it is currently becoming popular, offer solutions, but create problems elsewhere, as is always the case with loudspeakers.

There is not THE answer to loudspeaker design questions, there are only ever many answers with many compromises.

This applies to all loudspeakers, even if someone comes around the corner and praises Genelec or Grimm, for example, as acoustic marvels.
No, they are all just compromises that are sometimes more, sometimes less suitable for one or more applications.

I think the most perfect approach at the moment for a radiation system that can be customized in every respect is the Beolab 90

Beolab 90 – Ultimativer High-End 8.200 W Lautsprecher | B&O

Der vielleicht fortschrittlichste digitale Lautsprecher weltweit. Das Flaggschiff der Stereolautsprecher von Bang & Olufsen fügt sich perfekt in Ihren Wohnraum ein.

www.bang-olufsen.com

 

[PristineSound]

But it is - that is what a real point source would look like. It may not "seem" to you, but pure technically, it is.

Per ChartGPT. Looks like I gravely misunderstood what a point source is.

If based on this definitely of what a point source speaker is, then what is the difference between a point source and an omnidirectional speaker?

And why would anyone want a true textbook point source speaker? It would be horrific from a sound reflection perspective with virtually no imaging.

 

[olieb]

If based on this definitely of what a point source speaker is, then what is the difference between a point source and an omnidirectional speaker?

I would not trust ChatGPT for a single statement, and this is a good example. You will only ever get the generic internet "opinion" which might be BS (as in this case).
Here is what wikipedia has to tell about point source

A point source is a single identifiable localized source of something. A point source has a negligible extent, distinguishing it from other source geometries. Sources are called point sources because, in mathematical modeling, these sources can usually be approximated as a mathematical point to simplify analysis.

The actual source need not be physically small if its size is negligible relative to other length scales in the problem. For example, in astronomy, stars are routinely treated as point sources, even though they are in actuality much larger than the Earth.

In three dimensions, the density of something leaving a point source decreases in proportion to the inverse squareof the distance from the source, if the distribution is isotropic, and there is no absorption or other loss.

And that is about it. A point source is (practically) point like, that means it is small enough for all practical purposes that are to be considered. This results in all radiation coming out of the same small region in space (localised source). [If you are away far enough all sources become "small enough".]
This has little to do with omnidirectional radiation.
A dipole can be as point-like as a monopole, and the same holds for a cardioid, a quadrupole and so on. A point source can have any radiation pattern and it can change with frequency.

But there is a connection though. A perfect omnidirectional pattern over a wide range of wavelengths can only be created with a point source (monopole), because if two sources (or more) radiate at the same frequency from different points there will be interference.

The actual use of the terms might be different, as the answer of CHatGPT shows, in some bubbles "private" definitions exist.

 

[gnarly]

In my way of thinking, center-to-center (c2c) driver spacing is the starting consideration in any design.

When drivers' c2c's can be kept within 1/4 wavelength throughout their frequency summation ranges, they can radiate polars without ripple.

So the trick imo, in any of the ring/array type speakers that use mids around a central tweeter, is to first have a tweeter that reaches low enough to allow a c2c between the mids to be within 1/4 WL.
Then, the ring/array of the mids needs to have 1/4 WL spacing between them thru their operating range, and also reach low enough to pass the baton to the lows within 1/4WL thru crossover.
Here's a theoretical depiction of the above on a flat baffle, pulled from a Danley unity patent.

I think this is nearly impossible to accomplish on a flat baffle with regular drivers. And where 1/4WL c2c's break down there will be lobing. Then a design choice must be made where to aim the lobing. Ideally, any array type speaker would have no lobes, or destructive interferences between drivers.

The exceptions are arrays that achieve true line array behavior. Either vertically like conventional line arrays and ribbons, or both vertically and horizontally like electrostats.


I played with the ring radiator idea a little with spare drivers I had laying around...so not an all out attempt to optimize, at all.

I don't remember the crossover points, but they were all about the c2c comments above. Here's polars for 0-30 degrees in 10 deg steps, simply made indoors.
These were pretty bad in comparison to other speakers (MEHs) I had been working on as well, and only got worse past 30 degrees. (I use linear-phase tuning if folks are wondering how phase is so flat)

I gave up on the ring-radiator idea pretty fast.

I tend to take an abstract view of a MEH as way of putting rings of drivers around a central tweeter, but folded onto a pyramid to reduce the c2c distances of a flat baffle, such that 1/4 WL spacing become feasible.
here's one MEH, along with the ring speaker before it got its flush plate.

Doesn't all loudspeaker design come down to wavelengths and drivers spacings? And how much SPL & bass extension we want?

 

[thewas]

Didn‘t I already tell, KEF solves the problem entirely?

I own and really like current KEF loudspeakers but their designs do not offer the feature of different horizontal and vertical directivity which for some setups and preferences can be advantageous.

This applies to all loudspeakers, even if someone comes around the corner and praises Genelec or Grimm, for example, as acoustic marvels.
No, they are all just compromises that are sometimes more, sometimes less suitable for one or more applications.

Very true!

I think the most perfect approach at the moment for a radiation system that can be customized in every respect is the Beolab 90

And even this one only variates only the horizontal radiation, so there is still a long open way for the future.

In my way of thinking, center-to-center (c2c) driver spacing is the starting consideration in any design.

When drivers' c2c's can be kept within 1/4 wavelength throughout their frequency summation ranges, they can radiate polars without ripple.

Exactly, this would be also my personal way to differ between point source and not.

 

[PristineSound]

I would not trust ChatGPT for a single statement, and this is a good example. You will only ever get the generic internet "opinion" which might be BS (as in this case).
Here is what wikipedia has to tell about point source

And that is about it. A point source is (practically) point like, that means it is small enough for all practical purposes that are to be considered. This results in all radiation coming out of the same small region in space (localised source). [If you are away far enough all sources become "small enough".]
This has little to do with omnidirectional radiation.
A dipole can be as point-like as a monopole, and the same holds for a cardioid, a quadrupole and so on. A point source can have any radiation pattern and it can change with frequency.

But there is a connection though. A perfect omnidirectional pattern over a wide range of wavelengths can only be created with a point source (monopole), because if two sources (or more) radiate at the same frequency from different points there will be interference.

The actual use of the terms might be different, as the answer of CHatGPT shows, in some bubbles "private" definitions exist.

Yeah, I agree with this definitely over the definition of point source radiating sound equally in all the spherical direction. The way how I understand a point source speaker and how everyone else I know talks about it is that it's one single point of where the sound radiates from and radiating in all directions is not part of the common understanding.

 

[Arindal]

When standing up, the ‚eye‘ is better not directed up towards the listener‘s ears, clearly.

Not sure how tall you are, but in a usual home situation it is rather unlikely that the difference between being seated and standing is +20deg or more. And even if so, the dips are very very narrow and would not necessarily be leading to real colorations. Broadband colorations are much more audible and much more likely to get with conventional speaker designs, but in a completely different range.

Didn‘t I already tell, KEF solves the problem entirely?

Every easy solution comes at a price. In this case, it comes with very little midrange diaphragm area, limited excursion, overly broad dispersion in the midrange driver´s lower bands and very narrow dispersion of the tweeter due to inner waveguide/cone geometry. If constant directivity and higher dynamic midrange reserves are a design goal, I would not call this a solution without compromise, although from technical perspective I really appreciate how they mastered the interference issues.

I agree that coaxial is a very elegant solution to the lobing problem. But there are tens, if not hundreds, of different designs, and they all come with advantages and disadvantages. Would even refrain from any judging things solely by a label, as coherent, coaxial, pseudo-coaxial, virtually coaxial with array, cannot be clearly separated from each other and defy a precise categorization.

There is not THE answer to loudspeaker design questions, there are only ever many answers with many compromises.

I wholeheartedly agree, but would say there are better and less suitable solutions for a given application. And as SPL and listening distance for hi-fi is relatively limited, I can imagine a compromise which is getting pretty close to the imaginative ideal.

I think the most perfect approach at the moment for a radiation system that can be customized in every respect is the Beolab 90

Interestingly, I was thinking the same when I saw the technical details, the technology behind it and the options for beamsteering. When I had an extended listening session, sound quality in reality could not keep up, to say the least. Learned from that disappointment that tweeter arrays, particularly two-dimensional ones with domes and significant distance between the drivers, simply fail to deliver treble clarity, resolution and transparency.

When drivers' c2c's can be kept within 1/4 wavelength throughout their frequency summation ranges, they can radiate polars without ripple.

In the real world, some additional aspects factor in, like diaphragm geometry and resulting directivity, but as a rule of thumb this sounds very reasonable.

The main problem with these circular arrays following such strict rules is the tweeter. There is basically no way to have a tweeter array, so every midrange or midwoofer circle around it must somehow adopt to the directivity behavior of the single tweeter itself. But which one is the best?