Line sources

August 22, 2018
 by Paul McGowan

Kevin in Wappinger Falls New York (wherever that is) asked me an interesting question recently.

“Arnie Nudell popularized the line source loudspeaker design and many of his most revered speakers such as the Quantum Line Source, the IRS 1B, the IRS V, and the Genesis 1 (and apparently at least the two larger members of your forthcoming AN series loudspeakers) are all line source designs. Why don’t we see more line source loudspeaker designs? Is it strictly due to the cost because of the many more drivers, the larger cabinet, the increased manufacturing labor cost – or are there other attributes that have made them less popular?”

This is a really good question and one we don’t talk about much. Perhaps it’s a good idea to first get on the same page. The classic two-way or three-way loudspeaker has two or three drivers in a box. The multiplicity of drivers—tweeter, midrange, woofer—is needed to break apart the frequencies so each driver only has to handle a specific range: tweeters handle the higher frequencies about 2kHz and the woofer handles everything from that point down.

A line source handles the frequency divisions in exactly the same way—tweeters, midrange, and woofers—but instead of relying upon a single driver for each range multiple drivers are instead employed, typically with a line of many tweeters and sometimes many midranges. The advantages of multiple drivers in a line are manifold: each driver has fewer demands and the waveform comes out in a long, vertical, cylinder rather than a single driver’s ever-expanding circular wave.

The advantages of a line source vs. a point source can be summed up fairly easily. A point source sound radiates in all directions from the driver and quickly loses energy as it floods the room in a 180˚ plane. Worse, this expanded radiation pattern hits the ceiling, walls, and floor and reflects back into the room out of synch (time) with the initial launch. Only those listeners in a narrow sweet spot get to enjoy the best sound. A line source radiates a more focused pattern in the shape of a tall vertical cylinder that, above about 500Hz, has nearly no floor, ceiling, or sidewall reflections to dissipate energy and add to sonic confusion.

All that said, tomorrow we’ll look at some older Infinity designs.

Subscribe to Paul's Posts

42 comments on “Line sources”

  1. I concluded from your posts so far, Paul, that the major advantages of IRS V and it’s successor were the dipole character and the number of drivers both offering better room match and higher sensitivity respectively. There is no real line source speaker as there is no real point source. Both are idealistic assumptions. All existing “line source” speakers show heavy and most complex unpredictable comb filtering. And I guess the process of selecting identical drivers from a batch delivered is most time consuming and costly.

    1. Well it’s true there’s no perfect point source and the same can be said for a line source. You’re right. However, the IRS is closer to a true line source than any 2-way speaker or even a coaxial could be called a point source.

      There’s no comb filtering I am aware of in the IRS design though I understand it can be a problem in some quasi line source designs. I’d venture to say the IRS at 7.5 feet tall is about as close to a true line source as you’re likely to get.

    2. Modern robotic assembly generally means if you buy a pallets of drivers from vertically integrated European, US and Taiwanese driver manufacturers you get good matching over the batch.

      The comb filtering comes from the gaps between the drivers, which generally need to be spaced less than a quarter wavelength center-to-center to make a good approximation of a line source. Like the horizontal pattern of frequency response anomalies you get the worst perturbations around the crossover point. Further, line sources are subject to end lobing unless the line is extended to within 1/8 wavelength of the floor and ceiling boundaries so the reflected image meets the same criterion.

      You will still get comb filtering from the side wall reflections. Since the line source projects much further distance in geometrical focus this effect holds for more reflections and is stronger, with higher peaks and lower nulls.

      Further, since there are no line sources in the orchestra or Nature (besides waves hitting a straight shore at a 90 degree angle), your ears will have no prior experience sorting out the spatial geometry of line sources in the listening room. Far more common are point sources like songbirds and insects, and point source horns – human, canine and feline voices, Trumpet, Trombone, Tuba, Horn, etc.

      This is why big point source, full range horns are the best for amplifying singers – they approximate the geometric projection of singers who can fill 3,000 seats with sound from their organic pipes (roughly cardioid from 300Hz-8KHz). If you hang out with horn loving audiophiles, they are inclined towards opera or more full bodied pop singers to demo systems. I favor Danley’s Unity and the monster Klangfilm cinema speakers, which maintain concentricity. Horn concertos, orchestral brasswind ensembles and big bands are also good applications for the room filling geometry of horn based speakers.

      The majority of acoustic musical instruments, however, are membraneophones. They project from a flexible surface like a skin (drums, Erhu, Kora, Banjo) or sounding board: Tars, Sitars, Guitars, Lutes, Harps, Dulcimers, Psalteries, thumb harps like MBira, Kalimba and Saz, Harpsichord, Clavichord, Fortepiano, Pianoforte, fiddles, Viols, Violins, Musette and Hurdy-Gurdy, Cajon, and so forth.

      This is fascinating because there are no natural planar membraneophones. This spatial geometry of sound has come to be the epitome of music, especially accompanying human voice, and argues for electrostatic and planar speakers – except membraneophones are driven anisotropically by point impacts or bridge and braces.

      This led to my consort of Violin family speakers. They mimic the membrane shape, bridge and brace structure of the Cremonese design sound boards to project the same wave geometry. The next step was speakers for sound boards with closed back, which are generally plucked or struck instead of bowed (see list above). Reproducing the spatial signature eliminated the final difference between the sound of speakers and the sound of music.

      There are more complex spatial patterns. Distributed resonators include organs, marimbas, xylophones, etc. These re best represented by banks of speakers, like Cameron Carpenter’s “virtual pipe organ”.

      The most difficult are wind instruments that use tone holes to determine the pitch, like flutes, oboes, bassoons, clarinets, saxophones, etc. These have a different shape for every note! They require a microphone and speaker array to produce a matching projection geometry as the holes are opened and closed.

      1. I guess even one of the simplest instruments, a triangle, produces already a most highly complex wave pattern, isn’t it? How difficult is it to catch the best sound pattern of this moving instrument by a fixed microphone? 🙂

        1. Fortunately, I am going for psycho-acoustic equivalence, not mathematical. From the audience, a triangle is close enough approximation to a point source and the midrange driver contributes very little energy to that projection.

  2. I’m not sure about the rather brief theory that conventional box speakers are fundamentally flawed unless you are held in the sweet spot with a vice. If so, I might as well chuck my collection of Wilson, Focal and Harbeth speakers in the trash and would recommend everyone to do likewise.
    So far as I am aware, the only manufacturer of line source speakers in Europe are the Swiss company Piega, a very successful manufacturer indeed. I read they have sold over 100,000 pairs of speakers. Most of their products are conventional speakers, I’ve heard a pair at the UK dealer and they were very good. He does not stock the line stage speakers, as they cost about $200,000 and a baby version is about $70,000. A hop over to Zurich would be required for a demo.
    I note with interest that they got a red hot architect/product designer to design the cheaper ones and if you are going to have 6 foot speakers in your living room they might as well look like works of modern art. Heaven knows what the cases cost to make, but they’re Swiss so who cares? The problem of matching drivers seems a non-issue as they are all made in-house and when the Swiss make something, it’s perfect. Always. And on time. Always.

    1. Paul,

      My loudspeakers are tall, vertical, dipole line sources down to 40 Hz. I have one comment regarding your description of the basic dispersion pattern: in the higher than bass frequencies the dispersion pattern is much wider than from point source loudspeakers and the interaction with side walls is quite strong. But I have found it easy to control the sidewall bounce with furniture and other acoustic obstacles. In the case of my dipole woofers, bass energy towards the sides is strongly suppressed due to dipole cancellation.

        1. Some of that sidewall interaction is reduced in Infinity and most other designs above certain frequencies NOT because they are dipoles, but because for a given driver size, as frequencies rise the radiation pattern of most drivers narrows, thus reducing higher frequency energy to the sides. In vertical line drivers there is still significant lower treble energy to the sides.

          In one of the experiments that I witnessed, a flat panel of tweeters was operated well above human hearing range. At that frequency it produced a radiation pattern that was so narrow (nothing to any side) that the generated acoustic beam could be aimed at an individual. The high frequency was then modulated to a lower frequency so that the sound could then be heard, but only by the individual toward whom the beam was aimed.

  3. From what I’ve been reading the last 40 years or so about audio, this post makes no sense to me at all.
    If this post is right, most European speakers would be trash. And they’re not.
    On the contrary, many are among the best in the world.
    With excelent highs, mids and real bass in real life rooms.
    Not the amount of bass a bass junkie (like for instance Paul McG.) wants of course.
    Paul McG. who once stated B&W have no (real) bass. Of course they have real bass.
    (and make no mistake : I’m not a B&W advocate and I don’t have them myself !)
    Radiating in all directions, isn’t that what we want ?
    Hitting ceiling, walls, floor, isnt’that what we want ?. The sound of real instruments reflects from the walls as well.
    Without it, the sound would be “dead”.
    But maybe I’ve been listening to the wrong speakers with the wrong sound all those years.
    Time for another hobby ?
    BTW., I don’t know into which category Infinity speakers fell, but they were never very popular in Europe.
    Reading these posts I think much more in USA.
    A cultural thing I guess. Americans, Europeans, Japanese…They all have completely different expectations and taste regarding sound(quality).

    1. Actually jb4, I think that loudspeakers work best in real rooms because they are designed to work in real rooms. In a simple experiment I once tackled the problems of room bounce by taking my ordinary speakers outdoors away from most acoustic boundaries (even better than an anechoic chamber). This of course required huge amounts of frequency compensation, but I was able to attain reasonably flat reponse to 80 Hz.

      The stereo effect was unreal, literally. Virtual sources were solidly in place laterally and in depth. Depth perception was continuous, not layered. Micro dynamics seemed improved as well. Compared with indoors the stereo effect seemed exaggerated; many times sources seemed too large; at other times they were just right. Outdoors, any out of phase signals seemed infinitely in extent, which is what I think contributed to the stereo exaggeration. Then too, my speakers were consumer grade; better speakers may have produced better results.

      1. Yes, alanpayne, I get it.
        I did the same “experiment” (speakers in the garden on a sunny day) several times.
        But I don’t see what it has to do with the content of today’s post.

        1. Oh, I’m sorry, jb4. The point I was trying to make is related to your comment about the need for reflections from room boundaries. In my outdoor experiment I was able to hear more of the rooms in which recordings are made. Inside, this information is obscured by listening room contributions. But since most people listen in rooms, speakers are designed to operate inside. Paul has even made the point that he considers the room necessary for voicing his speakers (my take on what he said).

          1. No need to say sorry, alanpayne.
            But thanks for this explanation, now I understand
            better what you meant.
            But in my comment I was trying to express the fact that I’m not happy that in today’s post PaulMcG. gives the impression that in speakerland there’s nothing better than a line source. Point source is inferior.
            I know there are companies that will tell you the exact opposite.
            There’s more than one way to skin a cat.
            And PaulMcG. “corrected” that in his 6:22 comment.
            But I’m still skeptical when I read
            “The line source gives you more focused energy from the initial wave launch before reflections than the conventional patterns do”
            Well, that’s true of course, but for me the question remains : is it better..?
            The line source speaker may even give more “focused energy” than the real instrument(s).
            Could there be too much energy firing at your ears, compared to the real thing…?
            So, I have my doubts.

    2. If you look at the basic problem and challenge of authentically reproducing recorded music all speakers and recording techniques are “trash” or at least far from faultless. The crucial point is: which degree of deficiency and distortion are you willing to accept? Even sound systems based on WFS (wave field synthesis) require some hundreds elementary-source loudspeakers and a recording technique only catching the instrument sound and not the room sound in order to create a near authentic holographic sound including the desired room sound (reverberation). And those elementary-wave drivers a still far from being an ideal point source. My approach: keep it simple! Otherwise the complexity will create more problems than just being apparently solved by increasing the complexity.

      1. I have a question: why does WFS recording technique try to eliminate room sound in the recording; is it because a new synthesis of the room will be generated? THX

    3. I can only agree. I had to look up line stage speakers as in 35 years I’ve never seen or heard a pair. Without them, the European audio industry, which from what I’ve gleaned frankly knocks the spots off the USA industry (PSA products are the only USA items I’ve considered or bought, but with their UK pricing and lack of servicing/offers there are obvious better options), has done very well without them. The only USA speakers I’ve heard recently were the top of the line $100k Magico and I thought they were dreadful, unless you like your sound completely subsumed by bass.
      Design good drivers and you really only need one of each per channel. This has proven to be the successful approach of many of the leading European speaker manufacturers, who without exception (as far as I know) have their own proprietary drivers. You can go back to Tannoy dual concentric, Quad’s ELS, PMC’s 15″ bass unit, Harbeth’s 8″ Radial and various drivers from Focal and Dynaudio. These have all been the backbone of a successful speaker systems for at least 30 years in all cases, all of them in both consumer and professional audio. If line stage speakers suddenly knock all these off their perch, then so be it, I’ll be obliged to get a pair.

      1. I have auditioned every single brand you mention, and none of them are able to deliver realistic dynamics and transients without Doppler inter-modulation, slewing and Xmax limits. This is the total advantage of line source – enough surface area and swept volume to produce the sound without adding false frequencies, the most annoying form of distortion.

        Audiophiles are so used to the Doppler distortion and diffraction lobing of their speakers (including baffle step), they miss it when it goes away, or listen at such polite levels they are off the loudness curve.

        Further, most recordings are compressed in various ways so the inadequate 1″ dome tweeters and 8″ woofers of consumer systems don’t clip. My tweeters are equivalent to forty 1″ domes in surface area, and I use 8″ drivers as midrange.

    4. I would say this is a bit of an extreme reaction to what I was trying to say in the post, though I suppose also I could be accused of being too extreme in my simplistic explanation. So, we’re even. 🙂

      In a perfect world you want the initial wave launch to hit you first and without delayed reflections. This happens in both speaker types. The difference between the ever-widening radiation pattern of a conventional speaker and the cylindrical launch of a line source is one of focused energy. The line source gives you more focused energy from the initial wave launch before reflections than the conventional patterns do.

      That’s not to say conventional speakers are trash. They are not and I never wanted that to be what got taken from my words. There are many fine speakers – world class speakers – that are not line sources.

      1. Presumably a cylindrical launch will be radiating in the sale pattern (assuming the same drivers), so energy will be dispersing broadly the same way, only more of it. You might as well just stack two speakers one on top of another, which is what PMC did with their XBD speakers, but they still only have one mid and treble driver.

        For once I largely agree with AcuVox. Sound doesn’t come in lines, or from stereo point sources for that matter. The AvantGarde Duo horns do have a wonderful midrange and an astonishingly full yet relaxed sound, using active bass and DSP these days. All you need is lots of space an money. A chap near me sells the Cessaro horns, probably not the Gamma II as they cost $400k, but apparently sound very nice.

        Maybe if you want to build a line source speaker, it should be a bit more substantial, providing a wall of sound like this one:
        This was a one off and cost over $1million. Even their normal products are too expensive to merit a UK dealer.

        1. I would not exactly call this agreement. PMC and Avantgarde drive me from the room! Exponential horns have more uneven spatial patterns than point source 3 ways. I never liked Klipschorns, Voice of Theater or studio monitors like Westlakes, URIEs, JBLs (except for the direct radiator 4311’s). They typically have narrow band resonances from the mouth reflection, which led to my old aphorism, “They call them horns because they HONK!”

          Computer modeling is solving these problems with more complex profiles like semi-elliptical oblate spheroid and Tractrix, as well as the patents of the aforementioned Tom Danley.

  4. Paul,

    My loudspeakers are tall, vertical, dipole line sources down to 40 Hz. I have one comment regarding your description of the basic dispersion pattern: in the higher than bass frequencies the dispersion pattern is much wider than from point source loudspeakers and the interaction with side walls is quite strong. But I have found it easy to control the sidewall bounce with furniture and other acoustic obstacles. In the case of my dipole woofers, bass energy towards the sides is strongly suppressed due to dipole cancellation.

    I apologize for my displaced comment; I was unable to edit or cancel it.

    1. If you can think of a point source sound radiation pattern as half of a tennis ball, then the line source sound radiation would be half of the tennis ball can cut vertically. Hope that image helps.

  5. Line source speakers have been around for a long time and before Arnie Nudel’s IRS. One line source Hi Fi speaker that predates IRS by a few decades is Bozak Concert Grand.

    Line array speakers have been used in theaters and PA systems for donkey’s years.

    There is much to be said for line arrays. One advantage is load sharing. This means drivers share the load so no one driver has to output all of the sound. As the power for each driver is related to the square of the current and the heat generated dynamic compression is also related to the square of the current so multiple drivers at the same frequency substantially reduce this effect. They also keep each driver in the most linear portion of the magnetic field so distortion is also automatically reduced too. While 2 or more drivers operating at the same frequency will cause comb filtering the number of resoanant peaks and dips for N drivers = (N-1)!. what means is if you have two tweeters you get one set of comb filter frequencies. 5 sets and you get 4x3x2x1 sets. With the huge number of drivers in the Infinity IRS the number is so great that for all practical intents and purposes they don’t exist.

    For a point source in an anechoic environment (no reflections) the sound level of a single driver falls off with the square of the distance. If you double your distance it will drop by 1/4 or 6 db. For an infinitely long line array it will drop proportionately with distance so if you double the distance it falls off by 1/2 or 3 db. For an infinitely large surface array it falls of at 0 db, that means it will sound just as loud no matter how far away you get from it. Reflections mitigate this drop. Therefore in a highly reflective room where most of what you hear is reflections the sound level will remain more or less the same even for a point source. A concert hall is one such environment.

    In my bucket list among things to do before I die is to build a pair of large surface array dipole speakers just for the hell of it. Using dynamic drivers it will not be constrained in the way panel speakers are. The question it will answer is whether or not the special qualities of electrostatic and planar magnetic speakers are due to the fact of their different method of moving air or because they are surface arrays.

    Everyone seems to have their own idea of what “good sound” is and what the perfect loudspeaker and sound system is. I think for many people it is intuitive. Mine is based on my own unique physics and mathematical models of sound fields that reach your ears. So that in my optimal designs, that is the factor that drives the design, to understand and duplicate those fields produced by acoustic musical instruments. The design in those I’ve built not only took room acoustics into account, they incorporated the reflections as a necessary part of the design. It’s not easy. The most difficult design is one where you have to reconstruct the reflections of a very large room in a small room.

    Here’s one way to evaluate your own sound system or anyone else’s that has a pretense of high fidelity. Close your eyes and listen to the sound of recordings. If you had gone to a live concert and paid a lot of money for a ticket, say $100 or $200, and that was the sound you’d heard, would you feel you’d gotten your money’s worth?

    1. Your last paragraph was a great thesis of psycho-acoustics. I will be quoting you!

      The rest is statements of salient facts, except for the factor you are not hearing – the ability of one ear to determine the angles of multiple reflections. I think this may come from the difference in our live listening habits. I seek smaller, less reverberant rooms, smaller ensembles and seats inside the critical distance. We rarely go to Stern or Geffen, sit on the outside near the front in the Starr Theater and have been to the Met half a dozen times in 18 years.

      We replaced “Mostly Mozart” Summer season series with Caramoor, which is brilliant musical acoustics design. The stage is five sides of non-parallel reflectors that enable the performers to hear each other and project into the tent, where the deep catenaries provide broad diffusion. The sound level falls off, but articulation maintains to the back.

      Alan Kozinn, my favorite reviewer, always sat in the back of the hall, the “hoi polloi” seats. Your description of your system and references resemble that sound, 80-90% reverb. I prefer <50%, what everybody else calls "too dry".

  6. As has been alluded to elsewhere, there is nothing inherent in a line source that would result in a different horizontal dispersion pattern than a more conventional design. What is true is that the power response of a line source will come closer to matching the on axis response. This is because, unlike the horizontal response, the vertical response will be far more consistent. As a result, side wall reflections will be less problematical as the power response of the reflections will be smoother, but there will be no change in the effect of delayed reflected signals.

  7. Let me get this straight. Some dipole designs are point source and some are line source. Yet line source dipole will still have a different radiation pattern than point source dipoles ? I know of some dipole designs that are not line source which have minimal side wall reflection.
    Dazed and confused.

  8. Cost is a very big factor in making line sources. Also, room size matters. They dominate a small room and aesthetically just do not fit. They move too much air and easily overpower small rooms. They are more difficult to make then simple one, two or three way speakers. But given the right conditions are in a class by themselves. The sweet spot is much larger. The loudness remains the same whether one stands next to the speakers or much further away. The sound field is huge. In fact after listening to line sources going back to the regular speakers is quite a let down at least initially till such time that one gets used to the comparatively lilliputian sound of the regular speakers. Regards.

    1. “Getting used”, that’s a core aspect here! Those who got used to listen to horn speakers and their high resolution of fine details also never will accept the boxy sound of conventional loudspeakers with vibrating walls and internal room modes.

  9. Hi Paul,

    The village of Wappinger Falls, New York is along the east bank Hudson River in Southern NY. So named as it resides next to Wappinger Creek that drain Wappinger Lake into the Hudson. On NY Route 9, just south of Poughkeepsie and a little north of Newburgh. All this on the western edge of CT. Home of Pasquarella Recordings Studio. Though I have never been there it sounds like a lovely place.

    1. Yes, @ 80 miles north of NYC and just north of Rt 84 near Newburgh, NY. West Point is not far away (south and on the western bank of the Hudson). Quite beautiful area – I drive through it from time to time.

  10. Learned a lot by today’s posts. I’m after dynamics…large scale and especially also at lower volumes (which is one reason I use active speakers).

    From my previous experience with planar”s (in case that equals a line source, too in some regard), I didn’t like their missing dynamics, but their kind of radiation and impact of “wall of sound”.

    So while my speakers are just as high as the AN3 but with the volume of AN2, I’d buy everything bigger as line source. The smallest line source possible would be my favorite (for halfwise „normal folks“ compatibility reasons).

    Not to say the huge Van Schweikert d‘Appolitos haven’t been fascinating, too (as non line source) at the last HiFi show. So at least close to a Mill there are also some other nice speakers around 😉

  11. I believe another key to a line source is simply the total square inch area of transducer exciting the room. Think of the difference between one 10″ woofer and two 15″ subwoofers; the concept remains between one 1″ tweeter and a many square inch vertical column of tweeters (preferably ribbons).
    Can’t wait to hear your transducer, Paul!

  12. Being a ‘northeaster’ I know Wappinger falls. As far as all the comments posted so far I’m in the line source camp for speakers. They just sound Better to me. Not to say I haven’t heard horns or point source systems that sound excellent. I fell in love with Apogees and have both a set of the grands and the centaur majors. Sadly they have gone the way of the ghost. Based on my experience with the infinity sound (starting in mid 70’s) I can only imagine Paul’s speaker line will be stellar (no pun intended).
    I’m not sure I’ll Be ready to give up on what I have, but never say never… technology marches forward.

  13. Thanks to all for the interesting discussion today. I learned a lot. However as is the case with a good education, you end up with more questions than you started with.

    1. So did I. Also from here
      My impression is that the tend to be very big, hugely expensive, very heavy and solve problems that for many people are not problems. For example, most serious audio listeners have a chair in a sweet spot and good sound in that region is sufficient. Their ability to throw sound over a long distance is actually a negative in my, and probably other people’s domestic environments.

  14. All this talk about different speaker arrays for different voices and instruments is interesting, but to customize speakers for different voices and instruments is not practical for most of us. The diverse recorded music we listen to is typically 2-channel, recorded with a stereo microphone or set of microphones mixed down to a left channel and a right channel. The microphones are placed at points in space and, to my knowledge, do not detect the shape of the radiation pattern of the sound waves produced by each voice or instrument. Each microphone records only what it detects at its point in space. The final mix is effectively a single point left and single point right microphone. So, a single point left speaker and single point right speaker should reproduce very well what the mics recorded. A more complicated speaker array may result in a more satisfying, fuller, seemingly less distorted sound, but it is not a truer reproduction of what the microphones actually recorded. [As one poster commented, the only way a line array could give a more accurate sound would be if each driver were associated with a corresponding microphone in a line array of microphones.] To me, line array speakers are like vacuum tubes. They do something to the sound that some people may hear as a positive enhancement (such as bigger soundstage, thickening of the sound, liveliness and wow factor) while others may hear negative qualities (such as an exaggeration of scale, loss of some delicacy and micro-detail). Hopefully the IRS killer will have no negative qualities!

Leave a Reply

© 2023 PS Audio, Inc.

linkedin facebook pinterest youtube rss twitter instagram facebook-blank rss-blank linkedin-blank pinterest youtube twitter instagram