Optimal distance between loudspeakers

September 22, 2017
 by Paul McGowan

Before we get started on today’s subject I wanted to offer an apology as well as an explanation about the hats. I love these hats and wear mine proudly. I’ve been approached three times now, asking just what an audiophile is and the conversation is always a pleasure. Those of you that have sent me pictures wearing your audiophile hats have warmed my heart. Thank you. It’s a level of awareness we want to support. That said, I want to apologize for messing up.

In my original post to you about the hats, I told you they were hand-crafted by Legacy Athletics of Hannover, Pa. American made. That’s not entirely true. The hats themselves, known in the industry as “shells”, come from one of three approved manufacturing plants around the world: China, the Philippines, or Taiwan. The “decorating” embroidering work, bill forming, inspection, and packaging happen in the United States. My bad. Please accept my apologies for the misinformation. If you want a refund or wish to cancel your order just email me. Again, I am truly sorry.

It’s an unfortunate truth that few of us have the freedom to place our speakers where they sound best. Instead, we pull them out from the front wall as much as our living situations afford and call it good.

We need to know three basic things: how far out into the room is best, how close together they should be, and how much toe in.

With respect to the first question, we’d like to use the rule of thirds. This simple formula places the speaker pair 1/3 the way into the room as measured from the front wall (the wall behind the loudspeakers). The listener is then placed 1/3 the way into the room as measured from the rear wall (behind the listener).

The second question is how far apart should the pair be? Here, we want to form a cross between an equilateral and an isosceles triangle (depending on the speakers: the left and right speakers at two vertices, the listener at the remaining vertex of either triangle type or combination of the two.

Toe in (pointing the speakers at the listener) is really dependent on the type of speaker you have and its off-axis response. My advice is always start with the speakers facing straight ahead and toe in to solidify the center image without sacrificing stage width and depth.

These are great starting points for system setup. Depth of soundstage is controlled by front to back movement of the speakers (away from the front wall increases depth). Tonal balance changes with distance between the left and right channels (closer together increases midbass coupling giving a fuller sound).

Want more information? Check out this video I put together for you.

Subscribe to Paul's Posts

31 comments on “Optimal distance between loudspeakers”

  1. There is one limitation, Paul, concerning the listener‘s distance: the undistorted sound pressure level of the speakers should correspond to the listening distance. Small monitor speakers thus often require near field listening. Listening from a distance too far requires higher volume settings being prone to heavy distortion.

    1. Agreed. I think that at least with 2-ways, there’s probably some sort of rule of thumb for maximum separation before center image loss, like one foot per woofer diameter inch.

      Also, particularly with larger speakers, minimum shortest room dimension. Even moderate-sized speakers are choked by a low ceiling, for example – apparently excepting line sources. Though I suppose it would be possible to sit closer to big speakers, it seems like a waste. Certain types and sizes of speaker need space to open up, as Steven mentions about Quads.

    1. Same feeling here. Video kills the radio/chatting stars?
      In the long run it might both flatten AND open the discussions on this forum, as new voices come in here via the videos. Wait and tea.

    2. Same here, I think it’s a matter of preference. I’m too impatient to watch the same thing as video which i can read. At the beginning I thought the videos cover topics different to those here.

      1. I try and keep them a bit different where the post is specific to the written word and the video is an expansion. I was worried that by offering two completely different subjects each day that it would be too much.

        1. I think it is worth a try the way you planned it. The only possible drawback – for me – would be if the top competent people in this forum resign because all too simple themes bore them.

      1. My few points of input are meant in a constructive way:

        Such videos are successful IMO if…

        The speaker is eloquent, sympathetic, has humor and great presence (bingo, that’s you)

        Content is of much interest to most (so and so, depending on one‘s experience they are a bit too basic)

        Video is much more efficient for explaining than text would be and not stretching things (so and so)

        Content is not redundant to faster ways to obtain it (nope that’s not really the case)

        The filmed scenes are lively in terms of camera view or movement in front of the camera (nope that’s not really the case)

        Video helps to explain things by sketches or showing objects (nope that’s not really the case)

        So my current impression is, the positive aspects of video are not used to full capacity. But I also know that this would take much more time, maybe an unreasonable amount.

        My suggestion would be to make much less videos with more draft, industry visits, interviews or fair visits etc.

  2. My advice would be to follow the manufacturer’s instructions. For example, Tannoy and Audio Note are designed to sit in corners or flat against a wall. Larsen (doing quite well in the UK) are specifically flat against the wall. Quad ESL’s have to be well away from the rear wall. Harbeth can go very close to the rear wall with at least 10 degrees of toe-in. The instructions provided Harbeth specify the seating position 1.5 times from the distance between the speakers. Grimm LS specify 45 degrees toe-in, 30 degrees if 45 cannot be accommodated, and there are equalisation settings in the software depending on which option is chosen. PMC recommend equilateral triangles focused on a point about 0.5m behind the listening position. I think B&W recommend no toe-in at all.
    I would love a pair of ESLs, but neither have the room or the wife to accommodate them. Often the speaker choice is determined by the positioning options and I would hope it is one of the main issues discussed with the dealer before a speaker purchase.

  3. No worries from this purchaser about the hat I am very pleased with mine. I did notice the “made in China” label and was going to question that but I’ve been so happy wearing the product I didn’t give it a second thought. Is there really anything anymore that does not have some component of it not made in China?

        1. This conversation is getting a bit off-topic, but I’m quite sure as the first one was made at Hotel La Villa, Calvi, and we went back the next year when he was about 4 months old and he got us an upgrade. I can’t remember the audio system in that hotel, it wasn’t the focus of my attention. On the second visit he had a bright red baseball cap, I’ve got a photo of it, it probably was made in China but didn’t have “Audiophile” stitched on it.

  4. Two things I’d like to add re speaker placement, which I’m sure have been mentioned before in similar posts. If thirds doesn’t suit the room, choose an odd number fraction to minimize peaks/ suckouts. Mine are 1/5 from the front wall due to room configuration.

    Secondly, distance to sidewall greatly affects the tone and bass boom. A simple speaking test, starting from the sidewall will tell you at what distance your voice begins to sound normal.

  5. Didn’t we have this speaker positioning post, and discussion, a while back? Deja vu all over again;-)

    As I said last time the rule of thirds does not work in most listening rooms that were not purpose built. My room is about 19′ long. My speakers require a minimum of approximately 10′ to properly blend the six drivers on the front. If I brought them out a full 6′, I would be sitting 3′ from the rear wall. I have a table in the way. And I would lose access to my desktop. Without getting up to measure them, it too early to be measuring them. I think the front center of the speaker is about 4′ from the front wall, and I have the listening couch 10′ from each speaker’s front center. The room is 12′ wide. The speakers are about 8′ apart, with a gentle toe in. Final position was determined by listening, and using an uncalibrated SPL app on my tablet. My Radio Shack SPL gauge took a dump a few years ago. I know the app is not ideal, but it was only used to measure the test tones on the Stereophile test CD, from 1000hz down to get the smoothest bass. Although the speakers are measured by manufacturer to be 16- 30k hz +/- 2db. In my room, and with the ports plugged 20 hz is down by almost 20db. Otherwise bass is smooth, with no humps or suckout. I think that may be due to the room having a doorway to the kitchen, and an open door with two of the traps blocking it, in the right corner. Giving me a larger area, than the room itself. I know Paul does not like most speakers as they are bass shy, when it is in the music, bass can be felt, as well as heard.
    I think unless you have a large purpose built listening room, your best bet is to use the manufacturer’s guide as a starting point. I seem to remember reading Audio Physics set up was unique in that they recommended the long wall with the speakers quite far apart, while sitting very close, like 5-6′ from what would be the line drawn between the two speakers. A set up I never tried. I should mention that I have DIY bass traps, behind each speaker and in one of the two corners on the back wall. There is a door in one corner. I also have DIY Room Lens that are just ahead of the outer front edge of each speaker, at roughly 45° to the side walls. This set up works for my speakers, in my room.
    In my former residence, I read online, as many different speaker placement articles, as I could find. I took what I learned as a starting point, but using the manufacturer’s recommendations as the primary starting point.

    As a side note, I had loaded a bunch of music into JRiver last night, and as I was writing this, I was thinking about the school math problems of two trains traveling at different speeds, etc, when Taste came on with a song about ” there is two trains a coming”. Music as the soundtrack of my life.

    1. Your 19:12 room ratio is close the the Golden Mean of 1.618, which helps with bass. You are correct that bass leakage is important for flat bass. Doors are more effective than traps, and the corner is the best place to leak. Large windows leak bass too, and wood beam floors to some extent. Do you have any large furniture pieces, like tall cabinets, that add bass diffusion?

      Frequency response in manufacturer’s specs is in anechoic, quasi-anechoic or partially anechoic conditions. The marketing folks pick the position to measure the flattest (for example including floor bounce for low frequencies). There are NO commercially produced speakers that are flat at all angles, and yet the sound is projected somewhere between 45 degrees and 360 degrees depending on frequency and the non-flat angles bounce around your room and end up at your ears. Your walls, floor and ceiling also cause interference between the bouncing sound waves so the frequency response where you sit can be +-10dB even from perfect point or line source speaker.

      Your phone microphone could be down 20dB at 20Hz. Nothing below 100Hz adds to speech intelligibility, it is mostly environmental noise including wind noise so they filter it out.

  6. This set of observations is some of the data showing how far audiophiles are down the rabbit hole. Speaker location wrt walls and toe-in are indeed sensitive factors in the two channel FAKE stereo image.

    Real stereo works by triangulating the angles of arrival and timing of discrete boundary reflections. Speakers that have polar radiation patterns corresponding to acoustic sources are instantly locatable, and all sound will appear to come from the points where the speakers are in the listening room. This is what happens with my speakers – no phantom center.

    Fake stereo works by making speakers locationally vague, with uneven, incoherent polar radiation patterns that can’t be produced by striking, plucking, blowing, rubbing or scratching a physical object. Further, there are two copies of each sound allocated in different proportions to two speaker locations by “panorama” controls to increase the ambiguity. You hear sound but can’t tell where it is coming from – so your brain sets hard at work to grow, wire and program neurons so you can make sense of the sound. After thousands of hours a combination of learned cognition and imagination leads to “pinpoint imaging” and “spectacular soundstage, with realistic depth, width and height”.

    BUT, nobody has called the police to report that a group of musicians broke into their living room and are playing there. This “imaging” is compartmentalized into the fantasy room in your memory palace, an auditory “Holodeck”. Even professional musicians have one part of their brain allocated to music and another for reproduction.

    I went through this training process too, and a handful of times heard a solid phantom center from systems built by Ralph Glasgal, Brian Bilgore, Barry Berkowitz et al. Since I stopped listening to fake stereo, the ability to fool myself has faded and the artifacts of multi-miking are annoyingly apparent.

    The speaker response anomalies including cone modes, baffle diffraction, voice coil inductance, driver separation and mutual diffraction, the inherent spatial pattern of finite piston diameters combined with comb filtering from boundary reflections produce fake spatial clues that our brains latch onto to make a “hash table” into our true 3D sense of sound. What is worse, in absence of ear training to real (live, not reproduced or amplified) music, we have no neural circuits to interpret acoustic performances!

    This also explains why people are happy listening to inappropriate concert halls, like chamber music in Carnegie’s Stern Auditorium. If they always sit in the back, they never experience articulation or aural imaging.

    Speaker location and toe-in adjusts the sum of anomalies arriving at our ears until it conforms more or less to our ear training. The goal is to mask the acoustics of the listening room and transport us to a fantasy dream world where one dimensional objects like audio tracks have weight, size and substance. Toe-in adjustment works because our brains are warped from listening to two channel audio. Solid hours per day training to live music with no PA system makes toe-in adjustment sound like slightly different flavors of spatial distortion.

  7. Since we are not long from entering what I and some others call “the silly season” between around Thanksgiving and New Year’s day, here’s what two advanced theories that reach the same conclusion have to say about it. These theories are WFS and my own AEFTT arrive at. This is for accurate reproduction of live sound. Both in their ultimate conclusion require a dome of speakers virtually shoulder to shoulder numbering in the hundreds if not the thousands. WFS is based on the notion that any acoustic field can be described and duplicated by an infinite number of spherical waves while AEFTT is based on vector analysis. Each speaker requires its own amplifier and signal based on a mathematical algorithm for a particular set of sound sources and the acoustic space to be duplicated. Those algorithms are based on actual measurements of the space to be duplicated. No such machine has ever been built as far as I know due to cost and complexity. Both ideas have been tried on a severely compromised level but the compromises have been different.

    http://empac.rpi.edu/events/2017/spring/introduction-wave-field-synthesis

    http://en.wikiaudio.org/Wave_field_synthesis

    The compromise prototypes of WFS makes is almost invariably that it is restricted to the horizontal plane based on what is in my opinion the false assumption that the vertical components don’t matter much. It also has never been adapted for playing commercial recordings. That is why the system set up at Bell Labs needed to use their huge anechoic chamber for live performers. There were some speakers higher on the wall and a few on the ceiling in the listening room but not nearly enough.

    The listening space ideally should be anechoic as well to prevent unwanted reflections and the speakers themselves covered with sound absorbing material. There should be no reflections whatsoever in the listening space for this concept to work as a scientific instrument.

    The further away you get from this model, the more the results diverge from what it would yield if it were correctly implemented. In this regard, the level of distortion of any sound system you can buy is virtually 100%.

    1. SM & Acuvox, you are both, if not actually prophets of doom, at least messengers bearing ill tidings. I do not disagree with what you say for an instant. Instead I have long accepted that I should be pathetically grateful if, except at low volumes, the sound seems to come from somewhere other than the speakers. If I cannot tell that the third violin has positioned his chair two inches out of line then I am just going to be brave about it 🙂

      1. I can only speak for myself. Two entirely different analytic approaches to the same problem that come to the same conclusion from different directions. The complexity of building a machine as well as the cost to perform this function is staggering. Only its potential can be suggested by compromised versions. Note that these ideas while adapted for use in one case as a demonstration in a very large room for a large audience and in the other for playing commercially made recordings are not nearly the full potential of the concept which is really only useful as a scientific instrument for study of areas like acoustics and psychoacoustics. That either compromised version work as well as they do given their shortcomings based on the theory is rather remarkable and for me was totally unexpected. Even this much is no easy task.

        What my posting was meant to explain is the technical requirements and difficulty of the task of accurate sound reproduction. If there is bad news it is that refining an idea that does not address some of the most critical aspects of these technical requirements of the theory will inevitably have severe limitations no matter how extreme the refinement of the equipment itself or the installation is. An analogy is the difference between building the brightest sharpest black and white television possible when what you are looking for is a color television. It is the concept, not the execution that is the limiting factor here. This is not to say that you can’t to some degree create what seems like the illusion in small part under very special circumstances but the result will still fall far short of what is often claimed by the absurd hyperbole and impossibly high prices for the equipment.

      2. I am no prophet. I am a revisionist, insisting that the last 85 years of audiological research and the audio engineering based on that model of human hearing come from bad scientific observations. I had to go back before 1930 and seek out odd corners of the audio and audiological research to find confirmation of what I knew to be true: that we are measuring audio to faulty parameters, and that people who learn to hear music through audio systems can’t hear what is wrong with them.

        It is only in the last few years that it has been shown that audio listeners have stunted brains, and that comprises some 99+% of the population that can afford season tickets or systems based on better models of hearing.

        The upside is that my model of hearing explains every audio paradox and anomaly regarding the dichotomy between measurements and listening tests. The debate is settled – both sides are wrong. The only people that can tell realistic reproduction are ones who hear live acoustic music every day. According to them, my speakers are the only ones that sound like music so I suppose I am biased.. but that was my criterion for development – satisfying conservatory graduate musicians, not any textbooks about what we can and cannot hear.

    2. There is good literature support for one degree or better angular resolution in human hearing. If we look at a hemisphere with a tweeter every one degree, it calcs out to about 5,000 channels. Then there is the ray tracing. In optics you need to follow five or six bounces to get accurate light shading, assuming this is about the same for acoustics the computation required blows up. Further, to shape a vector so it appears to arrive from the right angle requires multi-megaflop phase coherent processing of adjacent tweeters to arrive at the correct angle at only one position listening position – it can’t be right for even two people sitting together…

      If I understand your system correctly, it is like a vector statistical reverb. The average angle, attenuation and time delay correspond to the statistics of a physical space. In that way you can roughly map a position somewhat beyond the critical distance, but can’t replicate the sound where I sit, between the 1st and 7th row depending on the hall and subscription.

      My system can’t produce the sound further back, which is what typical stereo pair recordings capture – 10th Row, high up in the air. I suspect your system is better for contemporary recordings made from hundreds of pieces of 96 track digital with enough processing to heat my house in January.

  8. A trick I picked up from Dan Meinwald is to set them at an angle, rather than in parallel, to the boundaries of the room. It works in rooms in which setting them up parallel to a wall is problematic.

  9. Nearfield is like aligning a military target using two outside points that with our GPS (ears) that must be zeroed in to where the music begins to sound like real music. Off by a small degree can make the clapping audience sound empty, yet getting it zeroed in puts “meat on the bones.” Its an intuition and skill that is not for the faint hearted who want a simple set it and forget it formula to follow. And, there are so many other things involved (cables, fuses, AC quality) in this that it gets scary to try to explain in one fell swoop. I think the best way is to give a live hands on demonstration. For the nay sayers need to be silenced by a demonstration that will silence their frightened cynicism.

  10. Paul, I understand you are attempting to offer broad advice but I believe starting off with the “rule of 3rds” is a mistake.

    The first step should be to consider the type of speaker in question. Placement will vary depending on radiation pattern, bass extension, etc. for different models. Dipoles, such as most planars, are best at least 5′ out from the front wall to differentiate reflected sounds from direct sounds, regardless of the percentage of the length of the room. Smaller speakers, such as monitors, often will have fuller bass when closer than the one-third distance. Some models were designed for at wall, near wall, or even corner placement. All this means — it depends.

    So doesn’t it make sense to begin with the specific speaker rather than suggesting a blanket distance into the room?

    I agree with Jeff, this seems to be a repeat post. So I had to repeat my criticism as well. ;^)

    1. Well, I have probably had several posts on this subject and likely over time there will be several more. Maybe my experience with speakers is quite limited but to date I haven’t found a system where the starting point of the rule of thirds wasn’t pretty close to where they wound up.

      In some cases, I had to move them closer to the front wall to improve bass, but in general, if imaging was to be taken into consideration, they were/are always away from the front wall enough so they can breathe.

  11. Regarding your caps, Paul, it’s just a wild guess, but I suppose your electronics use caps from foreign countries, too. Not a problem with either cap. It’s hard to find clothing that is made in the USA these days. That’s another good reason not to start another World War. Cut off the supply of clothing and we’d have to listen to music naked. 😎

Leave a Reply

Stop by for a tour:
Mon-Fri, 8:30am-5pm MST

4865 Sterling Dr.
Boulder, CO 80301
1-800-PSAUDIO

Join the hi-fi family

Stop by for a tour:
4865 Sterling Dr.
Boulder, CO 80301

Join the hi-fi family

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