Proper Speaker Placement

amplifiers, receivers and loudspeakers
Post Reply
Klaus R.
senior member
senior member
Posts: 378
Joined: 23 Apr 2004 09:57
Location: Netherlands

Post by Klaus R. » 01 Mar 2010 17:22

The few things that do seem to be broadband is to avoid early reflections by placing speakers away from room boundaries.
One common approach/recommendation is to lower the level of early reflections by 10-15 db. However, absolute perception threshold of lateral reflections for music is between 18 and 25 db below direct sound:

Schubert (1966), “Detectability of single reflections for music” (in German), Technische Mitteilungen RFZ, vol. 10, no. 3, p.124

Regarding room treatment, contrary to studio monitoring setups, the area behind the speakers and on the side walls should be dispersive. The area behind the listener should be absorptive.
Toole recommends to not treat side wall reflections. Linkwitz says: “Sidewall reflections should be diffused if treated at all. Absorbing them is like turning down the tweeter. Absorbers are not broadband and ineffective below a few hundred Hz.. Besides, lateral reflections are important for sound scene recognition.”

Toole further recommends to place absorbers on the wall behind the speakers. He bases this advice on investigation by Kishinaga, where it was found that for some descriptive parameters absorption was better than leaving the wall reflective and for some parameters there was no difference between absorptive and reflective. The overall judgement of Kishinaga was that for professional use absorptive and for music enjoyment reflective walls were better.

Kishinaga et al., “On the room acoustic design of listening rooms”, Audio Engineering Society preprint 1524

As for the wall behind the listener, Toole says: “The rear-wall reflections is of little value spatially; it is probably innocuous, but consideration may be given to absorbing or scattering it.”


I had another look at the AES paper where Audio Physic’s method and psychoacoustic background is explained. The basic idea of the method is to have no reflections within 5 ms after the direct sound. The basis for their reasoning seems to be precedence effect experiments using clicks and a single speaker as direct sound source and a second speaker for the simulated reflection. Clicks are used because of low temporal overlap. Echo threshold (the fused image perceptually splits into two images) for this setup is 5ms. In general, with increasing delay a broadening of the image is observed. If somebody could explain what a single speaker/single reflection setup using clicks has in common with 2-channel stereo speakers/right-and-left reflections using music, that’ll help me understand the approach.

Klaus

Klaus R.
senior member
senior member
Posts: 378
Joined: 23 Apr 2004 09:57
Location: Netherlands

Post by Klaus R. » 01 Mar 2010 17:39

ld wrote:So, I think its mistaken to consider studio sound treatment principles as different to domestic listening, the best practice is there to learn from, and the result a target to aim for IMO.
The problem is that studio control room acoustics is a moving target: in the old days where only 78 rpm were made control rooms were not treated at all or maybe some absorbers were randomly placed. There were Hidley designs, LEDE (apparently seriously flawed), various RFZ (reflection-free-zone) aiming at 20dB/20ms, 15db/15ms, 15dB/10ms, Newell's non-environment room. Which one do you choose?

Voetmann (2007), “50 years of sound control room design”, Audio Eng. Soc. Preprint 7140

http://www.wirelesstech.dk/C1256ED60045 ... 126205.pdf


Klaus

Guest

Post by Guest » 01 Mar 2010 20:08

Klaus R. wrote:There were Hidley designs, LEDE (apparently seriously flawed), various RFZ (reflection-free-zone) aiming at 20dB/20ms, 15db/15ms, 15dB/10ms, Newell's non-environment room. Which one do you choose?
Learn and apply to domestic listening ! I'd take any of them over most real living rooms :wink: But it's very different if you can start with a clean sheet as few of us can.

I had the privilige of hearing control rooms and listening rooms over the years. It just sets the standard. Yet seems so simple as to be achievable domestically. And listening rooms particularly can sort of sound good before there's any programme, just the room, when one talks for example. And have furnishings ! So I'm naturally inclined to apply those principles and goals first, variously.

albernbowen
Posts: 1
Joined: 23 Feb 2010 06:52
Location: newyork

Post by albernbowen » 02 Mar 2010 07:33

I just recently bought an Onkyo 7.1 htib. It was no problem for me to setup the 5.1 speakers, but don't know how to go about placing the two surround speakers that go behind you (135-150 degrees). My problem is that a wall is directly behind my couch and there's really no other way to arrange my furniture. Would it sound alright to put the extra two surround speakers on the wall right behind me or should I just keep it 5.1? I currently have two rears at 90 degrees on either side of the couch.

Klaus R.
senior member
senior member
Posts: 378
Joined: 23 Apr 2004 09:57
Location: Netherlands

Post by Klaus R. » 04 Mar 2010 18:24

ld wrote:
Klaus R. wrote:There were Hidley designs, LEDE (apparently seriously flawed), various RFZ (reflection-free-zone) aiming at 20dB/20ms, 15db/15ms, 15dB/10ms, Newell's non-environment room. Which one do you choose?
Learn and apply to domestic listening !
I try to learn, that’s why I’m reading the relevant technical/scientific literature. And what I note is that all those approaches for control rooms acoustics are based on exactly ZERO psychoacoustic evidence! That is, there is evidence cited by those who propose and favor those approaches, but this evidence is using single speaker/single reflections setups, is, with one exception, using artificial signals and speech. It is known that the phantom source created by two loudspeakers has a timbre different from that of a real source in the same position:

Theile, “Localisation in the superposed sound field”, PHD thesis, Berlin 1980

Lee et al. (2004), “Elicitation and grading of subjective attributes of 2-channel phantom images”, Audio Eng. Soc. preprint 6142

It is known that the more reflections there are the less the sound is colored:

Salomons (1995), “Coloration and binaural decoloration of sound due to reflections”, Thesis, Delft University
http://repository.tudelft.nl/view/ir/uu ... d6cc04fbf/

Bilsen (1995), “Binaural modeling of spaciousness and coloration”, Music and Concert Hall Acoustics, Conference Proceedings from MCHA95, Kirishima International Concert Hall, Japan, May 1995, pps.327-335, Editors: Y. Ando & D. Noson
But it's very different if you can start with a clean sheet as few of us can.
I could, in 2002. I choose room dimensions according to Bonello, now I know that this was useless, I included an acoustic ceiling to lower reverberation time, that ceiling was extremely useful indeed. Being our living room absorbers were out of the question and now I know that they are actually not necessary, in view of the good radiation behaviour of my speakers.
I had the privilege of hearing control rooms and listening rooms over the years. It just sets the standard.
Funny you should say that:

Fazenda et al. (2001), „The views of recording studio control room users“, Proceedings of the Institute of Acoustics 2001, vol. 23, pt.8, p. 213

tells a different tale.


Klaus

Guest

Post by Guest » 05 Mar 2010 01:34

Hi Klaus. Even in the implausible event that control/listening rooms were devised without regard to acoustic/psychoacoustic design principles and practice for some perverse reason, they remain the environment in which the reference sound exists. So understanding and learning from the methods employed there seems the way to fly if it is the intention to reproduce the reference sound domestically. That is my simple point. So if you elect to 'deviate', for example not to absorb front boundary reflections, makes sense to have a reason, i feel.

Otherwise, given the rhetoric of so many publications, it seems near impossible to extract 'best practice' and make something useful happen.

Of course, if you're inclined to the opinion that a studio listening environment is generally poor, you might prefer something like Mr Linkwitz' living room ! To each his own.

Be interesting to hear more about the room you created from scratch, Klaus, and where the best sounding room you ever heard was?

TA

Post by TA » 05 Mar 2010 06:47

As I've posted elsewhere, damping walls behind speakers and a reflective environment works excellent and o clean up the sound very much. That cannot be a bad thing.

If the model is that the the speaker wall side should be a stage, then it should be acoustically invisible. That model works just excellent in all installations I've heard, studio or homes.

Then there are many other things, adjusting for the fact that stereo is two speakers in an angle, which tend to shift the heard frequency response as compared to a source that is located in the middle, adjusting for room gain, etc , but that is speaker design.

Klaus R.
senior member
senior member
Posts: 378
Joined: 23 Apr 2004 09:57
Location: Netherlands

Post by Klaus R. » 06 Mar 2010 14:19

Hi Lucky,
ld wrote:Even in the implausible event that control/listening rooms were devised without regard to acoustic/psychoacoustic design principles and practice for some perverse reason, they remain the environment in which the reference sound exists. So understanding and learning from the methods employed there seems the way to fly if it is the intention to reproduce the reference sound domestically. That is my simple point.
Control room designs are based on perception threshold experiments of the type single loudspeaker/single reflection. Since it is proved fact that more than a single reflection results in less coloration I have serious doubts that the observations made using the above setups can be extended to the 2 loudspeaker/multiple reflection setup.
So if you elect to 'deviate', for example not to absorb front boundary reflections, makes sense to have a reason, i feel.
Again, there is evidence that the floor reflection is individually affecting timbre when using noise (single loudspeaker/multiple reflections), but no longer affecting timbre when using speech. No data for music.
Otherwise, given the rhetoric of so many publications, it seems near impossible to extract 'best practice' and make something useful happen.
In his book Floyd Toole write: “In a recent survey of recording control rooms (see link below) revealed a disturbing amount of variation in spectral balance among them. The differences were not subtle, especially at low frequencies.”

http://www.aes.org/e-lib/browse.cfm?elib=10075

Further, here’s an excerpt of an AES paper on sound localisation:

Hartmann, "Auditory Localization in Rooms", PROCEEDINGS OF THE 12TH INTERNATIONAL AES CONFERENCE 1993, pps 35 - 39

There are implications of this localization research for stereophonic sound recording and reproduction. Because the steady-state localization cues in a room are discounted, a listener who is positioned away from the loudspeakers will not respond to steady-state cues for localization, but will localize on the basis of transients only. This is especially true if the room is rather reverberant. The perceived source may be at one of the two loudspeaker positions, or, if speakers are about equally distant, the perceived source may be between the speakers. It may even be where the mixdown engineer wanted it to be. On the other hand, our experience with the Franssen effect suggests that if the listener is close to the speakers, especially if the room is dry, then steady-state cues may play an increasing role. One expects, for instance, that the mixdown engineer listening to small loudspeakers on top of the mixing desk in a dry engineering room is getting some plausible steady-state cues that will lose their effectiveness in a large living room. Because of the plausibility hypothesis, the consumer gets a different mix of steady-state and transient cues than the engineer.

The idea that for a given room configuration there is a minimum onset rate, measured in units of pressure per unit time, required to trigger the precedence effect means that sonic images will seem to be better localized when a recording is reproduced at high level. Based upon the numbers obtained experimentally, one does not expect this to be a large effect for recorded sources that are “up front.” One expects this to he a factor for background strings which do not appear at high level on the recording and do not have abrupt onsets. Again, if the mixdown engineer is listening at high levels, which is how we normally imagine him or her, then the engineer is getting localization cues that will be lost in most reproduction; this lime the cues are weak transients.


So, for a start, control room acoustics is a moving target and second, the listening conditions at home are almost certainly not the same as in any control room, you will hence end up with a task impossible to fulfil: recreate control listening environment.
Of course, if you're inclined to the opinion that a studio listening environment is generally poor, you might prefer something like Mr Linkwitz' living room!
I don’t think that studio environment is poor. Control rooms are working environments so no one will use “natural absorbers” you find in living rooms such as heavy carpeting, absorbing furniture, drapes etc. Therefore, acoustic treatment is a must. However, I think that by no means control rooms set the standard for the acoustics to be recreated at home. Show me convincing evidence and I change my mind.
Be interesting to hear more about the room you created from scratch, Klaus, and where the best sounding room you ever heard was?
That room was built as an extension to our house. I thought now is the occasion to get the acoustics right. Bonello’s criterion for room dimension ratios was nicely in line with the concept of the Schroeder-frequency, so I went and made the calculations. Afterwards it turned out that the builder had made a mistake and built the room a bit larger than on the drawings, thereby messing up with Bonello’s first requirement. What shall I say, there was no room mode disaster, and then I came across these:

http://www.aes.org/e-lib/browse.cfm?elib=13686
http://www.aes.org/e-lib/browse.cfm?elib=15044

The walls were built from bricks, the floor was to be tiled, windows with double glaze, so I knew in advance that without any treatment that room would be catastrophic. I initially intended to use the compound baffle absorbers of Fraunhofer Institute of Constructional Physics but these were heavy (40 kgs each) and expensive. Then I discovered the fabric under tension ceiling. Metallic profiles are attached to the walls, the room is heated to about 45 degrees C, the synthetic fabric is clamped into the profiles, when cooling down the fabric shrinks and is tensioned. With some 26 cm air space above the result is a huge membrane absorber with high absorption coefficient well down into the bass range, lowering reverberation time to comfortable values and treating both vertical and horizontal modes. When playing sine tones the modes can still be clearly distinguished but so far I have found only 3-4 tracks among my records and Ceedees where a mode is actually excited.

Best sounding room I’ve heard? No idea, worst sounding is easier, friends of ours have a dining room with just table and chairs, some light curtains, wooden floor. Speech intelligibility is extremely poor, they are French (which is not my mother tongue) so when I’m at one end of the table and all of us are talking at the same time, I have real difficulties to understand what the person at the other end of the table is saying: cocktail party effect is definitely struggling.

Klaus

TA

Post by TA » 07 Mar 2010 10:33

For non-published experiments, there has been one with using speakers and orchestra behing a movable wall, playing music either directly or via speakers.

In my own experience damping of the listeners "acoustic environment" leads to very unpleasant experience. Diffusion works better. While the speaker side is very much cleared up using proper damping.

Sidewalls should be reflected more than damped, but also special-made double-sided "angled" ones can be used (one side with absorption, one with reflection).

Since it is not possible to reproduce the recording venue in detail, one always have to deal with comprimises. In my mind, damping behing the speakers, heavy rugs on floor, diffusion in the rest of the room comes a far way to a higher listening experience, regardless of the audio system used.

In fine-details one could look at the speaker response in terms of reproducing a "Phantom image" between the speakers. The errors generated by the stereo system can be somewhat compensated for by having a different target response curve than "linear", leading to a more realistic presentation of the live image.

T

missan
senior member
senior member
Posts: 902
Joined: 26 Apr 2008 15:19
Location: sweden

Post by missan » 07 Mar 2010 10:52

wrote: In my mind, damping behing the speakers, heavy rugs on floor, diffusion in the rest of the room comes a far way to a higher listening experience, regardless of the audio system used
T
Agree, this is in my mind a good listening environment.
missan

Klaus R.
senior member
senior member
Posts: 378
Joined: 23 Apr 2004 09:57
Location: Netherlands

Post by Klaus R. » 06 Jul 2010 09:23

Klaus R. wrote:I had another look at the AES paper where Audio Physic’s method and psychoacoustic background is explained. The basic idea of the method is to have no reflections within 5 ms after the direct sound. The basis for their reasoning seems to be precedence effect experiments using clicks and a single speaker as direct sound source and a second speaker for the simulated reflection.

Since the Audio Physic method is currently being discussed in a thread on a German forum, I had yet another look at this AES paper (http://www.aes.org/e-lib/browse.cfm?elib=7529).

The main line of reasoning is that within the first 1-5 ms after the direct sound the reflections cause image spread and image shift. After that 5 ms window the precedence effect chimes in and the reflections are no longer disturbing.

Reference is made to Blauert "Spatial hearing", 1974 edition, chapter "Law of the first wave front". I checked this chapter in both the 1974 and 1983 edition and could not find anything about this 5 ms time window. I mailed Audio Physic asking where this 5 ms figure comes from, no reply yet. Anyone any idea about this 5 ms, any relevant literature?

Klaus

JollyJeweller
long player
long player
Posts: 2649
Joined: 02 Feb 2007 11:58
Location: At home !

Post by JollyJeweller » 13 Aug 2010 09:10

As some may know, a few months ago i bought a pair of Lumley Lampro L300s from Bigears.

https://i281.photobucket.com/albums/kk2 ... 230710.jpg

Whilst they sound excellent, in terms of detail etc, they weren't *quite* right, the bass was in the wrong place.
Where I sat was bass light, the best bass was about 3 feet behind me , and when I walked around the room, the bass was better in some places than others.
These are odd speakers , with a bass port and a huge rear facing speaker too.
Clearly the room was having a major impact on the sound.
last night I started from scratch again, moved the speakers further apart and further forward, just an inch or so, listen, move again. listen etc etc.
Eventually they ended up about 1 foot further apart and 4 inches further forward, and the difference is staggering.
It really does pay to experiment with the speaker placement, and get it right.
Good speakers will become excellent, and average speakers good, if you get them in the right place.
[/url]

Guest

Post by Guest » 13 Aug 2010 09:36

Hi Jolly

yes, Lampros placement in any room under 15 ft in length is a challenge, and they need to be at least 18 inches from the rear wall and slightly toed in. The rear units won't affect bass, but will add to image depth, so need experimentation. The rear adjusters won't add to bass as such, just alter the balance of the units. In a large room, they should drop to 35Hz with ease. In smaller rooms, you'll not get that extension no matter what amp is hooked up as its (as you've discovered) more a function of room size and interaction than power to the speakers. You can plot a graph of frequency response for your room, knowing it's length for a given frequency to see where the "peak" occurs and set your seating position to match. In my room (29ft total open length), I found the best position was 3.5 m out from the front plane of the speakers with them set 2.5m apart which gave a theoretical plus 3dB rise at 50 Hz (ie speating positioned at peak amplitude of 50 Hz wavelength in room).

Matt Lynch
senior member
senior member
Posts: 642
Joined: 10 Aug 2008 14:07
Location: Edinburgh, Scotland

Post by Matt Lynch » 26 Aug 2010 22:17

some interesting thoughts and theories in this thread, but it has become to me at least the golden rule that a change of speakers demands that you forget about where the last ones gave best effect and you start again, all the modeling and papers published can mean nothing unless a specific speaker and set of room conditions becomes the reference,even then it cant be real world type analysis, as a for instance , my mate phones me to say something is wrong with one of his speakers, can i come over? yes i said, i go in and he says listen to this, sure enough from the listening seat the sound was brighter in one speaker! i get up and pick up two magazines and put them under the seat , problem solved . another example, my system sounded quite good in my last flat, i decided i would decorate to keep her in doors happy, she chooses some paint and off we go, after the quick spruce up i carefully put back everything as it was and to my horror the sound had changed for the brighter and the soundstage was kinda confused, what the ~@:£% i say, so i go through the system and every thing is as was, turns out the old paint was matte and the new paint silk finish and that was it, i got the same colour in matte and re did it and all was back as was, would any of the positioning theories or analysis have fixed the issues?
granted i have only had about 20 listening rooms over the years and they have all been shared rooms with other purposes that i had my system in, but what i learned is that you should throw the rule books out the window and trust your ears and experiment, if you cannot find that sweet spot then you consider what might be the problem . to say that A should be this and B should be that is outrageously narrow minded ,
but with that rant over its up to you and your own personal OCD
have fun,
lots of love,
matt

LicoricePizza
senior member
senior member
Posts: 348
Joined: 02 Oct 2010 13:23
Location: 3rd. Rock From The Sun

Post by LicoricePizza » 04 Dec 2010 14:00

"Good speakers will become excellent, and average speakers good, if you get them in the right place."

Excellent quote, that.

Assuming that the listening environment is a basic living room, with furnishings, carpets and drapery, one should experiment with speaker placement until it sounds good to those who are listening therein.
I've found that, based on the flooring material, some carpet bits under the speakers can dampen them from coupling with the floor and help "tighten up" the bass response .
I'm a bit surprised that correct phasing hasn't been emphasised enough, though.
For the uninitiated, incorrect phasing can have a tremendous effect on the overall listening experience.

Post Reply