the home of the turntable

Proper Speaker Placement

amplifiers, receivers and loudspeakers

Proper Speaker Placement

Postby MonkeyBoy » 14 Feb 2010 02:02

Proper Speaker Placement

This is a guide to help those who have the space so they may place their stereo speakers in such a manner as to enable them to get everything out of the recordings they have. The purpose is to place the speakers in such a way that standing waves will not cause interference and cause “dead” ranges of the audio spectrum as well as trying to prevent accentuating those frequencies which would only serve to diminish the listening experience by overshadowing other frequency ranges. It won't help with those portions of the sound spectrum that are inherently weak in many systems, but it will help prevent deviation from what is on the original recording. That's what the musicians want you to hear and it can be assumed that you do as well. What we are aiming for is as flat a spectrum as possible enabling the entire recording to be heard well.

I would like to thank beforehand the nice folks at Cardas Inc. for their permission to use their graphics of speaker placement and the detailed information which has been a great help in putting this guide together.
12521

I would like to also thank our own TNTTNT for his insight and mathematical formula for figuring out where potential interference will probably occur for any given environment as well as his graph showing his own measurment for his own system in his listening area at home.

There are a few different methods for placement which will be detailed here: the 1/3 and 1/5 rules, the Cardas method and the Allen Perkins method. Near field listening placement will be taken up in another guide.

I will also delve into dealing with furniture and it's ability to affect the sound dispersion in your listening room but I will not do into great detail about that here. That is for a sticky on room treatment.

It must be stated that when initially placing your speakers you should place them so there is no toe-in until you have ascertained that they are in the placement that gives the best audio response. This may take days or even weeks or moths of tweaking. Patience is key, but the reward is getting the best sound out of your system that you can.

1) There are two rules for speaker placement which I will go into first, the 1/3 rule and the 1/5 rule. These rules state that after room measurements have been made the speakers will be placed either 1/3 or 1/5 of the distance from the back wall into the listening room. This will help mitigate standing (interference) waves which will cause dead portions of the audio spectrum to be a problem. There is also a school of thought that the speakers will be placed at 1/3 or 1/5 of the distance from the side walls at the same time. I must emphasise that it is the radiating cone or dome of the tweeter that is placed in the exact spot indicated by measurements.

If you have a room that is 20 feet (6m 10cm) long then the speakers would be placed 80 inches from the rear and side walls for the 1/3 method and 48 inches for the 1/5 method. Whichever method you use should also apply to placement of your listening platform (chair or sofa). In other words, if you have your speakers 1/3 of the way from the rear wall, your sofa should be 1/3 of the way from the wall behind the sofa. The same ratio applies to the 1/5 rule. Remember to place your listening seat evenly between the speakers. Also, measure where your ears actually are when you listen so they will be placed at the 1/3 or 1/5 junction, not just near it. You may need someone else to help you with this since they can actually measure distances as you sit in a comfortable listening position. Adjustments can then be made accordingly. Part of what you are trying to accomplish is the partial diminishing of reflections from the side, rear and behind the listening seat walls as well as from the ceiling. You don't want the music to come primarily from the perimeter but from the center of the sound stage radiating outward.

2) The Cardas method states that there are mathematical formulae which can be used to mitigate standing waves and promote a natural sounding soundstage emanating from your speakers. Simply put the speakers should be placed 44.7% of the distance of the total length of the room from the rear wall and 27.6% of the distance from the side walls. It should be noted that this ratio is intended for the woofers, not the tweeters. If your room is 20 feet long and 15 feet wide your speakers would be placed so the woofers are 107 inches from the rear wall and 50 inches from the side walls. Your listening position should therefore be placed 107 inches from the wall behind you. At this point there should be a triangle with angles of 60 degrees at each corner.

12479

12480

These diagrams shows the Cardas ratios for speaker placement regardless of actual room measurements. These ratios are for placing speakers in a room which is longer than it is wide and using the long portion of the room for front to back placement. The ratios are different for using the longer walls as the rear and behind the listening position walls. For those who place their speakers along the longer walls in the room the ratios are reversed; the placement from the rear wall is 27.6% of the total distance from the rear wall and 44.7% of the distance from each of the respective side walls. Again, if your room is 20 X 15 feet and you place your speakers so the longer wall is the rear wall then you would place them so the distance from the rear wall would be 50 inches and 107 inches from each side wall. Your listening position would be placed so your ears are 55.2% of the distance from the wall behind you (100 inches) and centered between the speakers. That would mean 80 inches from the wall behind the speakers.The angles of the corners of the triangle from your listening position will be different than if you used the shorter wall as the rear wall of your listening environment. Note that your ears should be placed at the intersection of the lines from the rear corners to the speakers cones. This should be just about at 55.2% of the length from the wall behind you.

The difference in apparent measurements in the room diagrams is from the “Golden Triangle”, which calculates ratios for a perfect listening environment of 26 X 16 X 10 feet. In each line below the first number is the ratio in the golden triangle. The second number in each line is the number which we are concerned with here. It is the actual percentage of room length or width which dictates where speakers are placed. These are the numbers which I am using for actual measurements in this guide. Note that in these diagrams the actual listening position is never numerically noted so I have given them to you in the text of this guide.

Distance/Ratio
Distance Ratio Numerical Percentage
Speaker to side wall:
RW x .618 5 .276
Speaker to rear wall:
RW x 1 8 .447
Speaker to opposite wall:
RW x 1.618 13 .724
Speaker to speaker:
RW x 1 8 .447

For those unfortunate enough to have a square listening environment you can still use these ratios to your advantage. Your speakers should be placed so the woofers are 27.6% of the distance from the side walls and 44.7% of the distance from the rear wall. Your listening position will be centered at the intersection of the line from the rear corners bisecting the speakers.

12481

3) The Allen Perkins method dictates 1)that the listening position will be near the wall behind you so that bass is accentuated while attempting to minimize sonic interference, 2) the listening position will be placed ¼ of the length of the room from the wall behind you and the speakers placed ¼ of the length from the rear wall and the side walls. It also dictates that minor speaker adjustments can be made along the length or width of the room to fine tune lower frequencies. To tune the mid-bass the speakers will be adjusted along the width of the room and for the lower bass the length of the room is the axis of adjustment. For the first method, the listening chair is placed near the wall behind it and the speakers are placed in a wide stance at approximately ½ of the distance from the rear wall and ¼ of the distance from each side wall. This gives a very wide soundstage after final adjustments have been made to speaker placement. The listening position not at a fixed ratio from the wall behind it but is placed a short distance away from the wall and it is recommended that some sort of sound absorbing material be placed behind the listening position. Generally speaking the listening position will be between 1 to 3 feet from the wall behind it. Make sure your system is wired in phase. This method is used when the rear wall is the longer of the walls in the listening room.

4) The second method dictates that the room be divided into sixteen equal sections with the length of each wall sectioned into quarters. This method is used when the side walls are the longer walls of the listening room.The listening position is ¼ of the length from the wall behind it. The speakers are placed ¼ of the length from the rear wall and ¼ of the width of the room from each of their respective walls.

5) This is what TNTTNT had to say about standing waves. There is also a graph to show the waves that are generated with a 40hz test signal.

40hz equates to a 14 foot distance between sound waves reflecting, which isn't far off my distance from wall to wall on listening axis. The other dimension is about 35 feet.

The way you get to 14 feet is simple. Sound travels at 1132 feet per second. It takes half the wave distance to generate the amplified node.
1132 feet divided by 40hz is 28 feet. Therefore the wavelength distance of a 40hz signal is 28 feet. We need only half this distance to generate a node, hence 14 feet. This matches the wall to wall measurement.
I bet you all the guys here have these nodes. Bigger rooms are better because you still get nodes, but they are lower and close to inaudible. Also the difference between the nodes generated (Several are not just at 40hz as in my example) get smaller, and the gives a flatter feel to the response.

I bought a sub and parametric equaliser to plump the the troughs in my graph, while cutting out 40hz before it gets to my sub amplifier.
I think this is why some speakers suit different room dimensions. A lot of speakers start fading in volume way before you get to 40hz. Perversely, this deficiency may be compensated by a 14 ft room.

12482

>> The line should be flat (The yellow line). The huge peak in volume at 40hz and 150hz is due to the room rebounding sound waves and amplifying these frequencies. They are call room nodes and determined by the speaker placement and shape and size of room.
> >
> > For speaker placement you need to consider: -
> >
> > - choice of lisatening room for size - Larger is better for flatter response
> > - Shape - Cubic rooms where height, length and width are the same are an acoustic nightmare. Rectangular rooms are better
> > - Listening on the longest dimensions of the rectangular room.
> > - Speaker placement - placing speakers at on third of room points reduces room nodes
> > - Furniture and room fabrics/draping will affect sound waves rebounding and room nodes, so moving this can have an effect on sound

6) here are other considerations before you can get every bit of music out of your system for which it was designed. First, there is the matter of final adjustments. When initially placing your speakers you should place them so there is no toe-in. After you have made all your measurements then you need to adjust the forward/backward placement of either one of your speakers. It doesn't matter which speaker you choose, but you will be adjusting one or the other in small increments as small as 1cm at a time. You keep adjusting until the music is correctly centered to your ears. This is necessary due to the way we actually hear. It may sound counter-intuitive at first, but it really does work. This is the part that tends to take the longest for most people. It helps if you have someone who will move the speakers for you or who you know has very good hearing while you adjust the speakers yourself. After you have the music centered then you can toe-in. Generally speaking you want to toe-in so the speakers are pointed to a point roughly 6 inches to 1 foot behind your head. This will be the fine focusing of your speakers. You want to avoid having the sound centered too close to your head so the soundstage will not sound squashed or too wide so the soundstage sounds unnaturally wide and disjointed. One thing I have forgotten to mention is that your tweeters should be placed so they are on approximately the same height as your ears when you are sitting in your listening seat. This is due to the direct sound of the shorter wavelengths in contrast to the less easily detectable sound source of longer wavelengths. This will go a long way toward making the soundstage in your final adjustments sound as natural as possible.

7) Other considerations are room treatments, furniture placement and subwoofer placement. Room treatments and subwoofer placement are too broad for proper detail here . Furniture placement can play a big role in your listening enjoyment. It matters greatly what kind it is, what it's made of and where it is in the room. Large cushioned furniture can absorb a lot of sound and will tend to absorb the higher frequencies the most. The same goes for large rugs, especially those with that are thick and/or have substantial padding underneath. Large bookcases can scatter sound waves helping to diminish standing waves if they are open but if they are enclosed they can often be considered to be the same as walls. Sound will bounce off the hard enclosures instead of being dispersed by the numerous books and objects held within. It is often best if you have a minimum of furniture surrounding the area where you place your speakers due to sound wave dispersion and absorption. Not all of us have the luxury of having a dedicated stereo room. It is also good if you can have a minimum of different sized furniture placed haphazardly around the room. This can create a sonic nightmare that is difficult to rectify without removing the offending furniture. Due to the aesthetic preferences of lovers, spouses and whatnot that isn't always practicle. Suffice it to say that large enclosed furniture can essentially change the dimensions of the room by becoming de-facto walls themselves. Speaker placement should be adjusted accordingly if they are in that side of the listening room. By large I mean any piece of furniture that is taller than the speaker height when they are in their final position and which takes a lot of lateral space. For example, I have two stacked bookcases which stand 5 feet high and are 8 feet wide. These can be considered to shorten the width of the room by 15 inches on that side since any waves won't be bouncing off the wall behind them but off the bookcases themselves. Ideally it would be nice to have nothing in your listening room except your equipment and a place to sit, but this is unrealistic for most of us.

8) I hope this guide has been of help to you. I would like to thank again TNTTNT, Cardas Inc. and Allen Perkins for their valuable information. I cannot include near-field speaker placement since I have never used it and know nothing about it's actual application.

If there is anything that can be done to make this guide better don't hesitate to let me know. I am always willing to learn anything new which helps make the music better for all of us and am willing to pass on any good information.
Music isn't an escape from life, it's one of the reasons to be alive in the first place!
User avatar
MonkeyBoy
long player
long player
 
Posts: 2115
Images: 64
Joined: 31 Dec 2008 21:44
Location: Maryland, USA

United States of America

Postby TNTTNT » 14 Feb 2010 16:22

No Problem mate, glad some of the stuff supplied was put to great use by you. I can imagine there are a lot of people new to hi-fi who will owe you a debt for the advice above, and especially given the time it must have taken you.

I ran measurements using a very good soundcard, measurement mic and acoustics software. I ran frequency sweeps at set volumes and took measurements.

The blue line remains unchanged, but since I discovered my sub was out of phase and volume. I have adjusted crossover, phase and gain on the sub, and am reasonably certain I have filled the trough between 50 and 120hz. I have only attacked this range by feeding my sub with a Parametric equaliser, so the other frequencies have been cut before they get to the sub.

Due to Vista 64 bit drivers, my beloved soundcard is now dead and defunct. I wish to get a new one, and will add the new sweep to this post. I might do a full range sweep as well.

I have used the meters on the Parametric equaliser to show me the trough between 50hz and 120 hz has been filled, and the other parts reasonable untouched.

Other thing to point out is look how low PMC speakers go, for such little cones. They claim 28hz +-3db, and the blue line shows this to be true. PMC Rock!

Great post Monkeyboy, and the next time someone wants a mild tweak to their system using cotton woven by twelve virgins, look at the magnitude of what the room is doing in many cases and speaker placement.
User avatar
TNTTNT
senior member
senior member
 
Posts: 488
Joined: 21 Sep 2009 09:50
Location: Kent

Postby JollyJeweller » 14 Feb 2010 18:15

>>" The Cardas method states that there are mathematical formulae which can be used to mitigate standing waves and promote a natural sounding soundstage emanating from your speakers. Simply put the speakers should be placed 44.7% of the distance of the total length of the room from the rear wall and 27.6% of the distance from the side walls. It should be noted that this ratio is intended for the woofers, not the tweeters. If your room is 20 feet long and 15 feet wide your speakers would be placed so the woofers are 66 inches from the rear wall and 50 inches from the side walls. Your listening position should therefore be placed 66 inches from the wall behind you. At this point there should be a triangle with angles of 60 degrees at each corner. "

Am I being a bit dim here, if the room is 20' long (240 inches) , then 44.7% is 107 inches off the rear wall and 50 inches in ? 66 inches is 27.6% not 44.7% ??
You'll hear Angel Trumpets and Devil Trombones...you are invited.
Delphi/SME/Shelter/Whest/Trigon Pre & Powers/McCormack/NEAT Ultimatums/Squeezebox
User avatar
JollyJeweller
long player
long player
 
Posts: 2659
Images: 2
Joined: 02 Feb 2007 12:58
Location: At home !

Postby MonkeyBoy » 14 Feb 2010 18:25

Letitroll, I had that stuff and had originally included it but then I thought that the subject of nearfield setups deserved their own sticky. Also, I've never used nearfield listening myself (well, not for long anyhow) so I didn't think I was qualified to expound on the subject. If anyone here has a lot of experience with it I'm sure many of our members would appreciate the information. Right now I'm investigating stuff for subwoofer placement. That's a subject that is dear to many of our hearts. Hopefully my computer won't crash on me this time. This was supposed to have been posted about three weeks ago, but with the crash and then the snow delaying everything in the D.C. area it just had to wait.

P.S. thanks for including the graphics. I lost them in the crash, as I lost so much other stuff.

JollyJeweller, thanks for the correction. Where in the world did I get that goofy figure? Even if I had accidentally used the 15 foot dimension it still doesn't work. What an embarrassing goof up! I'm such a schlemiel. #-o :oops:
Music isn't an escape from life, it's one of the reasons to be alive in the first place!
User avatar
MonkeyBoy
long player
long player
 
Posts: 2115
Images: 64
Joined: 31 Dec 2008 21:44
Location: Maryland, USA

United States of America

Postby JollyJeweller » 14 Feb 2010 19:11

Ok, I'm looking at cardas now, for a rectangular room, with the speakers backed up against the shorter wall, the speakers should be the length of the rear wall x 0.447 out (measured to the front of the woofer) and from the side walls , the length of the rear wall x 0.276 (measured to the centre of the woofer ).

So ....my room has a back wall of 360 cm, my speakers are now 161 cms out from the rear wall and 99.3 cms out from the side walls, measured to the front and centre of the woofers.
They are now a bit closer together than i had them and a lot further out, but it's a dedicated room, so will live with it a bit and see how i get on!
You'll hear Angel Trumpets and Devil Trombones...you are invited.
Delphi/SME/Shelter/Whest/Trigon Pre & Powers/McCormack/NEAT Ultimatums/Squeezebox
User avatar
JollyJeweller
long player
long player
 
Posts: 2659
Images: 2
Joined: 02 Feb 2007 12:58
Location: At home !

Postby MonkeyBoy » 14 Feb 2010 23:01

Neil, I look forward to your impression on the tonal qualities and soundstage with the Cardas setup. It's the setup I use and I have been very satisfied with how the music is presented. I've also used the 1/3 rule and I liked that as well. I haven't made up my mind as to which I prefer.
Music isn't an escape from life, it's one of the reasons to be alive in the first place!
User avatar
MonkeyBoy
long player
long player
 
Posts: 2115
Images: 64
Joined: 31 Dec 2008 21:44
Location: Maryland, USA

United States of America

Postby bauzace50 » 14 Feb 2010 23:06

@ Monkeyboy,

this packs substantial observations on the theme. Thankyou for a substantially valuable contribution!

Cordially,
bauzace50
As I watched the stars in the sky I wondered where the ceiling had gone.
User avatar
bauzace50
vinyl addict
vinyl addict
 
Posts: 7785
Joined: 11 Jun 2005 15:48
Location: Guayama, Puerto Rico

Puerto Rico

Postby JollyJeweller » 17 Feb 2010 11:10

OK, 2 days with the speakers placed in a Cardas styleee , and I'm not sure.
It seems to be that the bass has gone weird, become mushy and overblown , almost muffled as well.
Now the Lumleys are odd inasmuch as they are ported and also have a driver that's rear facing which you can dial dial in and out , so maybe they don't respond to rear wall placement like most "normal" speakers.
I might give the manufacturer a quick ring and see what he suggests.

(20 minutes later)....yup, the rear voiced speaker needs to be about 1 metre from the rear wall.
So, i'm afarid it's back out with the measuring tape....sorry!
I'm glad i wasn't imagining it, and there was certainly no placebo effect !
You'll hear Angel Trumpets and Devil Trombones...you are invited.
Delphi/SME/Shelter/Whest/Trigon Pre & Powers/McCormack/NEAT Ultimatums/Squeezebox
User avatar
JollyJeweller
long player
long player
 
Posts: 2659
Images: 2
Joined: 02 Feb 2007 12:58
Location: At home !

Postby goneawol » 22 Feb 2010 18:47

Nice one, Monkeyboy. =D>
goneawol
member
 
Posts: 4780
Images: 389
Joined: 25 Jun 2009 14:12

Postby Klaus R. » 27 Feb 2010 17:53

@MonkeyBoy,

Since you are asking, here some comments.


There are a few different methods for placement which will be detailed here: the 1/3 and 1/5 rules, the Cardas method and the Allen Perkins method.


Methods like these are based on the assumption that the room is empty, has no wall openings and perfectly rigid boundaries. Large absorbing furniture is capable of shifting mode frequencies and decrease mode levels (De Melo). Large reflective furniture is capable of splitting up modes, hence generating two modes instead of one (Bork). Wall openings are structural weaknesses and locations of pressure maxima and minima are shifted (Welti).

Bork, „Modal analysis of standing waves (in German)“, Progress of Acoustics, DAGA ’05, 31st Annual Convention of Acoustics. (German Society of Acoustics), Munich 2005

De Melo et al., “Sound absorption at low frequencies: room contents as obstacles”, J. of Building Acoustics 2007, vol. 14, no. 2, p.143

Welti, “Low-frequency optimization using multiple subwoofers”, J. of Audio Eng. Soc. 2006, p.347


I bet you all the guys here have these nodes. Bigger rooms are better because you still get nodes, but they are lower and close to inaudible. Also the difference between the nodes generated (Several are not just at 40hz as in my example) get smaller, and the gives a flatter feel to the response.


Bigger rooms are better because the Schroeder frequency Fs is lower than in smaller rooms, so the frequency range where problems might occur is smaller. Fs determines the transition from few well separated modes to many overlapping modes (at least 3 modes within the half-power bandwidth of a given mode).

Schroeder, “The Schroeder frequency revisited”, J. of Acoust. Soc. of America 1996, vol.99, no. 5, p. 3240


For speaker placement you need to consider:
- Shape - Cubic rooms where height, length and width are the same are an acoustic nightmare. Rectangular rooms are better.


It has been shown that square and cubic rooms are not necessarily worse than rectangular ones.

Fazenda et al., “Perception of modal distribution metrics in critical listening spaces - Dependence on room aspect ratios”, J. of Audio Engineering Society 2005, p.1128
¬
Wankling et al., “Subjective validity of figures of merit for room aspect ratio designs”, Audio Eng. Soc. Preprint 7746 (2009)


It helps if you have someone who will move the speakers for you or who you know has very good hearing while you adjust the speakers yourself.


In-ear frequency response may show large variations between individuals so having someone else listen on your behalf is definitely not a good idea.

Shaw, “Earcanal pressure generated by a free sound field”, J. of Acoust. Soc. of America 1965, vol. 39, no.3, p.465

Møller et al., “Head-related transfer functions of human subjects”, J. of Audio Eng. Soc 1995., p.300


I’ve made a list of technical/psychoacoustic literature that I consider as absolute minimum knowledge when it comes to acoustics of small listening rooms. Feel free to ask for copies of any paper you want to read.

http://www.hifi-forum.de/index.php?acti ... hread=2045


Further, measurements that are not correlated to human hearing make little sense. Human hearing has some in-built mechanisms that a measuring setup does not have, such as equal loudness contours or binaural decoloration.

http://lib.ioa.ac.cn/ScienceDB/JASA/jas ... /585_1.pdf

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



@ Letitroll,

You are mentioning Audio Physic’s placement method (as presented in Theiss et al., “Loudspeaker placement for optimized phantom source reproduction”, Audio Eng. Soc. preprint 4246). Also here some comments.


The way we locate sonic events in space is by the brain measuring the time delay of the sound between the two ears.


Wrong!! This in true only for low frequencies, for high frequencies it’s the amplitude difference between the two ear signals that is used for source location.

http://www.aip.org/pt/nov99/locsound.html


This spatial information is determined by the brain in the first 800us of the transient because this is the maximum time delay between the two ears.


Within these 0.8 ms summing localization takes place. When the second of two sources emits within this time frame, the apparent source location is midway between the two real sources. This phenomenon allows two-channel stereo to work. At delays above 0.8 ms the apparent source is located at the earlier of the two real sources, the precedence effect or law of the first wave front is operational. The upper limit (echo threshold) of the time frame depends on signal type and is about 50 ms for speech and about 80 ms for slow music.

Litovsky et al., “The precedence effect”, J. of Acoust. Soc. of America 1999, vol. 106, p.1633
http://www.waisman.wisc.edu/~litovsky/papers/1999-3.pdf

Blauert et al., “Acoustic communication: The precedence effect”, Forum Acusticum 2005, Budapest
http://intellagence.eu.com/acoustics200 ... /992-0.pdf


So, the first step to getting a good stereo soundstage is to eliminate early reflections of the leading transient as much as possible. Or, in practice, to have the sound from the speakers arrive at your ears before any reflections.


By definition any reflection will arrive after the direct sound. One must distinguish between reflections arriving with the first 0.8 ms and reflections arriving later than that. The former, caused by mounting hardware on the speakers’ front baffle and by cabinet edge diffraction, may disturb summing localization, the latter are merged with the direct sound and are said to have negative effects on timbre (comb filtering) and sound stage/imaging. In neither cases there is evidence that negative effects actually do exist.


According to a psychoacoustic phenomenon called the Haas effect, the brain prioritizes the first sound wave to avoid confusion, if the reflections are low enough in amplitude.


That effect is actually called precedence effect. Haas has found that a delayed signal (<35 ms) at levels of up to 10 dB louder than the direct sound does not disturb source localization. That is what is called Haas effect and is used in PA installations.

Other designations for the precedence effect that have appeared in literature are:

1849: limit of perceptibility
1927: threshold of extinction
1944: first arrival effect
1949: auditory-suppression effect
1952: law of the first wavefront



The close proximity (1-3 ft) of the head to the rear wall has two effects. At the room boundaries (walls) the room nodes are suppressed - because the sound pressure is high and the velocity is low. Sitting in the maximum pressure area gives the best perception of deep bass.


Directly at the wall surface the incident and reflected wave are in phase and pressure is doubled. At quarter wavelength distance the two waves are in opposite phase and pressure is zero. So at 1 ft distance you are in a null of 280 Hz, at 3 ft in a null of 85 Hz.

Waterhouse, “Interference patterns in reverberant sound fields”, J. of Acoust. Soc. of America 1955, vol.27, no.2, p.247


Secondly, the reflections are shorter than the circumference of the head, so the brain can not measure the time delay between the ears. When the brain cannot localize reflections it ignores them.


When reflections from the rear wall arrive within the summing localization window the apparent source can be perceived in front or behind, depending on delay. And again, human hearing uses both time and amplitude differences between the two ear signals for source localization.

Blauert, “Spatial hearing”, MIT Press 1983


Take the speakers along the axis you have chosen in step l and, ONE AT A TIME, have someone move them distally (toward the side walls) WHILE LISTENING.


The optimum angle between loudspeakers and listening axis is where the interaural cross correlation is zero. In order to obtain zero cross-correlation (maximum dissimilaritiy) in two-channel stereo systems the angles of the loudspeakers w.r.t. the listening axis should be 23, 67, 126, 158º (Damaske et al.), according to Ando the optimum angle for all sound sources is 26º.

Damaske et al., “Interaural crosscorrelation for multichannel loudspeaker reproduction” (in German), Acustica 1972, vol. 27, p.232

Ando, “Architectural Acoustics”, Chapter “Design of electroacoustic systems”, Springer editions, 1998

Klaus
Klaus R.
senior member
senior member
contributor
 
Posts: 349
Joined: 23 Apr 2004 09:57
Location: Netherlands

Postby TNTTNT » 27 Feb 2010 19:43

Klaus R. wrote:@MonkeyBoy,

Since you are asking, here some comments.


There are a few different methods for placement which will be detailed here: the 1/3 and 1/5 rules, the Cardas method and the Allen Perkins method.


Methods like these are based on the assumption that the room is empty, has no wall openings and perfectly rigid boundaries. Large absorbing furniture is capable of shifting mode frequencies and decrease mode levels (De Melo). Large reflective furniture is capable of splitting up modes, hence generating two modes instead of one (Bork). Wall openings are structural weaknesses and locations of pressure maxima and minima are shifted (Welti).

Bork, „Modal analysis of standing waves (in German)“, Progress of Acoustics, DAGA ’05, 31st Annual Convention of Acoustics. (German Society of Acoustics), Munich 2005

De Melo et al., “Sound absorption at low frequencies: room contents as obstacles”, J. of Building Acoustics 2007, vol. 14, no. 2, p.143

Welti, “Low-frequency optimization using multiple subwoofers”, J. of Audio Eng. Soc. 2006, p.347


I bet you all the guys here have these nodes. Bigger rooms are better because you still get nodes, but they are lower and close to inaudible. Also the difference between the nodes generated (Several are not just at 40hz as in my example) get smaller, and the gives a flatter feel to the response.


Bigger rooms are better because the Schroeder frequency Fs is lower than in smaller rooms, so the frequency range where problems might occur is smaller. Fs determines the transition from few well separated modes to many overlapping modes (at least 3 modes within the half-power bandwidth of a given mode).

Schroeder, “The Schroeder frequency revisited”, J. of Acoust. Soc. of America 1996, vol.99, no. 5, p. 3240


For speaker placement you need to consider:
- Shape - Cubic rooms where height, length and width are the same are an acoustic nightmare. Rectangular rooms are better.


It has been shown that square and cubic rooms are not necessarily worse than rectangular ones.

Fazenda et al., “Perception of modal distribution metrics in critical listening spaces - Dependence on room aspect ratios”, J. of Audio Engineering Society 2005, p.1128
¬
Wankling et al., “Subjective validity of figures of merit for room aspect ratio designs”, Audio Eng. Soc. Preprint 7746 (2009)


It helps if you have someone who will move the speakers for you or who you know has very good hearing while you adjust the speakers yourself.


In-ear frequency response may show large variations between individuals so having someone else listen on your behalf is definitely not a good idea.

Shaw, “Earcanal pressure generated by a free sound field”, J. of Acoust. Soc. of America 1965, vol. 39, no.3, p.465

Møller et al., “Head-related transfer functions of human subjects”, J. of Audio Eng. Soc 1995., p.300


I’ve made a list of technical/psychoacoustic literature that I consider as absolute minimum knowledge when it comes to acoustics of small listening rooms. Feel free to ask for copies of any paper you want to read.

http://www.hifi-forum.de/index.php?acti ... hread=2045


Further, measurements that are not correlated to human hearing make little sense. Human hearing has some in-built mechanisms that a measuring setup does not have, such as equal loudness contours or binaural decoloration.

http://lib.ioa.ac.cn/ScienceDB/JASA/jas ... /585_1.pdf

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



@ Letitroll,

You are mentioning Audio Physic’s placement method (as presented in Theiss et al., “Loudspeaker placement for optimized phantom source reproduction”, Audio Eng. Soc. preprint 4246). Also here some comments.


The way we locate sonic events in space is by the brain measuring the time delay of the sound between the two ears.


Wrong!! This in true only for low frequencies, for high frequencies it’s the amplitude difference between the two ear signals that is used for source location.

http://www.aip.org/pt/nov99/locsound.html


This spatial information is determined by the brain in the first 800us of the transient because this is the maximum time delay between the two ears.


Within these 0.8 ms summing localization takes place. When the second of two sources emits within this time frame, the apparent source location is midway between the two real sources. This phenomenon allows two-channel stereo to work. At delays above 0.8 ms the apparent source is located at the earlier of the two real sources, the precedence effect or law of the first wave front is operational. The upper limit (echo threshold) of the time frame depends on signal type and is about 50 ms for speech and about 80 ms for slow music.

Litovsky et al., “The precedence effect”, J. of Acoust. Soc. of America 1999, vol. 106, p.1633
http://www.waisman.wisc.edu/~litovsky/papers/1999-3.pdf

Blauert et al., “Acoustic communication: The precedence effect”, Forum Acusticum 2005, Budapest
http://intellagence.eu.com/acoustics200 ... /992-0.pdf


So, the first step to getting a good stereo soundstage is to eliminate early reflections of the leading transient as much as possible. Or, in practice, to have the sound from the speakers arrive at your ears before any reflections.


By definition any reflection will arrive after the direct sound. One must distinguish between reflections arriving with the first 0.8 ms and reflections arriving later than that. The former, caused by mounting hardware on the speakers’ front baffle and by cabinet edge diffraction, may disturb summing localization, the latter are merged with the direct sound and are said to have negative effects on timbre (comb filtering) and sound stage/imaging. In neither cases there is evidence that negative effects actually do exist.


According to a psychoacoustic phenomenon called the Haas effect, the brain prioritizes the first sound wave to avoid confusion, if the reflections are low enough in amplitude.


That effect is actually called precedence effect. Haas has found that a delayed signal (<35 ms) at levels of up to 10 dB louder than the direct sound does not disturb source localization. That is what is called Haas effect and is used in PA installations.

Other designations for the precedence effect that have appeared in literature are:

1849: limit of perceptibility
1927: threshold of extinction
1944: first arrival effect
1949: auditory-suppression effect
1952: law of the first wavefront



The close proximity (1-3 ft) of the head to the rear wall has two effects. At the room boundaries (walls) the room nodes are suppressed - because the sound pressure is high and the velocity is low. Sitting in the maximum pressure area gives the best perception of deep bass.


Directly at the wall surface the incident and reflected wave are in phase and pressure is doubled. At quarter wavelength distance the two waves are in opposite phase and pressure is zero. So at 1 ft distance you are in a null of 280 Hz, at 3 ft in a null of 85 Hz.

Waterhouse, “Interference patterns in reverberant sound fields”, J. of Acoust. Soc. of America 1955, vol.27, no.2, p.247


Secondly, the reflections are shorter than the circumference of the head, so the brain can not measure the time delay between the ears. When the brain cannot localize reflections it ignores them.


When reflections from the rear wall arrive within the summing localization window the apparent source can be perceived in front or behind, depending on delay. And again, human hearing uses both time and amplitude differences between the two ear signals for source localization.

Blauert, “Spatial hearing”, MIT Press 1983


Take the speakers along the axis you have chosen in step l and, ONE AT A TIME, have someone move them distally (toward the side walls) WHILE LISTENING.


The optimum angle between loudspeakers and listening axis is where the interaural cross correlation is zero. In order to obtain zero cross-correlation (maximum dissimilaritiy) in two-channel stereo systems the angles of the loudspeakers w.r.t. the listening axis should be 23, 67, 126, 158º (Damaske et al.), according to Ando the optimum angle for all sound sources is 26º.

Damaske et al., “Interaural crosscorrelation for multichannel loudspeaker reproduction” (in German), Acustica 1972, vol. 27, p.232

Ando, “Architectural Acoustics”, Chapter “Design of electroacoustic systems”, Springer editions, 1998

Klaus



I agree that human hearing curves need to be a factor when analysing room response. This doesn't negate the act of measurement, but affects the act of interpretation. Even in its simplest linear understanding, room measurements can be useful to quantifying very large room modes.

I am glad that you agree and confirm that bigger rooms are better.

Given you have taken a very scientific approach in dissecting the material, all good scientific papers should have a conclusion. Further, given this is a sticky, it may provide guidance to many future readers. With this in mind, what practical simple advice and summary would you advise on speaker placement and room effects.
User avatar
TNTTNT
senior member
senior member
 
Posts: 488
Joined: 21 Sep 2009 09:50
Location: Kent

Postby Klaus R. » 28 Feb 2010 13:56

@ TNTTNT,


I agree that human hearing curves need to be a factor when analysing room response. This doesn't negate the act of measurement, but affects the act of interpretation. Even in its simplest linear understanding, room measurements can be useful to qantifying very large room modes.


Without appropriate correlation measurements to human hearing you cannot use measurements to predict problems, yet this is what many people do: they measure, look at the graph and think they absolutely need room treatment. Measurements are useful once you have detected a problem while listening to music.


I am glad that you agree and confirm that bigger room are better.


I forgot to mention that concepts like the Schroeder-frequency are based on the assumption that the sound field is diffuse, which it is not:

Meyer, „Definition and diffusion in rooms“, J. of the Acoustical Society of America 1854, vol. 26, no. 5, p.630

Gover et al., “Measurements of directional properties of reverberant sound fields in rooms using a spherical microphone array”, J. of Acoust. Soc. of America 2004, vol. 116, no. 4, pt.1, p.2138

Therefore, the actual transition frequency is higher than what is computed using Schroeder’s formula:

Baskind et al., “Sound power radiated by sources in diffuse field”, Audio Eng. Soc. preprint 5146


…..what practical simple advice and summary would you advise on speaker placement and room effects.


Speaker placement depends on many parameters, so there is no straightforward advice.

Allison effect: small speakers without good bass extension can be placed closer to the walls/corners than large speakers that go down to 20 Hz. Large speakers can be placed close to walls/corners provided they have in-built equalization. If not, out-board equalization is needed to avoid bass boom.

Early reflections: reflections coming within 2-3 ms after the direct sound produce high interaural cross-correlation and are less preferred than later reflections, so if placement to side walls is required, absorption may be beneficial.
Ando, “Subjective preference in relation to objective parameters of music sound fields with a single echo, J. of the Acoustical Society of America 1977, vol. 62, no.6, p.1463

Room modes: since common placement rules are based on circumstances you generally do not encounter in domestic situations, they cannot guarantee optimum results, so simply don’t rely on them. You may use them as starting point, and it may turn out that that position is the best in your room, but equally well you may ask your spouse where she’d like to have the speakers placed in the living room, maybe that that position is the best in that room. If you have a dedicated listening room, you may use one of the many placement rules or just place speakers and listening chair somewhere and look how it sounds.

In rooms with hard reflective walls (brick ‘n mortar, concrete) those placement rules may give better results than in rooms with flexing walls (plasterboard on studs). Since all room modes have a maximum in corners, corner placement will excite all modes equally loud, but you’d have to use equalization.

Furthermore, in order to become annoying room modes have to be excited first and then they have to be perceived. Whatever placement method is used, play music (not test signals) and listen while being seated. Move loudspeakers and listening chair if necessary until it sounds ok to you. In our living room I’ve placed speakers and listening sofa where the wife told me to. When playing test tones the modes are clearly audible but in 7 years I have found only 3 tracks which actually do excite them.


@Letitroll

Thanks for some great links and info Klaus, they are of great value and very interesting to me.


If you want to read more, I have all the articles of that list I posted the link to, so feel free to ask for a copy.

One, I clipped a couple of paragraphs out of the original articles and I'm sure I did them a disservice in this, you should read the entire text of the links I posted.


Well, I’ve read both links and it won’t surprise you that I have some more comments:

NSM Audio

>The following comments apply to truly phase coherent speakers, the majority of which tend to have simple (or no) crossovers.<

It’s perfectly possible to build truly phase coherent speakers with complex crossovers. The answer is FIR-filter. The result looks like this:

Image

> If the positioning causes standing waves or "suck-outs" you can bet the whole tonality of the presentation will be off…<

Any position of a loudspeaker in a room will generate standing waves!!! A standing wave is the result of the superposition of an incident wave with its reflection from a room boundary. Happens all the time and at all frequencies.


Audioasylum post

> In this position [speakers at the two center points of an ellipsoid touching the walls of the room, best listening position 1 to 3 feet from the rear wall.] the sound from the speakers reaches the ears before any reflections coming from the side walls resulting in better soundstaging and an unaltered perception of the speakers tonal balance.<

The direct sound comes first, yes, but the reflections do come and will be merged with the direct sound (precedence effect), the sum will contribute to sound stage and tonal balance. If soundstage and tonal balance are altered by the reflections is not known. I looked at the literature relevant for this issue and while there are indications that some imaging parameter are affected by reflections (blur (Kuhl); phantom image width (Ringlstetter); phantom image location (Linkwitz), there is no thorough and detailed research on this. I've prepared a literature overview, if interested, drop me a mail.

Kuhl, „Effect of a Loudspeaker Radiated Diffuse Sound on the Hearing Event” (in German) Acustica 1978, vol. 40, no.3, p.182

Ringlstetter, „Investigations on the directivity index of loudspeakers „ (in German), Fortschritte der Akustik, DAGA ’96, 22nd Annual Conference on Acoustics (German Acoustical Society), Bonn 1996

Linkwitz, "Room Reflections Misunderstood", Audio Engineering Society preprint 7162


> Imagine the situation of being in a somewhat noisy public place and conversing with the person next to you.<

This is known as “cocktail party effect” and I’m not sure that it is relevant in the present context (direct sound – early reflections).

http://www.media.mit.edu/speech/papers/ ... effect.pdf


> Our brains do this automatically all the time to, for example, filter out the annoying natural resonance of a room to facilitate speech…<

It has been shown that early reflections actually improve speech intelligibility:

http://scitation.aip.org/getabs/servlet ... yes&ref=no


> The point is that, by placing the speakers and/or listening chair at an even division of the room, you will get natural bass reinforcement from the room.<

When I look at the two first axial modes, the 1st order mode has pressure minimum at room center (i.e. an even division), the 2nd order mode has pressure maximum at room center. At any even room division I will have pressure minima of some modes and pressure maxima of other modes. How this results in “natural bass reinforcement” is a mystery to me.

Letitroll wrote: Unless the signal is modulated as in musical signals, then ILD is a factor up to 4,000hz. This is noted in the article you linked.


I think the following passage nicely sums it up:
> To summarize the matter of binaural differences, the physiology of the binaural system is sensitive to amplitude cues from ILDs at any frequency, but for incident plane waves, ILD cues exist physically only for frequencies above about 500 Hz. They become large and reliable for frequencies above 3000 Hz, making ILD cues most effective at high frequencies. In contrast, the binaural physiology is capable of using phase information from ITD cues only at low frequencies, below about 1500 Hz. For a sine tone of intermediate frequency, such as 2000 Hz, neither cue works well. As a result, human localization ability tends to be poor for signals in this frequency region.<

Both ILD and ITD are hence used for source localization, so the statement “The way we locate sonic events in space is by the brain measuring the time delay of the sound between the two ears” is wrong.

Letitroll wrote:
Klaus R. wrote:The former, caused by mounting hardware on the speakers’ front baffle and by cabinet edge diffraction, may disturb summing localization, the latter are merged with the direct sound and are said to have negative effects on timbre (comb filtering) and sound stage/imaging. In neither cases there is evidence that negative effects actually do exist.


This would be in direct opposition to many successful speaker designers who go to great lengths to reduce early reflections. And in opposition to my anecdotal experience in speaker design and modification. But you of course are entitled to your opinion and have every right to state it.


Like all reflections also these very early reflections cause comb filters. To the best of my knowledge there is no research available on the effect of these reflections and hence no evidence, other than anecdotal, that these reflections are a problem. One obviously can always choose to play safe and design cabinets without very these early reflections.


You are quite correct here, the Haas effect is included in the group of theories called in general the Precedence effect. The article should have been more precise. From my reading of it there is no exact hard and fast rule that applies in all conditions, and the effect varies in intensity and delay times depending strongly on environment and type of signal.


The upper limit (or echo threshold) of the time window of the effect depends on signal type. What is relevant for the context of music listening in small rooms is that first reflections generally come within the 80 ms time window where the effect is operational. On rare occasions like hitting the rim of the drum with the stick the time window is possibly closed before the reflection arrives so that this reflection might be perceived as echo. So far I haven't noticed anything resembling that.

From my reading of the entire article and from the diagrams posted, as well as my own experience, it is very important to be as absolutely close to the rear wall as possible, within inches. And if there is any separation it should be filled with highly absorptive material.


Our listening sofa is actually placed against the rear wall so distance between ears and wall is about 20 cm, unless I decide to slide down into a more relaxed position:-) I tried absorbers I had at hand behind my head and could not hear a difference.

Thanks again for taking the time to post so comprehensively Klaus, I'm very grateful for your interest. You have provided me with many hours of engaging reading for the next few weeks.


Back in 2007 I came across Floyd Toole’s paper “Loudspeakers and rooms for sound reproduction – a scientific review”, J. of the Audio Engineering Society, p.451. Some of his conclusions are exactly opposite to what is considered as acquired truth and wisdom in the field of small room acoustics, so I went and read the literature he cites and many more. And yes, there are lots of myths, misconceptions, half-truths out there in audiophile circles.


Klaus
Klaus R.
senior member
senior member
contributor
 
Posts: 349
Joined: 23 Apr 2004 09:57
Location: Netherlands

Postby goneawol » 28 Feb 2010 18:06

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.

Why would it be different, studio monitoring practice is as good as it gets, virtually by definition ? Doesn't make sense.
goneawol
member
 
Posts: 4780
Images: 389
Joined: 25 Jun 2009 14:12

Postby goneawol » 28 Feb 2010 18:13

double post
goneawol
member
 
Posts: 4780
Images: 389
Joined: 25 Jun 2009 14:12

Postby goneawol » 01 Mar 2010 14:32

Thanks LIR. Those links don't address principles and practice in studio monitoring setups at all, and are silent on the point you made reference to ? Rather they seem to be an exploration of the author's hypotheses of why he perceives two different speaker rigs sound similar in his living room ! Recording monitoring gets mentioned twice. Once as a goal. And once the author seems to suggest deliberately introducing living room type reflections into recordings, which is entirely the wrong way round to me ! 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.

Those papers highlight a simple, interesting test though. Record a programme using an omni located at the listening position, then play it back through the same system. Just to hear how much 'room' there is in the sound that reaches your ears. It's amazing. How good a job one's brain/senses/perception does of sorting out the crud and inventing a stable sensation of reality from a pair of speakers !

I didn't get what part of room relflections was 'misunderstood' by the author either, FWIW.
goneawol
member
 
Posts: 4780
Images: 389
Joined: 25 Jun 2009 14:12

Next

Return to Amplifiers and Loudspeakers


Design and Content © Vinyl Engine 2002-2014

faq | site policy | advertising