Home Theater Forum and Systems banner

HTS Two-Channel Speaker Setup Guide for a Deep Soundstage

74K views 123 replies 33 participants last post by  mark3141 
#1 ·
Abbreviations

CCC - Cedar Creek Cinema, Sonnie's home-based cinema and 2-channel listening room
GM = Golden Mean, or Golden Ratio, which is 1.61803398875:1 ~ 1.618:1
GMP = Golden Mean Positioning, or positioning of speakers using Golden Mean guidelines
LP - Listening Position


Introduction

This will be a developing thread. Here is the quick start version, diagrams will follow.

The data and methods given here represent the best we have learned from Sonnie's Two-Channel Speaker Evaluation Events hosted by Sonnie at his home-based Cedar Creek Cinema (CCC). This guide was put together to help answer questions about speaker setup techniques learned during those events (links below). As further events occur and as pertinent additional information becomes available, it will be added to this guide.

$1,000 Speaker Evaluation Event, August 2013
$2,500 Speaker Evaluation Event, November 2013
$3,000 Speaker Evaluation Event, February 2014

Those events have been documented extensively, so they will not be discussed in this thread except where that discussion relates directly to stereo speaker setup for music.

Data from speaker setups in other rooms might be included if it clearly fits the approaches discussed.


Conditions and Caveats

These results occurred in certain room(s) with certain listening priorities. To summarize:
  • The primary goal is a wide, deep, detailed, cohesive, lifelike soundstage with sharp imaging and, where achievable, sharp soundstage depth acuity (few speakers can achieve really precise depth acuity, but when they do, your heart will stop - hopefully only for a moment).
  • The speakers go wherever they need to in order to sound their best, no consideration for WAF or convenience or pets or kids or anything else.
  • Minor sacrifices in frequency response may result in following these recommendations:
    • Toe-in angles help accomplish the desired soundstage and usually put the listener off the tweeter axis, resulting in minor high-frequency rolloff above 5 kHz - down 2 to 3 dB at 10 kHz, more above that.
    • Speaker positioning meant to minimize sensitivity to room modes has, up to this point, pretty much been ignored.
One might very well ask if there is any validity to assuming that these dimensions and ratios are at all universal. I have done detailed speaker setup work and achieve the desired soundstage and imaging results in many very different rooms and at multiple different LP/speaker arrangements in some of those rooms and the basic ratios and results have always ended up very close to those given here. In discussions with the evaluators at the CCC evaluation events they have indicated having similar experiences. So far there have been no major departures from these patterns. When there are, we will do our best to incorporate them in an informative way.

We are doing our best to convert the data into a useful rule-of-thumb setup guide format. But remember it all comes from the way certain speakers sounded to our ears and listening tastes in certain rooms. We hope our attempts here are useful, but they will certainly not apply to all. This is an experiment with the best of intentions, please take it as such.

There are many guides and approaches to room and speaker setup. This guide is specifically for reaching the goals stated above. We have ignored "equilateral triangle" rules and "aim at the listener" rules and rules that reduce room mode interaction and other considerations that common setup guides focus on, because they have different goals and do not help - in some cases they work against - the accomplishment of our defined goals. In putting this guide together, it was seen that the Golden Mean guidelines for reducing room mode interaction could be incorporated in determining logical setup starting points for various rooms. But adjustments from those starting points will usually be necessary, negating those benefits partially or completely.

We appreciate meaningful feedback, but will limit folding in data to that which we are confident meets our criteria and is compatible with our approach to finding that ideal speaker location. For starters, there is one data point from work in my home listening room. The setup was determined totally by ear without referring to notes about setting up the same speakers at CCC a few weeks earlier. The configuration arrived at independently matched very closely what we came up with at CCC in parametric terms, according to our rule-of-thumb approach. A single data point of confirmation, hopefully more will follow.


Defining our Terms

There are two parts to our process:
  • Determine the Starting Points
  • Make Adjustments
There will be a diagram for this eventually. For now, imagine a big letter T. The top of the T is the line from center to center of the fronts of the speakers (B in our table below), along the speaker plane. The "stalk" of the T (A in our table) is the line from the ears at the Listening Position, starting at the ear plane, to the center of that first line, the top of the T.

If your speakers are very close to front wall of your room (more guidance below), you will need to consider one more measurement. To illustrate it, we magically convert the letter T into a big + sign, with its vertical line going from the ear plane to the wall straight in front of the listener (C in our table).

Here is what our data shows:
Average B/A Ratio: 1.4 (ranges between 1.08 and 1.57)
Average C/A Ratio: 2.2 for NON-DIPOLES (ranges between 1.74 and 2.61)
Average C/A Ratio: 2.0 for DIPOLES (ranges between 1.59 and 2.31)

But we are going to use a value of 2.4 for C, regardless of speaker type, not 2.0 or 2.2. The reason follows...


Incorporating the Cardas Golden Mean Guidelines

In our experience, the C dimensions vary more than the A and B dimensions, and appear not to be super critical as long as C is "long enough." By increasing C only slightly from the values arrived at below, the speaker locations - if they can be set first and the LP can be placed accordingly, as in Option 1 below - line up with the popular Cardas Guidelines for room setup, minimizing room mode interaction. This is why the value for C is 2.4 for both dipoles and non-dipoles, rather than 2.0 and 2.2 seen below. From here forward, the value of C=2.4 will be used just to simplify matters.


Decision Point: The Simple Way or The Complex Way?

I have been told that I tend to over-complicate things. That might be true.

At CCC, the LP is fixed by room design, so that is our starting point when setting up speakers for evaluation. We have looked at Golden Mean Positioning of speakers, but the speakers almost always end up somewhere else for the best soundstage, so Golden Mean Positioning (GMP) of speakers is not a top priority.

Starting from the fixed LP and determining the starting point is easy for us. We have marked the spots where most speakers end up or thereabouts, and that is where we start. Easy. And, guess what, those spots fall right where A=1 and B=1.4. No surprise, our experience is what defined those values, as shown in the chart below.

The only reason the extra GMP calculations were added is this: if you have to figure all of this out from scratch for your room anyway, and have the flexibility to place the LP and the speakers wherever you want to, why not make a few extra calculations and make your starting points fall on GMP spots for your room? Maybe you will be lucky and with your speakers you will get the desired deep soundstage with the speakers on those very spots, and then you will have nice, even bass responses, too. Probably not, but maybe. It is worth the few minutes of extra effort to at least try, is it not?

But it does add extra complication that will probably be of little benefit to most. So my suggestion is this:

First time through, do it The Simple Way, with no GMP calculations, just to see how it works out with your speakers in your room and if you even like the results. If you do, then you can go about it the more complex way with the GMP calculations and maybe get smoother bass response, too.


Determine The Starting Points - The Simple Way
  1. Pick a spot for your LP.
  2. Measure from there (center of head, right between the ears) to the wall in front of the LP. That value is C. A good value for C is 12 ft in a medium-large room.
  3. Calculate A: C / 2.4 = A, the distance from the ear plane to the speaker plane, 5 ft in this example.
  4. Calculate B: A x 1.4 = B, the center-to-center distance between the speakers (center of the baffle plate, the tweeter is a good reference), 7 ft in our example.
  5. Place the speakers according to the definitions for A, B, and C.
  6. Now skip down to Make Adjustments below.
That's it, The Simple Way. When you want to try The Complex Way, it is given next.....


Determine The Starting Points - The Complex Way

Option 1 - Best - No Restrictions, Start with the Speaker Positions

Starting with your room width as 3.130 units ((1 + 1.618 + 1) x (1.4 / 1.618)), a weird number, the logic follows) place the speakers spaced between the side walls using units of 0.866:1.4:0.866 (this would normally be expressed as 1:1.618:1, but we have multiplied these numbers by 1.4 / 1.618 or 0.866 in order to keep with our A = 1, B = 1.4, C = 2.4 starting point) and 1.4 units (C - A) away from the front wall (measure from the center of the tweeter). This is good for reducing room mode interaction. Use C = 2.4, which also makes A = 1.

For example, in a room 15.66 feet wide, speaker spacing would be 4.33 feet from side walls, 7 feet from the front wall, and 7 feet apart (B). The LP would be 12 feet from the front wall (C) and 5 feet from the speaker plane (A). Following the Golden Mean (1.618:1), this places the speaker positions with left/right room symmetry where they will be least affected by room modes.

This all works out in most rooms anyway. We suggest always keeping the LP at least 3 feet from the back wall and its reflections, so in some smaller rooms the A, B, and C values must be reduced, keeping the same ratios, until that 3 feet guideline is satisfied. Then the Golden Mean guidelines are no longer satisfied.

Now skip down to Make Adjustments below.

Option 2 - 2nd Best - No Restrictions, Start with the Listening Position

Place the center of head of the LP equally between the side walls, as room symmetry allows, and make the distance from the front wall - dimension C - equal to 2.4 units of measure and the distance from the back wall 1.48 units of measure. Set the distance from speaker to speaker at 1.4 units of length (dimension B), and the distance from the speaker plane to the LP at 1 unit (dimension A).

For example, if the room is 19.4 feet long, locate the listener's ears 12 feet back from the front wall (C), leaving another 7.4 feet to the back wall. The speakers will be spaced 7 feet apart (b) and the LP will be 5 feet from the speaker plane. Following the Golden Mean (1.618:1), this places the listening position with left/right room symmetry where it will be least affected by room modes.

This all works out in most rooms anyway. We suggest always keeping the speakers at least 3 feet from the side walls, so in some smaller rooms the A, B, and C values must be reduced, keeping the same ratios, until that 3 feet guideline is satisfied. Then the Golden Mean guidelines are no longer satisfied.

Now skip down to Make Adjustments below.

Option 3 - 3rd Best - Other Conditions, Exceptions

If the LP or speaker location is set, use it as your starting point and determine A, B, and C the best you can given that limitation, also following the other guidelines in Options 1 and 2 the best you can. Remember that C is not critical as long as it is "big enough," so if conditions force it to be bigger than 2.4, that is fine. The same applies to mid- or near-field monitoring in a large room.

When speaker spacing gets overly wide, the soundstage and center imaging fall apart. Large rooms might also force you to abandon Options 1 and 2 for that reason (this is also covered in the next section.) If so, use a smaller-spaced configuration, focusing on A and B and ignoring C if it goes above 2.4.

Now skip down to Make Adjustments below.


Really Big Rooms

Our experience has been in rooms with ceilings under 10 ft (the CCC ceiling is 8.5 ft high) and with side walls and front wall within "spitting distance" of the speakers. Soundstage development as we like to hear it depends on diffused reflections from these surfaces, along with the more diffused later reflections from the back of the room behind the LP. In very large rooms, very long rooms, or outside, it is unknown how well any of this applies.


Make Adjustments

With dipoles, point the speakers straight at the ears at the LP. While listening, adjust the angles outward by equal increments (check angle symmetry with a laser pointer along a flat surface of the speaker cabinet to reference points on the LP chair) until the soundstage with a strong center, tight imaging, and good depth - it will not take much of an angle change from the starting angles for this to happen with dipoles, probably only a few degrees.

With a non-dipole design, start with the speakers pointed straight forward, not at the LP at all, but at the back wall (0 toe-in), and adjust them inward by equal increments while listening until the soundtage appears with a strong center, tight imaging, and good depth. This could occur anywhere from a few degrees of toe-in to 25 or 30 degrees, highly dependent on the speaker dispersion pattern.

You are listening for a soundstage that extends well beyond the width of the speaker spacing, tight, clear imaging that does not move on different notes or smear on sibilants, and - here is the magic element - you are not done until you achieve this - soundstage depth almost as deep as the width. With rare speakers, the acuity of the depth of positioning of voices and instruments will be down-to-the-inch precise in the depth dimension. In the width dimension this kind of precision is a minimum requirement for acceptable speakers, although many do not achieve it. Also pay attention to high-frequency response changes, as too wide an off-axis angle can cause excessive rolloff.

Make angular adjustments a few degrees at a time, or estimate an amount of change successive-approximation style. A $4 protractor might help, but mainly verify symmetry. Measure, measure, measure. A laser pointing device is a must, a laser measuring device is better. (Verify proper alignment of the laser within the body of the pointer.) Be precise to the inch, check angle symmetry to reference points at or around the LP to the inch, or - even better - to the back wall for greater sensitivity, assuming the relationship of your reference points to the LP have been measured or are known to be symmetrical by room design & construction.

After making an angle change, if you do not appear to have located the sweet spot at the LP, see if you can go and find it. Move around slowly while listening so you can determine how close you are to the sweet spot. Check the sound a few inches higher or lower, maybe much higher, or even in standing position. Be sure you have fond the right listening height for your speakers. Lean or move forward from the LP keeping your ears at the ideal height, like a "strutting rooster," move way up between the speakers, try behind the LP. Move your head side-to-side to see how sensitive the soundstage is to that movement. Experiment. Once you have located the best sweet spot for the current speaker setup, estimate how far the next adjustment should be to move it to the LP.

If the deep soundstage never happens, or if the center of soundstage seems weak or empty or has poor image clarity, start over with the speakers spaced closer together. If the deep soundstage occurs but you want to try to make it deeper, start over with the speakers spaced farther apart. If the angle adjustments seem too sensitive, start over with the speakers closer to the front wall, or the LP farther back from the speaker plane. If the angle adjustments do not seem to make much difference, start over with the speakers farther from the front wall, or the LP up closer to the speaker plane. If you notice excessive high-frequency rolloff, the off-axis angle is excessive (or the LP height might not be right), and you will need to adjust accordingly.

Tip: Drape a folded-up plushy blanket over the back of the LP seat if it has a back that approaches ear height. It will reduce reflections, improve the imaging and soundstage, and make them less susceptible to head movement and position.

When you find a setup that works, make careful measurements and write it all down, using reference marks for easy checks or resetting if something gets bumped or moved. Gaff tape for floor position and reference dots on the back wall for checking angle symmetry with a laser make a 10-second setup check possible. Do a quick setup check before every listening session.

Enjoy your great soundstage!:sn:


Table of Setup Data

Here is the data used to generate the spacing ratios.

The calculation of the A/B Ratio includes data for all speaker types. The calculation of the C/B Ratio is done with an averaging of values for only dipoles and a separate averaging of values for all non-dipoles.

No code has to be inserted here.
(18.21+1.46)/14~1.405=1.4
(22.65+1.81)/11=2.22364~2.2

No code has to be inserted here.
5.85/3=1.95~2.0

** Conforms with Cardas Golden Mean Guidelines under some conditions, so C=2.4 is always used for convenience and simplification. Explained above.




Early Reflection Guidelines:

Perfect room symmetry is not absolutely necessary. General symmetry plus a few treatments probably do the trick. A room with longer RT60 times might have "unbalanced spaciousness" effects where one side sounds bigger than the other. My own listening area is far from symmetrical, but RT60 is quite low and I do not notice anything like this. With longer RT60, absorptive treatment can help balance this effect out.

Early reflections can really throw off the imaging and soundstage. Treatments here are very beneficial. My own rules for early reflections:
  • Less than 5 mS (of additional delay relative to the delay of the direct wavefront from speaker to LP; on an impulse response plot, it will be easy to read that relative delay directly): disruptive to image clarity, must be absorbed.
  • Between 5 mS and 15 mS: must be VERY summetrical (direction and delay time) and must be perceived as coming from the vertical line of sight from speaker to LP; if the reflections are NOT coming along the vertical line of sight from speaker to LP, or if not symmetrical, then either absorb or scatter them. Scatter them with a reflective surface pattern that does not reflect ANY of the sound directly at the LP (I have gotten into long arguments about this last point, but experiments have proven to me that it is very important for the best imaging, and I do not care what expert says otherwise, they are WRONG).
    • Example 1: ceiling reflections (flat ceiling) would be on that vertical line of sight, and might be only delayed by 8 mS or so, and would probably be highly symmetrical. They are OK, though, will not smear the image L or R, and will add a small amount of spaciousness to the sound, as well as some height info to the soundstage.
    • Example 2: in a symmetrical room, reflections off the side walls, if only a few feet from the speakers, would be symmetrical, and would fall within the 5 to 15 mS window, but would NOT be on the vertical speaker to LP line of sight, so imaging would be smeared and softened, so they are not OK. Absorb or scatter (see above).
    • Example 3: Reflective panels on the front wall on the vertical line of sight from LP to speaker could be angled so the reflections are on or very close to that line of sight and are very summetrical and fall in the specified delay range. These are OK, can enhance the soundstage and imaging. However, the natural reflection points on the front wall just inside those panels would not fall on that line of sight and would mess up the imaging. Absorb or scatter (see above) at those natural reflection points.
  • More than 15 mS: less critical. In general, better scattered, and the direction of reflection is not important, although randomized is better.



Related Articles

A Soundstage Enhancement Experience, refining and enhancing the soundstage and imaging with MartinLogan ESL hybrid electrostatic speakers in Sonnie Parker's Cedar Creek Cinema, March of 2015.
 
See less See more
#124 ·
Thanks for putting together this information. It looks like the recommendations are largely empirically based, but can anyone explain why these positions tend to work best? I've seen various places recommending placement based upon the golden ratio, but I haven't seen any explanation of why beyond vague statements like it 'reduces room modes'.

In addition to being an interesting question, I'm curious because another forum recommended I deal with my odd shaped room by placing the speakers centered the room's diagonal instead of centered between two parallel walls. This has been much better, but the placement was largely arbitrary - I looked for a spot that got the speakers away from wall while leaving enough room for my listening position.

Unfortunately, you cannot directly mimic the ratios provided here for this arrangement. For example, if I keep the distance between the speakers the same as the distance between the midpoint of the speakers and the wall, then the line between the speakers will have dimensions 1:2:1 (moving along the line between the speakers, distance from wall to first speaker is 1, distance between speakers is 2, and distance from second speaker to wall is 1).

Any thoughts on what I should be looking to do in my case? Thanks!

And for what it's worth, I have REW measurements from the current position as well as a lay-out of my room available. I'd love some feedback and thoughts on how to start putting together room treatments, but should this wait until I find a better spot for placement, and in either case, is this something I should place in another thread (if so, recommendations on which thread?).

Thanks again!
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top