The 38% Rule from Wes Lachot

[Eric.Desart]

Several years back Wes Lachot (a very nice and gentle guy) came with a rule of thumb to define a distance from the front wall of a control room, in order to minimize x-axis (length) disturbances of the length axial modes.
Hence the idea is not to sit in dips or explicit peaks from these first order modes.

This also dated from a period that the length axial modes were considerer to be the, or at least a defining common factor.
Wes was aware that there could be other problems as wel, but it was a rule of thumb which indeed could help people to find a start point.

Since a real soundfield is much more complicated, and other modes and boundary interference come into the picture as well measuring and playing with speaker and listener position is always the safest way to go.

Now Wes did NOT sell this as the "Holy Grail" himself, but as a rule of thumb, to at least exclude some typical problems.
The problem from lots of stuff is that what starts as a well meant and useful aid, fast diverts in an axiom on the net. 38% is easier to remember and repeat, then bothering where this 38% originated from and what it's boundaries were.

I can't remember exactly how Wes told me at the time but I vaguely remember that he superposed the 4 or 5 first length order modes and somehow extracted this 38% as a point where, at least taking these length modes into account, where no explicit dips and peaks, giving a good response on all the modes involved.

I saw a discussion here now between S.F and R.G. where I find Scott to use to much a slogan like: "Forget about these numbers", and R.G. disagreeing telling this was a good starting point.
1) Indeed S.F. did not clearly distinguish between types of rooms by rejecting this 38% as some nonsense. If this is a general applicable comment then I have my question marks with it.
2) R.G. rightfully defended his point and then later edited his post after noticing that the room in question was a tracking room rather than a recording room where indeed positioning of whatever is an entirely different thing.

Anyhow having such a starting number for control rooms feels as a good thing to me, knowing that a serious amount of reasons can cause this number to be adjusted and knowing that measuring is always the safest way. The base idea of this rule of thumb NEVER was, also not by its author, to substitute room acoustics.
Even using such a number helps people to understand what it stands for. This 38% originates from something. The idea behind it is educational.

This discussion however triggered me to really calculate the effect of these length modes ( the typical Eric approach). I don't see Wes as a typical mathematical guy but as an experienced designer.
In fact I doubt that Wes really calculated this, but just played with with these modes (possibly in a graphical manner) finding this 38%. If not I really should like to know the mathematical approach Wes used.

I made some graphs now in which I calculated this effect, but it's not still 100% clear to me what to do with it in function of interpreting this data.
Hence I come back to this and should appreciates Terry's thoughts about it.
Also clear is that this 38% is not that obvious. As interesting is that to be avoided spots become visible. (all within the limitations of this approach of course)
The first page is just an overview to see how all these graphs relate, but hardly readable in function of details.
The following pages become clearer:
SWSP-01 is the situation when only the 2 first order modes are taken into account. (as you can see Rod that's not 37.5% as you told but around 33%)
SWSP-02 is the situation when only the 3 first order modes are taken into account.
SWSP-03 is the situation when only the 4 first order modes are taken into account.

The full lines assume equal energy (power) in all modes.
The dotted lines assume that I increased the power of the subsequent modes with 2 dB per increasing order.
The HOLLOW dotted lines assumed that I decreased the power of the subsequent modes with 2 dB per increasing order.
And for these last 2 approaches I'm still wondering, since with my tired mind know and the way I calculated this by increasing the power of the higher order modes, I have the temporary feeling that I favoured the fundamental mode (sounds reversed, I know).



Later more

[JohnPM]

Was the 38% based on speakers being flush mounted in the wall, or was there a similar recommendation for their position?

[Eric.Desart]

Was the 38% based on speakers being flush mounted in the wall, or was there a similar recommendation for their position?

John,

Wes made this just in function of the listening position, in order to help people to define a good sweet spot.
He nowhere defined anything of the speakers.

But analysing Wes reasoning it's easy to extend this to much more.
As far as I see it (because Wes never described the exact method he came to that number), and from what I could understand from Wes, Wes started from equal exited modes (the first 4 or 5, I can't remember).
It's clear that depending on the position of the speakers (whether flush mounted or not), this is a theoretical approach which will immediately be influenced when these modes are excited asymmetrically. For this I wanted to see the effect with my increased/decreased power on these subsequent modes.

But basically, what happens here is that this also defines the spots where you have equal coupling with the speakers themselves (always taking the limitations into account that we only speak about few x-axis modes).
Whether you see this as length, width or height, whether you see this as optimum position for speaker placement it's all the same.
You just define the spots where the symmetry of the pressure between the modes involved is optimal (in as far as possible), by definition that's also valid for the coupling of speakers with the modes.

Did this answer your question? Take into account that I also had to read between the lines in my discussions with Wes at the time. This 38% became stronger I think on the net than Wes ever thought it should be. It was just so handy for many to have some handle that it became some axiom.

[Rod Gervais]

everything that Wes has concluded he does through the use of music rather than acoustic mathematics.

Everything he discusses in this regard is done using both Architectural and Musical Scales.

Here's an article he wrote that should help you to understand his approach.

http://www.weslachot.com/new/articles_scales.html

I do think you'll find his approach interesting.

On the issue of the 38% -- you're very correct in your thinking......... this is not inteded to be a solution for modal issues within a room - nor a magical location where every type of acoustic interference suddenly dissapears, but it really is a pretty good starting point for people to begin checking their room out.

Whenever I catch someone on the web suggesting that this is the magical sweet spot - I don't take very long to explain to them that it is just a beginning point - nothing more - nothing less.

Heck - you have to start somewhere - and to leave (especially) a novice blind inside a room and just suggest that they begin anywhere and move around until they find the sweet spot is going to make life very difficult for them.

Seeing as even very small changes in location (as well as temperature and humidity levels) can introduce some fairly drastic differences acoustically - I can see where it would be very frustrating for a novice to just begin anywhere within a room.

I try to convince them t omake perhaps 10 sets of at least 10 tests each for a location - and then to average those tests - and generally see where they have taken just 2 or 3 tests....... and then they have to move the microphone and speakers 1/8 or 1/4" to begin all over........

So I do lead them in this direction - for what I consider good sound engineering practices , one of which happens to include the suggestion from Wes in this regard (that after discussing with him all of his reasons why in some depth).

Wes is a very busy studio designer for a variety of reasons - not the least of which is that he designs rooms that work acoustically (the fact that they are also architecturally beautiful is secondary).

Sincerely,

Rod

[Rod Gervais]

double post

[bert stoltenborg]

What drastic things does difference in humidity and temperature cause?

[JohnPM]

Eric, I'm not sure how valid these rules of thumb are in actuality, even if we lived in a 1 dimensional world . I have a feeling people tend to forget the contribution of the direct sound and look at the behaviour of the modes alone. By coincidence I knocked up an axial mode simulator this weekend to generate some test impulse response data for new LF analysis routines I'm working on. The frequency responses from the model are perhaps not what one might intuitively expect, though I don't think I made any particularly rash assumptions in generating it. Here are the responses for 1% steps from 33% to 41% from the rear wall in a room 5m long, with a speaker flush mounted in the front wall, plotting only the length effects (modal and boundary) assuming a frequency-independent absorption factor of 20% for both walls.


For comparison, here are the responses for 38% and 50%, one would expect 50% to be pretty poor perhaps but...


Maybe I just did something stupid in making the model, but it seemed pretty straightforward.

[Rod Gervais]

What drastic things does difference in humidity and temperature cause?

Lot of things - for example - the absorption values of room treatments - the absorption value of the air itself - the location within a space of phase cancellation or enhancement - the speed of sound itself (which is going to cause some of the other problems) the decay rate within a space, etc.

That is why the ISO and ASTM standards place an importance (and not a small importance) on temperature and humidity when testing. This from ASTM C423:

6. Interferences

6.1 Changes in temperature and relative humidity during the course of a measurement may have a large effect on the decay rate, especially at high frequencies and at low relative humidities. The effects are described quantitatively in ANSI S1.26. These effects due to temperature and relative humidity changes can be accounted for by the procedure in 6.2.

6.2 It is advisable to make measurements in the room when it is empty and in the room when it contains the test specimen under conditions of temperature and relative humidity so nearly the same that the adjustments due to air absorption do not differ significantly. In any case, the relative humidity in the room shall be greater than 40 % during the test. Unless the conditions given in Table 1 are satisfied, decay rates for the measurements in the 1000 Hz one-third octave band and above in both the empty room and in the room containing the test specimen shall be adjusted by subtracting the decay rate due to air absorption, determined according to ANSI S1.26, from the decay rate calculated according to 11.4 (see especially 11.4.1 and 11.4.2).

Some interesting reading on the subject, entitled Environmental Effects on the Speed of Sound* written by Dennis A. Bohn

http://tinyurl.com/Effect-of-environment-on-Sound

All of this is why my designs for HVAC systems in studios always include humidity control.

Have a great day:

Rod

[Eric.Desart]

John,

I'll check later back.
Don't know now in how far it matters but my graphs are no response curves.

Come back later on this, also to understand your graphs.

[Scott R. Foster]

While I address this post to Eric, I would appreciate everyone else's thoughtful input as well – and I want to thank Rod in specific for precipitating this discussion and thank Eric for expanding it – I think this issue is worthy of more examination and that this forum is the right place to do it. Please forgive in advance the length of this post – I really am interested in where this discussion might lead in improving our collective ability to advise others – especially my own ability to do so in this context.

================================

Eric:

Very interesting - we see again that when your brain [even as tired as it may be] turns to these nuggets of "common wisdom" we find in the net-universe of small room acoustics, the operative word becomes "common" and the reference to "wisdom" inherent to the nugget recedes. I would be very interested in knowing your suggestions – if you can make any – for a rational "starting place" for the room depth dimension of the defined listening position when normal folks seek to layout a critical listening suite [control room / audiophile listening room / prime seat for home theater] as part of their project.

I realize you might well feel that the cumulative caveats you would attach onto any single number "rule of thumb" for listener position room depth dimension will render the proposition a less than brilliant idea - but I run into this question 100's of times a year dealing both with Ready Acoustics customers and with DIY acoustic treatments enthusiast who have tracked me down to ask a few questions about their project.

Many of these folks arrive to the discussion with "38%" stamped on their foreheads arising from their reading of non-scientific sources on the internet - and I find this irksome – just as you suggest in your initial post of this thread. Despite my disdain for the idea that expressing the suggestion of a starting place for the LP as a single digit % of room depth I want to be clear that I often make suggestions on how to get started with setting the LP for critical listening purposes. Also, I want to be clear, I would be delighted to be able to make such a precise suggestion for the LP if a man skilled in the art such as yourself were to advise me on what that number might be.

Background - I have never thought that advising folks of a starting place for the LP was a bad idea - nor have I, or would I, chastise Mr. Lachot for having developed one - or for his having shared such in an attempt to help others. I gather Mr. Lachot's 38% suggestion is based, at least in part, on his empirical knowledge [which I find compelling]. Also, per some followers of Lachot who have advanced the 38% suggestion, I gather there is an element of trying to apply the ratios of the even-tempered musical scale to the geometry of room acoustics [which I find much less compelling]. I have no idea of Mr. Lachot's strength of feeling on the "38% suggestion", but I do know that the concept of a suggestion has been elevated to a "38% rule" via assorted net-gurus. I don’t blame anyone per se for the elevation of the number 38 to “rule” status – but without regard to guilt or motivation, a week does not go by that I am not invited to a discussion on room setup by someone who is convinced that 38% of room depth is the THE PLACE where their listening position MUST be. It is not uncommon for a user to relate to me their frustration that their LP is forced to 36.5% because of some practical matter and express deep concern for how they might adjust for this flaw [should I build a new wall?].

As you know - I hold that expressing suggested LP as a single digit % of room depth dimension is a misleading description of the problem and I have long felt that the 38% suggestion was destined to become an erroneous idée fixe. For whatever it might be worth - I have always suggested that the place to start in setting the LP is a zone – or range of values – not a single point.

To make an example of my concern and understandings – take for given a 14' deep room. The difference between 38% and 36% is 3.36"... I move my head about 4 times that far between sitting back in my chair versus leaning slightly forward to place a hand on the mixer in my own control room. We do not listen from fixed positions even in the most controlled environments, and it just gets worse in a real world application… folks change posture over the course of a listening session… some even go so far as to sit in a chair with wheels! Also people come in different sizes and they work in different postures – then, as suggested above, they move around a bit. Therefore as a matter of practical reality, there is zone that designers of a room must consider as potentially being occupied - not a point. Expressing the LP as a point is therefore wrong – it should instead be expressed as what it is – a zone or mappable phase space of probable location.

As such, it is not that 38% strikes me as an unreasonable place to start - it is that 38% conveys a false sense of exactitude and a false sense of the goal. If conveyed as a point, a sense of calculated and verified suitability is conveyed. This obviously ignores the fact that the user will occupy a zone – not a point – but also, I feel this ignores the well established scientific concept of never reporting experimental results with greater exactitude than the experimental process can justify. In other words, best practice is to report the precise result with % uncertainty - worst case is to over state the accuracy of the result by using something like a single number to report results when the experimental methods are nowhere near that precise - and thereby mislead. I feel the “38% rule” that has come to pervade on the net ignores the former and has predictably lapsed into the latter.

Oddly enough, it occurs to me that many who have advanced the 38% rule know there is a problem with the degree of exactitude conveyed by the figure 38. This is evidenced by the fact they will mention the figure 37.5%, as the calculated result and then go on to admit that such a degree of exactness is overstating the case, and segue into the 38% figure. In essence I agree that 37.5% is silly – the nit I have picked is that 38% is still so exact that it does not escape the silliness shadow of 37.5%. Moreover this form of expressing the goal is flawed precisely because it defines a point for the “sweet spot.” Which oddly enough, when the “sweet spot” is graphically presented, is typically not shown as a spot at all, but rather as a circular or oval shaped area where the listener is “in the good zone.” .

To me it has always seemed that the whole explanation is upside down – it is rational to instead explain that there are some single points worth discussing – namely, that one should avoid 25 and 50% - for both the LP and the speakers – and there is a good area for proposing to place the LP. In short - stay away from here, and here, and instead gravitate to elsewhere, for example, this area here.

The other flaw I find in the 38% rule, is that it addresses a very small set of the relevant and rational elements of room acoustics that play on the suitability of any given LP – but this is too long already, so I’ll just leave this comment to stand naked.

To put all of the above in my own personal context, using a 14’ deep example room and the trend of discussions I am all to familiar with - I have suggested that the user should slap LP chair down somewhere between 5 and 5.5' off the front wall - set up the rig rationally given fixed realities such as the size of the mix desk, etcetera - try hard to keep the speakers at least a few inches away from 42” off the front wall [25%], then start listening and/or measuring. If you have a significant problem that can be address by changing layout at all, you can probably find satisfaction by moving the speakers and/or the front/rear wall treatments. Change one thing at a time – be systematic. Find the best spot you can for smooth low frequency response. If not delighted with the results, move the LP and try again.

To instead express these thoughts in a numerical context versus the 38% rule see the number string diagram below.

room depth.jpg (116.02 KiB) Viewed 3029 times

I would very much appreciate your thoughts on a single digit suggestion – or in lieu of that any advice you might be able to grant with regard to how I might improve making suggestions for seeking a range of “known good” starting positions.

Thanks in advance.

[avare]

This thread is great. Thanks Eric.

When I talk with people about the 38% guideline and they get get adamant about the implied precision by the two significant digits, I suggest that they think about it as 3/8 rounded. The precision, or lack of it becomes obvious then.

Andre

[Rod Gervais]

Background - I have never thought that advising folks of a starting place for the LP was a bad idea - nor have I, or would I, chastise Mr. Lachot for having developed one - or for his having shared such in an attempt to help others. I gather Mr. Lachot's 38% suggestion is based, at least in part, on his empirical knowledge [which I find compelling].

As do I.....

Also, per some followers of Lachot who have advanced the 38% suggestion, I gather there is an element of trying to apply the ratios of the even-tempered musical scale to the geometry of room acoustics [which I find much less compelling].

Actually - at least in Western Cultures, I find this rather compelling......... first of all - everything we lean towards in the understand of room modes and other acoustic anomalies has to do with frequency as it relates to the room - and it is really the even tempered musical scale that controls (for the most part) the music in western culture.

With the exception of perhaps the pedal steel, violin (and it's cousins) and some fretless guitars - the even tempered musical scale pervades - and even with those fretless instruments - with the exception of the slide to position - they adhere to the convention as well.

So the only difference between the approaches is that one defines the problems as frequencies with particular lengths - and the other as musical notes (which contain the same particular lengths).

It is not so difficult a reach for me.

I have no idea of Mr. Lachot's strength of feeling on the "38% suggestion", but I do know that the concept of a suggestion has been elevated to a "38% rule" via assorted net-gurus.

Wes does not view this as an absolute - but certainly as a decent place to begin testing.

I don’t blame anyone per se for the elevation of the number 38 to “rule” status – but without regard to guilt or motivation, a week does not go by that I am not invited to a discussion on room setup by someone who is convinced that 38% of room depth is the PLACE where their listening position MUST be. It is not uncommon for a user to relate to me their frustration that their LP is forced to 36.5% because of some practical matter and express deep concern for how they might adjust for this flaw [should I build a new wall?].

Simply continue to educate then as to the error in their knowledge.

As you know - I hold that expressing suggested LP as a single digit % of room depth dimension is a misleading description of the problem and I have long felt that the 38% suggestion was destined to become an erroneous idée fixe. For whatever it might be worth - I have always suggested that the place to start in setting the LP is a zone – or range of values – not a single point.

I do not necessarily see that there is an issue with either approach - as long as the person who is on the receiving end of the suggestion understands the importance of small movements when moving through the testing process.

To make an example of my concern and understandings – take for given a 14' deep room. The difference between 38% and 36% is 3.36"... I move my head about 4 times that far between sitting back in my chair versus leaning slightly forward to place a hand on the mixer in my own control room. We do not listen from fixed positions even in the most controlled environments, and it just gets worse in a real world application… folks change posture over the course of a listening session… some even go so far as to sit in a chair with wheels! Also people come in different sizes and they work in different postures – then, as suggested above, they move around a bit. Therefore as a matter of practical reality, there is zone that designers of a room must consider as potentially being occupied - not a point. Expressing the LP as a point is therefore wrong – it should instead be expressed as what it is – a zone or mappable phase space of probable location.

Unfortunately I cannot completely agree with this....... although it is true that the listener will exist within a zone- the fact is that the testing mic does not.

The microphone is not capable of independent movement - only of repetitive measurements - and (if the person performing the tests is patient enough) it will eventually yield a single spot which is the most advantageous place to listen from.

This spot is then the center of the sweet spot zone you refer to - and it is within that area that a human being is going to fit.

As such, it is not that 38% strikes me as an unreasonable place to start - it is that 38% conveys a false sense of exactitude and a false sense of the goal. If conveyed as a point, a sense of calculated and verified suitability is conveyed. This obviously ignores the fact that the user will occupy a zone – not a point – but also, I feel this ignores the well established scientific concept of never reporting experimental results with greater exactitude than the experimental process can justify. In other words, best practice is to report the precise result with % uncertainty - worst case is to over state the accuracy of the result by using something like a single number to report results when the experimental methods are nowhere near that precise - and thereby mislead. I feel the “38% rule” that has come to pervade on the net ignores the former and has predictably lapsed into the latter.

But that is the fault of some people - and not the fault of the 38^.

Something does not become less useful as a starting point simply because some people mistake it for the finish line.

Oddly enough, it occurs to me that many who have advanced the 38% rule know there is a problem with the degree of exactitude conveyed by the figure 38. This is evidenced by the fact they will mention the figure 37.5%, as the calculated result and then go on to admit that such a degree of exactness is overstating the case, and segue into the 38% figure. In essence I agree that 37.5% is silly – the nit I have picked is that 38% is still so exact that it does not escape the silliness shadow of 37.5%. Moreover this form of expressing the goal is flawed precisely because it defines a point for the “sweet spot.” Which oddly enough, when the “sweet spot” is graphically presented, is typically not shown as a spot at all, but rather as a circular or oval shaped area where the listener is “in the good zone.” .

Again - that circular or oval space has a perfect center - and it is that center that people are striving to find - you know that a move of even inches or fractions of an inch introduce some fairly significant differences in readings.

Is it so difficult to accept the idea that a "most perfect" single spot does exist - and that this should become the center of the listening position?

To me it has always seemed that the whole explanation is upside down – it is rational to instead explain that there are some single points worth discussing – namely, that one should avoid 25 and 50% - for both the LP and the speakers – and there is a good area for proposing to place the LP. In short - stay away from here, and here, and instead gravitate to elsewhere, for example, this area here.

What I find with the vast majority of the people I deal with at the various forums is this:

They do not know - or care to come to know - all of the reasoning and rationale behind any of this - they do not want to try and understand where they shouldn't (in order to arrive by themselves at a place where they should (therefore) go - they only want to know where to begin - knowing that somewhere in that area they will find the truth of things.

The other flaw I find in the 38% rule, is that it addresses a very small set of the relevant and rational elements of room acoustics that play on the suitability of any given LP – but this is too long already, so I’ll just leave this comment to stand naked.

To put all of the above in my own personal context, using a 14’ deep example room and the trend of discussions I am all to familiar with - I have suggested that the user should slap LP chair down somewhere between 5 and 5.5' off the front wall - set up the rig rationally given fixed realities such as the size of the mix desk, etcetera - try hard to keep the speakers at least a few inches away from 42” off the front wall [25%], then start listening and/or measuring. If you have a significant problem that can be address by changing layout at all, you can probably find satisfaction by moving the speakers and/or the front/rear wall treatments. Change one thing at a time – be systematic. Find the best spot you can for smooth low frequency response. If not delighted with the results, move the LP and try again.

Again, I see no problems associated with this approach - and do not feel that it is either better or worse than the approach I use - only a different way to reach the same ultimate goal.

You say poe-tot-toh - I say Poe-tay-toh.......... you say toe-mot-toh - I say toe-may-toh............. they still taste the same do they not?

I would very much appreciate your thoughts on a single digit suggestion – or in lieu of that any advice you might be able to grant with regard to how I might improve making suggestions for seeking a range of “known good” starting positions.

Scott, I don't see a single thing wrong with your approach - and I don't believe that (within the context of this discussion) it needs you to improve on it.

Your advice is sound - it is (from my perspective) just that you appear to be rigid when it comes to seeing other equally good methods of finding a starting point. Remember - it is neither the points fault - nor is it Wes' fault - that this gained some sort of weird internet status - it was and still is - as viable a starting point as your 5' to 5'-5".

Why does it have to be a case of choose one best method. Some architects build their entire career with building designs based on Roman Architecture - other base it on the French Provincial Era......... which one is right and which is wrong?

Sincerely,

Rod

[JohnPM]

Sometimes bad can be good. For example, if you sit exactly half way down the length of the room you are unaffected by any of the odd order modes, so you no longer need to worry about the 1st length mode (which is too low for treatments to get hold of it in anything but a broom cupboard) and can ignore the 3rd, 5th etc as well. If you then plonk your speakers 1/4 length from the wall you will (almost) eliminate the 2nd mode and also the 6th, 10th etc. The first mode you then have problems with is the 4th, which if the room is not huge is high enough in frequency that your acoustic treatments can effectively deal with it, and there isn't another problem mode until the 8th so again easily handled by treatments.

[Bob]

I thought the 38% rule was a fairly simple thing: you add up the first 4 modes, and 38% is a minima, in the sense that it's far away from bad nulls you can't fix with EQ.
I remember looking at Harmon graphs of several modes, and it seemed about right.
e.g. Figure 18, pg 21, of http://www.acoustics.org/smallrooms.pdf
http://www.harman.com/xls/Room%20Mode%20Calculator.xls

Idea: If you have a very large room, then you throw away the lowest mode.

The limits on 38% are:
a) does the room have modes there, or are they different frequencies (the usual modal formula prediction problems)
b) do the virtical/width modes have too large an effect in that room
c) do the tangental oblique modes have too large an effect in that room

Would anyone put a subwoofer at 38% ? I can't recall it ever being recommended as a start point.

Basically what you're looking for is the best frequency response location.
www.harman.com/wp/pdf/multsubs.pdf pg 9

[timo]

Me

[bert stoltenborg]

godammit, words.
lots of it.

[Scott R. Foster]

3/8 rounded

Nice.. that sounds about right to me.


Rod:

What note is your kick tuned?

[JohnPM]

What drastic things does difference in humidity and temperature cause?

For the purposes of this discussion no effect at all, only the dimensions matter. You could fill the room with Helium and 38% would still be as good or as bad as a room full of air, only the frequencies of the modes would change. Would also be quite a giggle

[timo]

wow this is very interesting, If humidity and temperature can cause things. I'm wondering.... what ours emotions, body (very low) vibrations can do...?
I'm not kidding, seriously.

[jcgriggs23]

Not to muddy the waters even more, but didn't Stanley Lipschitz (sp?) suggest "roughly 1/3 of the long dimension" as a similar starting point for locating an optimal listening position in (roughly) rectangular rooms when he was at NRC. IIRC the paper in question was a survey of "professional listening spaces" done for the AES (or possibly JASA) and Dr. Lipschitz was the co-author - I don't recall who the other author was, but I think they were not from the NRC. I recall hearing this referenced several times in the '80's when discussing control room set ups (Dr. Lipschitz was a consultant in the design and construiction of a (locally) well known Ottawa studio - Marc Sound - in this time period.

Regards,
John