Room dimensions

Tried to find some info on room dimensions here, but failed. I hope I don't post questions that have been beaten to death already. Here goes: 1) My control room will be diamond shaped. That is, the side walls are splayed outwards in the front 2/3rds of the room, and then splayed inwards in the back 1/3rd. When looking on optimum room dimensions, they are always calculated for rectangular rooms. My question is - can I assume (or at least sorta assume) that the average width of my room more or less corresponds with the widht of a rectangular room? Can I expect a similar spread of room modes? 2) If so, I'm in luck. My room length is 4,7 m and the average width is 3,9. Now if I made my ceiling 3,38 m tall I'd get very close to the classic 1 - 1:14 - 1:39, which I understand is Eric Desart's favourite dimension. The only thing is, I had planned to make the ceiling 3 m tall, as that would give me a space for storage of stuff on top of the CR roof (the existing ceiling of my location is about 4 m tall, so I would get about 80 cms of storage height). My location is small and I have two drum kits, so storage will be a bit of an issue for me. But nevertheless - do you think I should change my plans and make the ceiling 3,38 m? I could use the space between the CR roof and the existing ceiling for a huge bass trap for the live room. Your advice is much appreciated! Best, Henrik

C'mon - nobody?!

Most rooms when the acoustic treatment goes in are what you have described, however with your room being shaped as such (the main wall construction) how much bass trapping will you be able to acommodate. take a look at the room design I used. Hope this helps. The angles are actually the treatments. the outer rectangle is the initial room. the spaces between the treatments and the main walls is used to trap bass. :)

Using a 3m ceiling height gives you a "hole" around the 50 hZ range - going back to the 3.38m ceiling lessens (but doesn't eliminate) the 50 hZ hole; your length and width cause an axial mode coincidence at 220 hZ. This would be minimised by the fact that your side wall dimension is an AVERAGE of the splayed walls. Having nulls in a low range like that can cause you to try to compensate by boosting the sub - this can be a real "can of worms"; placing a speaker in a null and then trying to compensate by cranking it up is just asking for headaches - blown woofers, overheated amps, and STILL not an accurate sound field... Hope that helped... Steve

A recent paper published in the Journal of the Audio Engineering Society described a fairly sophisticated method for modeling room modal response. Room Sizing and Optimization at Low Frequencies; Trevor J. Cox, Peter D'Antonio, and Mark R. Avis; JAES, Volume 52 Number 6, June 2004 They found that there is not simply one optimal set of dimension ratios. Rather there are several ranges of good solutions. Furthermore the optimal ratios depend on the room volume. Using their published ratio maps and the nominal dimensions you gave, I came up with a few good ratios. Some may not be practical. The second in the list is probably the best, since it is centered on the largest local minimum of the optimization map.

0.71  1.00  1.12
1.00  1.55  2.60 
1.00  1.96  3.00
0.56  1.00  3.30

The next question is whether these ratios apply at all to your situation. You have splayed walls. These ratios are for a rectangular room. And the modal response isn't simply derived from the average distance between the splayed walls. So, I don't know? Thomas

They found that there is not simply one optimal set of dimension ratios. Rather there are several ranges of good solutions. Furthermore the optimal ratios depend on the room volume.

Thomas: How do the results compare with Walker's paper (BBC RD 1993-08)? The first two sentences I quoted could be right from Walker's conclusions. The last sentence (room volume) is new to me. Thanks Andre

"And the modal response isn't simply derived from the average distance between the splayed walls." - Thomas, I'm not sure if this was aimed at my comments or not - if so, that's not what I said or meant - I only meant that the coincidence at 220 hZ would be a minor thing, since the walls are splayed and that was calculated using a parallel wall scenario. I've yet to see a reasonable way to calculate non-parallel walls, especially for math-challenged peeps like me - the average thing is admittedly a (weak) band-aid... Steve

Steve, > the coincidence at 220 hZ would be a minor thing, since the walls are splayed < I'm glad you brought this up, because a mode that high is probably not very important anyway. In that room those 220 Hz modes are the fourth and fifth modes for length and width respectively. Once you get that far into the series the mode "harmonics" are not contributing nearly as much as the first few in the series. What a lot of people overlook is that a room is simply a bandpass filter with 6 dB per octave slopes on either side of the center frequency. (Or, more correctly, three such bandpass filters in parallel, with one for each dimension.) So the farther up the series you go, the less resonance is being contributed. --Ethan

What a lot of people overlook is that a room is simply a bandpass filter with 6 dB per octave slopes on either side of the center frequency. (Or, more correctly, three such bandpass filters in parallel, with one for each dimension.) So the farther up the series you go, the less resonance is being contributed.

I have never read or seen anything presenting that point. Can you provide some references so I can read up on it? Thanks

AVare wrote:

They found that there is not simply one optimal set of dimension ratios. Rather there are several ranges of good solutions. Furthermore the optimal ratios depend on the room volume.

Thomas: How do the results compare with Walker's paper (BBC RD 1993-08)? The first two sentences I quoted could be right from Walker's conclusions. The last sentence (room volume) is new to me. Thanks Andre

http://www.acoustics.salford.ac.uk/acou ... izing3.htm Check Excel Files too. This is based on the Room Sizer calculator results. And indeed it rizes some questions. I'm still busy with it. Warm Regards Eric

Andre, > I have never read or seen anything presenting that point. Can you provide some references so I can read up on it? < Sure. This basic property of rooms was explained to me a while ago by an engineer friend of mine, Bill Eppler, who's been my mentor for many years. I'll see if he has any "real" references in the form of textbooks, but in the mean time, a quick Google search for "room acoustic filter -comb" (-comb excludes the numerous references to "comb filter") turned up these links: http://pcfarina.eng.unipr.it/Aurora/SAW/RoomSim.html http://www.techonline.com/community/ed_ ... icle/20119 http://www.far-audio.com/panels.html Even aside from links and references, if you're versed at all in electronics it's not much of a leap to see why a room is in fact a bandpass filter. Like an electronic filter there's a natural resonance caused by the exact same phase shift you get from one inductor and one capacitor. That's why the room's filter has a 6 dB per octave slope up and down - where one slope corresponds to each reactive component. I'll call my friend Bill tomorrow and see if he has anything else to add. --Ethan

Thanks Ethan. Looking back my quote was misleading. What I was referring to was the last sentence.

So the farther up the series you go, the less resonance is being contributed.

BTW, I have worked as an audio electronics engineer designing filters, amplifiers, etc. the electronic analogies are quite familiar to me.

Andre, > So the farther up the series you go, the less resonance is being contributed. < In my posts elsewhere "Do room modes even matter" and "Room modes part deux" I showed the results of ETF measurements at varying distances from the rear wall in my living room and my garage, respectively. At the lowest frequencies the room modes clearly dominate the response, contributing less and less at higher frequencies where simple boundary interference takes over. This should be easy to measure using an impulse test, to see what frequencies linger and at what relative amplitude. --Ethan

Thanks Ethan: To paraphrase then "By the time you move up to 4th and 5th harmonics, of the axial modes, the non-axial modes and SBIR have a much more significant effect than the room modes." Is that close? Is this a simplification of concepts like the Schroeder frequency?

Thomas,

barefoot wrote:

Room Sizing and Optimization at Low Frequencies; Trevor J. Cox, Peter D'Antonio, and Mark R. Avis; JAES, Volume 52 Number 6, June 2004 They found that there is not simply one optimal set of dimension ratios. Rather there are several ranges of good solutions. Furthermore the optimal ratios depend on the room volume.

Actually... That paper re-presented concepts that have been understood for a long time. The mappings provided were in the same style as those presented by Milner and Bernhard in their paper: An investigation of the modal characteristics of nonrectangular reverberation rooms, JASA 85 (2), Feb. 1989. And their work, in turn, was based on previous algorithms; primarily those of Louden. And Louden, in turn, was based on Bolt (ca. 1938). What is often omitted in the presentation of Bolt's famous ratio graph is the supplemental graph that gives the volume dependence of the "ideal" ratio bubble. (Don't you just love when something "new" is "discovered" that's actually been on the books for over 60 years!!!??? :lol: ) Best regards, Jeff D. Szymanski Chief Acoustical Engineer Auralex Acoustics, Inc.

Andre, > To paraphrase then "By the time you move up to 4th and 5th harmonics, of the axial modes, the non-axial modes and SBIR have a much more significant effect than the room modes." Is that close? < It seems to me that non-axial modes would diminish in significance just as quickly. But I admit I've never measured that specifically. I'm not convinced non-axial modes really account for much anyway, outside of a theoretical completely empty room having very rigid boundaries. Clearly, boundary interference is at least as important as anything else, whether you consider the distance between the speaker and the boundary or the listener and the boundary. In fact, I've argued that boundary interference is the fundamental cause of peaks and nulls in small rooms, and modes are just a subset. The following is from the Room Modes article on my company's web site:

Acoustic interference occurs outdoors against a single boundary, and so doesn't even require a room. As you close off one wall, then two, and so forth until the room is completely enclosed, the interference patterns simply become stronger and more complex due to interaction, reinforcement, and resonance within the enclosed space. But simple boundary interference is the basis - the parent property, if you will - and the additional subset properties of room modes develop when the space becomes enclosed.

> Is this a simplification of concepts like the Schroeder frequency? < I don't know what Shroeder's frequency is so I can't comment on that. --Ethan

Andre,

AVare wrote:

To paraphrase then "By the time you move up to 4th and 5th harmonics, of the axial modes, the non-axial modes and SBIR have a much more significant effect than the room modes." Is that close? Is this a simplification of concepts like the Schroeder frequency?

I hope you don't mind my jumping in here. :) By time you move up to the 4th and 5th order modes in a room, the modal density will have increased considerably. Likewise, the average modal spacing will have decreased significantly. This usually leads to an audible "evenness" in an untreated space. As the modes are damped by treatments in the room, some persnickety frequencies (often axials, sometimes tangentials, rarely obliques) could cause problems. It simply depends on what frequency range is being considered versus what frequency range is being damped and where in the room the damping is placed. E.g., if you have treatments that damp the modes in a room down to 150 Hz, but there's a 4th order axial at 100 Hz, then it might be a much more noticeable problem after treatments. SBIR is fully explained in myriad references. Suffice to say, at any given listening position in an otherwise empty room, you will have 6 "first order" SBIR frequencies, 6² "second" order, and so forth. The first order frequencies are usually the most troublesome. In fact, it is often only the bounces from the front and back wall that are even noticeable. And their effect in a typical listening room can usually be sufficiently damped with good broadband treatments on the front and back walls. Finally, I am not particularly fond of the Schroeder frequency for small rooms, since Schroeder developed the formula for this frequency with large rooms in mind. A better frequency to use, IMO, is the "Davis" frequency (from Sound System Engineering, by Davis & Davis): 3*c/Lmin where c = speed of sound in air and Lmin = smallest room dimension. It is worth noting that: 1. This is equivalent to the highest 6th order axial mode. 2. A comment that accompanies this frequency in the Davis text is "Do not view it as a rigid fixed frequency." Best regards, Jeff D. Szymanski Chief Acoustical Engineer Auralex Acoustics, Inc.

Finally, I am not particularly fond of the Schroeder frequency for small rooms, since Schroeder developed the formula for this frequency with large rooms in mind. A better frequency to use, IMO, is the "Davis" frequency (from Sound System Engineering, by Davis & Davis): 3*c/Lmin

It is interesting to note how certain practices become formalized over time. In Acoustics and Noise Control , Rettinger calculates modes up to the fifth harmonic. Just a quick in case takes that the wrong way. Davis and Davis helped lead the world of electronics and acoustics from one one plateau to the next one up in changing it from an art to a science. Andre[/i]

Andre,

AVare wrote:

Davis and Davis helped lead the world of electronics and acoustics...

One of my favorite quotes:

Don Davis wrote:

In audio and acoustics the fundamentals are not difficult; the physics are.

Best regards, Jeff D. Szymanski Chief Acoustical Engineer Auralex Acoustics, Inc.

lovecow wrote:

A better frequency to use, IMO, is the "Davis" frequency (from Sound System Engineering, by Davis & Davis): 3*c/Lmin where c = speed of sound in air and Lmin = smallest room dimension.

Digging up... Is this metric or imperial ? Thanks !

msikio wrote:

lovecow wrote:

A better frequency to use, IMO, is the "Davis" frequency (from Sound System Engineering, by Davis & Davis): 3*c/Lmin where c = speed of sound in air and Lmin = smallest room dimension.

Digging up... Is this metric or imperial ?

Yes. Use either consistently. Speed of sound in m/s and Lmin in meters: or, Speed of sound in ft/s and Lmin in feet. Andre

lovecow wrote:

And Louden, in turn, was based on Bolt (ca. 1938). What is often omitted in the presentation of Bolt's famous ratio graph is the supplemental graph that gives the volume dependence of the "ideal" ratio bubble. (Don't you just love when something "new" is "discovered" that's actually been on the books for over 60 years!!!??? :lol: ) Best regards, Jeff D. Szymanski Chief Acoustical Engineer Auralex Acoustics, Inc.

I'm going to try to find a way to fit this into my website. I have always been engaged by the golden ratio, the room ratio and the ratios presented or discussed by all you guys that have real knowledge of that. Thanks for your time Jeff, even if it is from such a long ago post. Brien