Jump to content

Room Acoustics - Large Room and Small Room


mas

Recommended Posts

{Please Note: Associated graphics in next frame!}

(In partial response to the Speaker Placement thread in the Architectural forum begun on 4/26/07 http://forums.klipsch.com/forums/1/906285/ShowThread.aspx

<?xml:namespace prefix = o />

Only two walls of the room have been sheet rocked, still have two walls left and the ceiling. No flooring has been laid (still concrete). So I should be taking measurements in the room even though the construction phase of the room is not complete?

First, you will please pardon me, when you mentioned the state of your room, I mistakenly thought that you were saying that because it was small and only partially had sheetrock that perhaps measurements didnt matter or were not useful! Oops! Pardon my misunderstanding...

A few thoughts and general comments some of which do not specifically apply in your specific situation

There still seems to be a general/common confusion among some folks between sound transmission through walls (and other flanking path vectors) with the behavior of the sound fields and component reflections within the room and the role of measurements. For the most part, they are two separate and distinct topics, despite the fact that they both involve walls. Aside from the potential for building some traps, etc. into wall surfaces, the easy way to think of this is that transmission from one area into another has to do with wall mass and construction techniques, while the rooms sonic character is a function solely of its surface (albeit conditioned by the previous qualification).

Techniques for minimizing sound transmission are actually rather well understood, despite our penchant for reinventing the wheel. I think what causes the most hassle for most is that such considerations are entered after a portion or a majority of the project is complete, rather than planning from inception which BTW is MUCH easier folks! ;-) And as such they are not the focus of this post.

Instead lets agree to focus on the in room response, characterized by the geometry and by the room surface.

Once the skin of the room is installed (as this is not a modeled room, meaning its behavior and response characteristics were not first modeled and anticipated in software) we can take preliminary measurements. The effect of embellishments such as carpet can be predicted relatively easily. This can be done in a very cursory manner by calculating the thickness of the pad and carpet calculated at a 45 degree incident angle. And then by allowing this distance to correlate to ¼ wavelength, determining the frequency to which this correlates. (Note: This will vary from the actual, as the two desperate densities and acoustic impedances of the carpet and the pad will not act as a single boundary. And the pad will also act as a reflective layerthus reinforcing the much more complex real world of non-linear surface behavior.)

Thus a ¾ inch carpet and pad would correlate to a 45 degree incident thickness of 1.31 inches which correlates to a ¼ wavelength of 10,308 Hz. Well above the frequencies where we will be worried about dense specular reflections. Thus you can see that most carpet and Spiderman beach towel tapestries, etc. will have little positive effect on the acoustics. To amplify and qualify this, assuming that the carpet is ¼ inch and the pad is ½ inch, the ¼ inch carpet would correlate to a frequency of 38,200 Hz. And the 1.2 inch pad to frequency of 19.097 Hz. So it should become pretty apparent that we are dealing with frequencies that are not only extremely easily controlled and damped, but we are also not dealing with frequencies that really need control.

All of this is to show that some of the more commonly accepted treatments will have effects, but they are not the ones that we really want to focus upon. They will tend to damp frequencies that we really do not need damped and ignore the lower frequencies containing much more energy that constitute more problematic specular (focused) reflections.

But lets take a step back before we get caught up in lots of minutia and lets look at the major divisions of approaching a small acoustic space:

Generally speaking, lets summarize the various groups of acoustic distortion:

Below 300Hz:

Problem 1: Room Modal Response

Solutions: Room Dimensions, Speaker Placement, Tuned Absorption (Corner Bass traps, Helmholtz resonators)

Problem 2: Speaker Boundary Interference;

Solutions: Speaker/Listener position, Tuned Absorption

Above 300 Hz:

Problem 1: Specular reflections (1st order early reflections)

Solutions:Surgically placed Absorption, Diffusion (Create an effectively anechoic Initial Signal Delay Gap)

Problem 2: Specular reflections (later arriving reflections)

Solutions: Diffusion, Selectively placed Absorption (rare) (Create a well behaved diffuse semi-reverberant sound field)

Derivative Problem 3: Comb filtering/Polar Anomalies (Superposition of direct and reflected signals)

Solutions: Signal alignment, Selectively placed Absorption, Diffusion,

Solets start by looking at room modes and standing waves

The room calculator programs for standing waves can give us a few ideas, allowing for their assumption of the room as a perfect space consisting of 6 parallel surfaces with no doorways, irregularities or dimensional variances. Not exactly any room that you or I can get into or out of! Especially as they do not allow for doorways! But its a good place to start!

From this, the average room will have doorways, alcoves, open passageways connecting adjoining rooms, windows, perhaps with inset frames, and a myriad other surface and structural irregularities. The net effect of this will be to modify the modal structure of the room.

The reason for this is pretty simple, and a basic understanding of tuned spaces is helpful here.

<?xml:namespace prefix = v /><!--[if !vml]--><!--[endif]-->

The useful information illustrated above is in the location of the nodes and antinodes, the peaks and the nulls of the standing pressure wave. Remember, a standing wave is a wave that basically reflects back and forth with the high and low pressure zones reinforcing itself at particular resonance frequencies. Beginning with a fundamental frequency and repeating at harmonic intervals (multiples of the fundamental frequency).

<!--[if !vml]--><!--[endif]-->

Please note. In a closed end of a resonant space as illustrated in the diagrams, there is a node, a high pressure zone. If the room is open, the open end ends in an antinode or a low pressure zone. If the room is closed on both ends, you have high pressure zones at both ends, with an antinode in the center (in the most basic form), or a series of distributed nodes and antinodes.

If a room is not a perfect rectangular space and exhibits the common asymmetrical anomalies mentioned briefly above, these well behaved patterns are modified. And in the more complex examples, alcoves, adjoining rooms, be they joined by large openings or simply doorways act variously as multiple resonant spaces, summing in very complex ways.

This phenomena is referred to as a coupled space, and its behavior is anything but simple. Oh, some may be simple, but most become very complex very quickly! The significance of this fact being not so much that they are complex, but to reinforce the understanding that predicting them from simple ideal calculators is almost impossible. And thus the easiest way to address them is by the actual measurement of them. Then the focus is not so much oh how they occur, as it is on simply what is occurring!

Oh, and while we are here, I will also mention that the modal response is not one dimensional. We are dealing with a 3Space space, and as such, we have 3 primary modes. Room modes consist of three different types of resonances; these are known as axial, tangential & oblique modes. Axial modes consist of waves resonating only along one dimension: the length, width or height of the room, Tangential modes involve two dimensions, the length & width, length & height, or width & height. Oblique modes involve all three dimensions in each mode of resonance. Normally the axial modes have the most strength while the oblique modes have the lowest strength.

The concept of dimensioning a room is to attempt to more evenly distribute the frequency centers of the various modes in an attempt to minimize their summation and hence greater peaks and lower nulls, resulting in a more evenly distributed series of peaks ands nulls exhibiting lower peak and higher null SPL levels.

A very simple example of an axial mode would be that formed along the length an example room: L=25ft, W=16ft, H=8ft. The first axial mode along the room length would be: SpeedOfSound/(2*RoomLength) = 1130 ft/sec/(25.0 ft*2) = 22.15 Hz. Additional modes would exist at integral multiples of 22.15 Hz : 44.3 Hz, 66.45 Hz, etc. The width dimension gives a first axial mode at 1130/(16*2) = 35.31 Hz, 70.62 Hz, etc. The height dimension gives a first axial mode at 1130/(8*2) = 70.62 Hz, 141.25 Hz, etc.

You will quickly note that just for this axial mode, that the frequency centered around 70 Hz will be a very popular frequency for many of us! And common multiples of any dimension will tend to create additional nodes at common frequencies that will sum! And just as in a Dickens novel, these coincidences portend future issues that will require addressing.

What all of this means is two things. One, there will be zones in the room where the low frequencies will tend to be much more pronounced than at others. And two, the bass will tend to overdrive the small space and render it boomy and not very distinct and tight. And this quality is the primary problem encountered in small acoustic spaces. Higher order reflections, etc tend to get more attention, but room treatment MUST begin with the fundamental treatment of room modes and Lf energy.

<!--[if !vml]--><!--[endif]--> <!--[if !vml]--><!--[endif]-->

<!--[if !vml]--><!--[endif]-->

So, for most rooms that exhibit closed ends, there will be a node at the rear corners. This is further reinforced by the corner placement of speakers. Additional nodes may be spaces at intervals in the room, usually ½ and 1/3 the distance of each lateral wall. And for the additional modes, at the intersections of each of the lateral surfaces walls, floors and ceiling in all of the corners.

We can address additional room boundary concerns which affect alternative speaker placement at a later date, but personally I try to avoid them if at all possible either via corner placement or via acoustically isolated soffet (in the wall) installation...

Generally speaking, corner traps consisting of any of a variety of tuned absorptive traps are effective. These are most commonly of a tuned tube or pipe, or a modified version thereof, or in the form of panels. The fundamental physics model which is employed is that of the Helmholtz resonator. This is a tuned enclosure that resonates at a particular frequency. A very common example that most have experienced is that of blowing across the top of a coke bottle or a jug. Likewise the use of a tuned pipe to create a sound in a traditional pipe organ. The varying volume of the enclosed space will result in a particular resonant frequency. A trait that is most helpful when tuning a jug for use in the weekend jug band festivals that are fast approaching!

And if we refer back to our knowledge of impedance, a properly terminated source with a load of identical impedance will result in the maximum power transfer, and all of the incident energy will be absorbed and none will be reflected and returned back into the system. And since a room exhibits an acoustic impedance, and each surface likewise exhibits an acoustical impedance, they will absorb and reflect incident energy. And since nothing is ever simple, each surface exhibits a complex impedance, resulting in a non-linear absorption of some frequencies, and a reflection of other frequencies. So each wall surface will absorb or reflect energy in varying ways, and with varying frequency spectrums, in effect EQing the reflected sound. (And this is a critical consideration, as many will erroneously believe that an absorber simply absorbs! NOT TRUE!)

The Helmholtz absorber is a precisely tuned enclosure, meaning that it tend to be a high Q device, efficiently absorbing a very narrow band of frequencies and tending to ignore others (although lower Q, broader absorptive band designs are possible, they are correspondingly less efficient over the frequency band).

Again, many variations in the design of Helmholtz enclosures

are possible, from tuned pipes, to panels exhibiting various resonant/flexure, porous, or slotted designs.The design of each of these is beyond the scope of this diatribe. But they are very precise, and able to be tuned to quite a variety of frequencies. Often the most critical aspect is not the design, but of obtaining materials with known qualities (density, porosity, mass, etc) allowing the construction of a trap that correlates closely with the design!

So, to make a very general recommendation In a room that is generally closed, LF traps in all of the corners consisting of a combination of corner and lateral flanking tuned LF absorptive traps is almost always advantageous. The trick is to know what frequencies are problematic and then responding by designing and building traps sufficient to accurately and efficiently trap them. Additionally, room furnishing can have an impact, both good and bad on the room response as well. So as things become more complex and more variables are introduced, its nice to know what is actually happening. Thus measurements knowing exactly what you are dealing with comes in much more handy then a calculator that provides a very pretty graph and a guess.

Oh, and since many are using Heritage speakers, this means LF traps placed above the corner speakers in the front of the room as well!

Oh, and while we are here, and since there is so much more to any of these topics, at the risk of introducing as many or more problems as the technique actually solves in the complete overview of the topic, I might note that variable subwoofer placement offers a mixture of both good and bad options! The good is that dependent upon its placements about the room, the subwoofer will tend to shift the primacy of the modes from axial to tangential to oblique. With this shift, different modes will be reinforced differently, thus offering the potential to reinforce different modes at differing fundamental and subsequent harmonic frequencies and reinforcing various nodes and antinodes and their placement according to the dynamics of the particular modes. Additionally, additional units can be placed in positions so that certain modal peaks and nulls are cancelled via 180 degree out of phase superposition. So, varying the position of the subwoofer can potentially be used to your advantage. This is the focus of the oft cited Harmon whitepaper on subwoofer placement.

But there is a cost! Despite the mantra that LFs are omnidirectional, the notion that LFs are non-localizable is optimistic at best, and simply incorrect at worst. Oh sure, at sufficient distances this becomes more valid; witness the attempt to localize a approaching low flying helicopter. But in a small acoustical space, where the Henry precedence effect is still valid, the difference in arrival times at each ear still allows quite a bit of localizable cues, despite the frequencies at hand. Additionally, the lack of signal alignment in the time domain also causes radical signal offsets (significant and perverse group delay anomalies) and crossover anomalies manifested via superposition, ALL of which are audible effects that ideally should be avoided. But, I mention this as you do have a variety of variables at your disposal, but it must be remembered that these have negative ramifications as well as positive ones.

And generally speaking, aligning signals in the time domain minimizes frequency response and intelligibility anomalies and should be your first goal. After all, once signals are aligned within the time domain in a minimum phase relationship, many well behaved options, including EQ, are available for adjusting the character of the direct signal.

There is much more than can be addressed, but I prefer to avoid going into too much arcane detail if it is not productive. So lets start with this and see where the needs are.

Oh, and let me make a request. At this point we are still talking about the basic concept. And my goal is to try to present an opportunity for folks to understand the basic issue at hand. I am not going to go off on lots of tangents here chasing lots of how do I solve thi or that at this point. After we get the basic concept down, there will then be plenty of opportunity to pursue specific solutions. And hopefully by then, most will have a basic idea of how to begin to go about doing this.

Besides, I will have my hand full trying to avoid getting off on tangents myself. Trust me, I dont need any help finding tangents to chase! So help me stay focused and I will try to help youwith lots of help from some very able folks here!

Focus Mark, please focus.

Link to comment
Share on other sites

  • 3 weeks later...
  • Replies 118
  • Created
  • Last Reply

Top Posters In This Topic

Top Posters In This Topic

Posted Images

Mark...

When I start working in my downstairs room (where my "stuff" is going to end up), are you going to come on up here & give me some pointers & bring your measuring stuff?

I'll even put you up for the weekend and most special of all... let you take Chloe for her evening walk (sorry though, she's still gonna sleep with ME! [:P])

I'm not going to ask the significance of the two martians in your avatar... [:^)]

Link to comment
Share on other sites

Mark...

When I start working in my downstairs room (where my "stuff" is going to end up), are you going to come on up here & give me some pointers & bring your measuring stuff?

I'll even put you up for the weekend and most special of all... let you take Chloe for her evening walk (sorry though, she's still gonna sleep with ME! [:P])

I'm not going to ask the significance of the two martians in your avatar... [:^)]

Is Chloe your dog or your wife?

Link to comment
Share on other sites

  • 4 weeks later...

Ok, it's time to start in with some serious study. I'm going to order the basic text and start in on this process. Hopefully Doc can help me along with some of the formulae. ( Perhaps it's not that important to understand all the math, just the main precepts).

Any idea where to get this at a discounted price?

Michael

Link to comment
Share on other sites

Ok, it's time to start in with some serious study. I'm going to order the basic text and start in on this process. Hopefully Doc can help me along with some of the formulae. ( Perhaps it's not that important to understand all the math, just the main precepts).

Any idea where to get this at a discounted price?

Michael

Michael,

I assume you are talking about the Davis & Patronis book. I picked up a copy from Amazon for $63 and shipping was free. The math presentation is a bit funny, some fairly basic concepts are discussed at an introductory level, while some (not all) of the advanced stuff is thrown at you. Another problem is that many of the references are not readily available at University or public library. In fact many of the references are talks given at "their" Syn-Aud seminars. I suspect these are not freely available (this is a pet-peeve of mine) but I might be wrong. Don't even get me started on the number of graphs that are presented that do not even have their axes labelled.

Other than that ....

Good Luck,

-Tom

Link to comment
Share on other sites

Michael, I bought a copy a few months ago and have had time to
look it over. Lots of great info, but maybe a little to
deep for me. I can't see me needing this book for any of my projects in
the near future so I am offering to send it to you for you to keep as
long as you wish. Maybe Doc would like to see it as well.
If you are interested, just say the word.

Garth

Link to comment
Share on other sites

  • 2 weeks later...
Sound System Engineering, 3rd Ed.; Don Davis, Eugene Patronis ISBN: 9780240808307


Price is $57.27, but the Amazon price is quite reasonable.


Please do not get the 1st or 2nd editions. SUBSTANTIAL updates have been incorporated.



I'm a little confused by some of the above comments...so I will attempt to address them with a bit of wry humor.


Which axes are not defined? Note: most graphs are screen shots from instruments. And most of the measurements have axes that are defined by the nature of the measurement or derived quantities. If you have any problem with this, PM me and I can help.


As far as the math being "thrown at you", it is not a book that is intended to hand hold you through basic math. Nor is it a book that is totally reliant upon math. In fact, for an engineering text, it is one of the most NON-math intensive books that you can find. Chapter One includes 30 pages that are dedicated to explaining the basic non-calculus math tools. Unfortunately, math is the language of physics and engineering ...they are funny that way. ...And neither do Porsche GT3s come with basic drivers ed guides; it is assumed that you can drive ;-) So while an understanding of the basic math will add greatly to your understanding of the concepts, math is not necessarily required to understand the major concepts. Chapter One includes 30 pages that are dedicated to explaining the basic non-calculus math tools.


...And all references are not available for free? You mean like AES, ASA, IEEE, APS and other references for which you must pay $20+ per article for non-members? Libraries and individuals can certainly subscribe to Syn-Aud-Con. No, its not free, but neither are the expenses necessary to run the association, publications, and directed research, nor the seminars. But the information has been freely released to the public domain, UNLIKE many other associations. And they are hardly to be held responsible for texts and documents that are out of print. SynAudCon Tech Notes are technical articles distributed with the SAC newsletter. They are not Workshop notes. Workshops were typically accompanied by a notebook of technical addendum. For the major journals and Tech Notes I would suggest contacting a library which participates in an academic inter-library loan program as many libraries do have them. Despite this, most referenced AES and ASA papers (of all types) are not available via the journals and must be ordered from the online libraries at a cost of $5 for members and $20 for non-members.


If anyone would like to see examples of Dr. Patronis' presentations at a SynAudCon function, I can dig a few out for your perusal. They do reflect his 50+ years experience as a Georgia Tech Physics professor! Also, the source contributions by Dr. Peter D'Antonio (based upon Dr. Manfred Schroeder's research in mathematics) and Don Keele and Dick Heyser will terrify you as well (especially as your realize their primary backgrounds from which the audio applications came as a result of a secondary derivative interest) Trust me, you are getting off easy in the text! (And you will also begin to appreciate why the various online 'calculators' are so strikingly simplistic and limited.) Materials were presented at the level necessary to convey the complexity of the topic, despite the workshops being attended by persons of varying backgrounds.


Oh, and have you priced a recent Vienna AES seminar (and sampled the even more 'off the wall', at best, list of presentations)? Historically, the SAC seminars were oriented 50/50 towards both applied research and instruction. Current SAC seminars are much more instructionally oriented, with occasional focused applied workshops. The Syn-Aud-Con seminars and workshops are much more reasonable in cost than a new pair of Heritage speakers! In fact they are, and were, cheaper than what many here find quite reasonable for a CD player or a fancy set of designer interconnects or a 'required' passive crossover! But they are nevertheless like taking a sip from a fire hydrant. ;-)


All in all, a trip to Brazil doesn't absolutely require proficiency in Portuguese, but a basic understanding of the language does make the experience much more rewarding. But you can still get a round and appreciate quite a bit...And I not sure that this is due to a shortcoming of Brazil. ;-)


So, may I suggest that you not let the math scare you. The concepts can be understood sufficiently without the math. Of course I will leave you to decipher the DSP and crossover sections as you wade through the math, as that is simply the language of the territory, but the other areas are understandable with a good grasp of algebra.


;-)


Note: The primary chart/measurement in chapeters 7,8 & 9 are ETCs (Envelope Time Curves). These are strictly defined. They are also presented and defined prior to their use. If one is looking at them after their initial presentation and does not recognize them or their use, one needs to go back and find out what they are looking at. Useful interpretation is based upon info presented in the body of the measurement (e.g.: dB/division). Their generic value is established in comparison with the other ETCs and the basic paradigm for a desired room or speaker response.. All of the other details are superfluous in these cases. That is why the scales are intentionally omitted!

Oh, and to put a subscription to the SynAudCon newsletter into perspective, just ask me what benefits 30 years of membership in AES, ASA, IEEE have provided. And they don't even provide a membership card whose lettering doesn't smudge and migrate to the wallet sleeve! I have walls of journals (all except for the totally absurd waste of paper from IEEE, with the sole exception being the Fall 2003 step by step instruction guide on how to condtruct an impulse weapon from a microwave oven Klystron!) that have cost me over $9000 and I STILL have to buy the listed articles as the VAST majority are not in the Journals!


Link to comment
Share on other sites

MAS, I am not going to get into an argument about this, but I think you misconstrued some of my comments. Having written articles and chapters for a technical audience, I am very picky about making sure that arguments are developed and that references are provided etc.

Specifically:

A university library will have JAES and JASA, they will most likely not have proceedings from a SynAudCon workshop.

Have a look at chpts 7, 8 & 9, there are many examples of units, scales and axes not being labelled. With careful reading the specifics can be figured out, but the reader should not have to work that hard (this is a personal pet peeve).

Regarding the math: choosing an appropriate level is always a challenge, especially when the different readers will have such different backgrounds, However, the contrast between describing decibels in chapter 2 and the presentation of concepts in Fig 6.5 - 6.7 is staggering. A mathematically sophisticated reader who might want to follow up on this would be at a loss (certainly an appendix would have helped).

My goal is not to get into an argument. The book covers a number of good topics. Any book would have difficulty in tackling this range of topics, especially for a diverse audience. I will grant you that. I actually think the book is pretty good, but it does have some problems. These problems could have been resolved (at least in part) with some better editing etc.

However, this may give us a common ground and a common vocabulary to start our group-conversation (even though much of the math is asserted rather than explained).

So if other interested folks can get the book, maybe we can roll up our collective sleeves and get started.

Good Luck,

-Tom

Link to comment
Share on other sites

Michael, I bought a copy a few months ago and have had time to

look it over. Lots of great info, but maybe a little to

deep for me. I can't see me needing this book for any of my projects in

the near future so I am offering to send it to you for you to keep as

long as you wish. Maybe Doc would like to see it as well.

If you are interested, just say the word.

Garth

Garth- I got the book today! THANK YOU! WOW, that's a great gift from a Forum brother I have never even met! Very Cool! [Y]

It all looks terribly complicated, I think I'll have to skip the formulae the first go around or just skip to the most applicable chapters. It certainly is... well... thorough!

Michael

Link to comment
Share on other sites

Michael, glad you got the book. Hope it helps you in
your studies. I must admit that although I found about half the
book to be very interesting and fairly easy to understand, the
rest of the book is probably more suited to someone with and
engineering or physics background.

Enjoy and I hope you can "pay it forward" someday.

Garth

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


×
×
  • Create New...