Soundman2020 - Studio Design Forum

Hey everyone, i thought it would be great to have a thread concerning methods used to predict acoustical effects and results. There are various facets of studio buildings, isolation, frequency response, wiring, ergonomics, sightlines, ect ect. When faced with a blank slate, or existing conditions, acoustic engineers and studio builders are often called on to make things sound "good, quiet, and/better".

One question ive found difficult to answer is exactly how good, how much better, is the nd result going to be? Isolation, is somewhat more straightforward (to me at least) to give a good estimate of, though technical knowledge and attention to construction detail is necessary, the mass law, budget, and existing conditions or exptations, can get go a long way.

Frequency response seems far more evasive. In a different thread on this site Stuart made a response curve (under 300hz) for a room he was involved with. My question is how did he arrive at that curve. Identifying the modes is straightforward with a mode calc. If i reacall correctly modal spacing is ideal between 5-20%, so spacing doesnt seem super complex, at least on the basic level. So my first question regarding modal response prediction is:

How was the relative levels of the modes predicted?


-----

Another question is how do you predict the effect of a bass trap? i.e- i add X amount of cubic ft. of typical fluffy fiberglass to a corner, how much reduction of Y frequency should i expect at the listening postion.

Ditto for helmholtz resonators, tuning them to a problem mode is fairly simple via a calculator. But what happens when your looking for a certain amount of decibels in reduction at the listening position? How do i know how big to make the resonator?

-------

Quickly fwiw about where im coming from with this thread. Ive been into playing and recording music since the late 90s, and became a pro engineer (live/studio) in the mid 00s. Around that time i was asked to build my cousin a home studio, and did so after reading Rod Gervias' book a couple times. I have experience working in contracting/construction as a child hobby, and during my high scool/college years.

This first build spawned an obsession in acoustics and studio construction, and i proceeded to read the usual books by Philip Newell, F. Alton Everest, Marshall Long, the usual forums GS, JS, and R.O, as well as watch hundreds of physics and acoustics classes on youtube from Stanford, the US Army, and MIT. Ive watched interveiws with acousticians. I've also built 2 sucessful commercial studios, a few project studios and i high end home theater. Ive also assisted online with hundreds.

My point if that babble was to establish my context, and illustrate that a basic understanding of the fundemental principles, following acoustic assembly plans precisely, and a bit of experience has gone surprisingly far. Ive found about equal demand for my studio construction skills (meager as they may be), as the audio engineering side.

Ive gotten great advice from people online and phone like Rod and Andre Avare, and others when i got stuck.

But still..... Its been difficult to develop a method of prediction. No book has outlined a method ie a build it like the pros, but for acoustics, and the web is peicemeal. This is excaclty what i would like this thread to do. This way me, and others can give reasonably accurate predictions for our designs and improvements, and have our own and client/friends expectations and budgets aligned with reality.

This would be a welcome addition to my skill set, as opposed to the old "as much bass trapping as you can fit", "a little trial and error", "as much mass as you can afford", ect, ect. Basically the transition from a journeyman to black belt has been allususive so far, largely due to inability to predict certain results..

Anyway i apollogize for rambling on and boring the world with personal details. Its akin to a mucisian playing an E chord, and knowing that they are playing an E chord, then knowing when an E chord could sound right at a certian point in the riff. Or knowing how to play instruments seperately, but then applying that to writing a song.

Thanks, i hope this grows into a valuable experience for everyone, and a good resource.

-Kyle

Hey everyone, i thought it would be great to have a thread concerning methods used to predict acoustical effects and results.

Excellent idea! I love it!

One question ive found difficult to answer is exactly how good, how much better, is the nd result going to be?

One thing I learned early in technical design class, way back when (don't ask how many years ago... I prefer to not remember! ), is the importance of "design with the end outcome in mind". The instructor was always harping on on that! It seems logical and obvious when you think about it, but lots of people don't seem to get it! Just look around at some of the crazy designs you see for commercial products all over the place, and its clear that some designers just don't think things through very well: even professional product designers.

So, that's a round-about-way of answering your question: Any decisions you take in the initial layout and design of your room can have consequences on the final acoustic outcome. The can also have an impact on the operation of the studio, which can be another big issue, especially for commercial studios. For example: doors that open into furniture, equipment or instruments; convoluted traffic paths between rooms; poor sight lines; insufficient storage; etc.

For acoustics too: taking the final desired outcome into account when choosing wall locations, ceiling heights, shapes, etc. can have an effect, so it's important to take into account... to a certain extent. If you are designing a studio from the ground up, with just an empty piece of ground to start with, this can be very important. But most home studio designers do not have that luxury, and have to live with existing rooms, with fixed shape and size. Even then, there might be things that can be done to improve the acoustics....

Basic rules:

- Make the room as big as possible! Sound lives space, and the more space you can give it, the better. So if you have a built-in closet at one end of your room, then take it out to gain that space, for example. Or if you have a choice between two rooms, take the bigger one.

- Make the room as high as possible! High ceilings are great. The higher the better. So if you have a drop ceiling in your room, take it out and get all the space above you. (caveats: drop ceilings often hide HVAC ducts, electrical wiring, plumbing... )

- Stay away from cube/square/curved rooms, if possible. Also stay away from long, thin rooms. There's a range of "good" ratios, and a whole bunch of bad ones: try to avoid the bad ones, and get close to the good ones. That said, there's no need to go crazy on your room ratio, thinking that if only you can get the "perfect" ratio, your room will sound just like Abbey Road! It won't. There is no perfect ratio. There's a series of a few dozen good ratios, but here too you don't need to nail one of those perfectly: There's way too much noise and hype on the internet about room ratios, with some advocating for this ratio or that one, and claiming that perfectly following some strange scheme will guarantees perfect sound: it wont. Just stay away from the bad ones, get close to one of the good ones, and that's it.

- Make the room rectangular. Why? Because you get the largest floor area and greatest room volume like that! Angling walls reduces the siz of the room, and rule #1 says you should aim for the largest size possible.

I can probably add more "rules" to that later, but those are the basics.

Frequency response seems far more evasive

Frequency response is a direct consequence of room dimensions. Many people don't realize that it is the modal response of the room at the listening position, that defines the frequency response curve. It really is that simple. Even if the modes are well damped, it's still the modal response that defines the FR curve for any given location in the room. A well-damped room will have a generally flatter response, but the peaks and dips will still coincide with the modal response (along with other things).

In a different thread on this site Stuart made a response curve (under 300hz) for a room he was involved with. My question is how did he arrive at that curve. Identifying the modes is straightforward with a mode calc. If i reacall correctly modal spacing is ideal between 5-20%, so spacing doesnt seem super complex, at least on the basic level. So my first question regarding modal response prediction is:

How was the relative levels of the modes predicted?

As I hinted at above, that depends on your location in the room. Another common misconception is that the room has one single frequency response everywhere. Not so. The frequency response that you hear, depends on where your ears are, and where the sound source is. For example, in a control room, if your speakers are located at the null points for a certain mode, then they likely will not trigger that mode, and you won't hear it anywhere in the room. (But the null for one mode is often the peak or another....). On the other hand, in a typical room, as you move your head around from place to place you will be moving into locations where one or another mode is peaking, or nulling, or not doing anything. Simple example: the first lengthwise axial mode will null at exactly 50% of the room length, and peak at exactly the walls (assuming we are taking about the pressure component of the wave). At all locations in between, the level will follow a simple sine function (actually, cosine, but it's the same thing...). So it is trivial to figure out what the intensity of that mode will be like at any point between the front wall and back wall. The same applies to all the other modes.

Thus, it is simple to calculate all of the modes for any given room to find out what frequencies they will form at, and it is relatively easy to then figure out the intensity for each of those modes at any given location in the room. The overall frequency response at any point in the room is simply the sum total of all the modal pressures a that point, taking into account also the set of which modes are triggered, and how much they are triggered, from having the speakers at an any give set of points.

That's what you are seeing on that graph you mentioned. REW does all of the math for you. Use the "Room Sim" module for that. You define the room dimensions, the locations of the speakers in the room, and the location of the listening position. REW does all the rest: it figures out the frequencies of all possible modes in that room, then takes into account how much each of those would be triggered by the speakers, then how much of that you would hear when located at the the spot you choose.

It's a very useful tool... and quite accurate too, as you can see from the comparison of the actual measured response, which I posted directly below the predicted response. For an empty room with fairly massive, solid, rigid walls, the prediction is usually pretty good. However, if your walls are not very massive, not very rigid, not square and plumb, angled, or have holes in them (windows, doors, fireplaces, etc), or large alcoves, projections, etc., then the predictions might not be so accurate.

Another question is how do you predict the effect of a bass trap? i.e- i add X amount of cubic ft. of typical fluffy fiberglass to a corner, how much reduction of Y frequency should i expect at the listening postion.

You can use tools such as these to help with that: http://www.acousticmodelling.com/ Especially the "multi-layer" one. That will help you understand what any given design of bass trap should do, under ideal circumstances. But! Don't take that as Gospel truth. It's not as accurate as REW, for the simple reason that modelling acoustics absorption and resonance is a lot more complex than modelling the modal response in a rectangular room. Those tools use simplified versions of the calculations, that make a lot of assumptions.

Also, take into account that, even if the trap you design does actually perform perfectly, identically to what that tool predicts, it won't do anything at all if you put it in the wrong place. For example, hypothetically, if you look at the REW prediction and see that you are going to have a width-wise modal problem at 100 Hz, so you design a tuned trap to hit 100 Hz exactly, but then you put that trap on the back wall... it won't do much! Because the mode is forming sideways, ACROSS the room, so there's not much happening on the rear wall. On the other hand, if you put that trap on one of the side walls, then it would have a good effect....

... provided that it is big enough! Obviously, if you make a trap the size of a matchbox, it ain't gonna do much, even if you position it perfectly at the modal peak, and it is perfectly tuned. And conversely, if you make it the size of the entire wall, then it will have a major effect on that mode...

... but even then, maybe not enough! It all depends on how effective your device is. Let's say you manage to cover the entire wall with a tuned trap that can reduce the level by 6 dB. In theory, if your mode was peaking at 20 dB, the you'd still have 14 dB of mode left. A device capable of reducing it by 20 dB would, in theory, kill it completely.... but that's a tall order...!

Ditto for helmholtz resonators, tuning them to a problem mode is fairly simple via a calculator. But what happens when your looking for a certain amount of decibels in reduction at the listening position? How do i know how big to make the resonator?

Tuning Helmholtz resonators isn't as easy as the text books make it seem! There's lots of variability in materials: Eg., let's say you figured that perf panel 16mm thick with a certain arrangement of holes will exactly hit your 100 Hz mode (hypothetically!), so you go to the hardware store and buy a sheet of 16mm plywood... but it turns out to be only 15.5mm thick, but you didn't notice, then you sanded both faces because the surfaces were rough and it went down to 14.2mm thick, and the air is very dry where you love so it shrinks a bit more, and it's now only 13.8mm thick... For a Helmholtz device, that's a big difference! The tuning could be quite a bit off from where you thought it would be....

But anyway, your question is more about how big the device has to be, in order to have an effect. This goes back to the same issue as above: if it is tiny relative to the surface area of the wall, it wont do much. If it covers a large percentage of the wall area, it will be more effective. Unfortunately, here's no simple rule that says "You need 27.38% wall area coverage for a perf panel Helmholtz resonator to kill a mode that peaks 10 dB at 83 Hz and rings for 500 ms". It's not that easy. I would be nice if it was... About the best advice I can give is to make the device as big as possible, and locate it as close as possible to the pressure peak or velocity peak (depending on the type of device). Thus, the general recommendation to treat 100% of the rear wall in a control room, with treatment designed to get down to the lowest axial mode.

But still..... Its been difficult to develop a method of prediction. No book has outlined a method ie a build it like the pros, but for acoustics, and the web is peicemeal. This is excaclty what i would like this thread to do. This way me, and others can give reasonably accurate predictions for our designs and improvements, and have our own and client/friends expectations and budgets aligned with reality

I'd say that you are running into a lack of information about this, precisely because it's not that easy at all to calculate!

Under ideal conditions, predicting frequency response for a rectangular room is simple, as I outlined above. That's not a problem. You can know exactly what the modal response will be for any location, and since the modes are directly related to the walls (where I'm using "wall" to also include ceiling and floor), you can know where you will need to place treatment to deal with each issue. That's already established from above. The issue, then (your basic question) is: "How much treatment area do I need to deal with each mode?" The answer is "Just enough to do the job." That seems like a cop-out, I know, but it's the truth. For low frequency issues, that very often implies treating a very large area of the wall, often at least 25% (eg, side walls), and for the rear wall at least, as much as 100%. If you look at most pro studios, the entire rear wall is treated in one way or another, often with several layers of treatment, doing different things.

If there's one thing you can do in your room, then treat the hell out of the rear wall. All of it. 100%. The ceiling is often heavily treated too (let me suck some numbers out of thin air: 50% to 75%), the front wall less so, the side walls less too, and the floor usually has no treatment at all, for psycho-acoustic reasons.

There's also the issue of design concept: If you are going with the original LEDE concept (hopefully you wont do that! Not such a good design..), then the decision is already made: one half reflective, one half absorptive. Ditto for other design concepts: RFZ, FTB, NER, MR, CID, etc. Each has it's own "rules", and those tend to dictate how much of each wall will be treated, and in what way it will be treated.

But if you are going with your own design, then it gets a little more complex to figure out. I wish there were an easy way, a simple set of rules or equations, but unfortunately, there isn't. There's just general guidelines.

About the only way you can do this predicatively, is with software. There are FEM/FEA and BEM based programs for predicting the acoustic response of rooms, and they can do a good job of helping you... but once again, you have to give them valid data, such as defining the characteristics of each surface in the room. Some of them are very fancy, and will also "auralize" the results, which basically means that you can choose a listening position in the room, give it a song, and it will produce an audio file that simulates what that song will sound like at that location in the room, with the parameters you specified. Interesting tools! But you'll need a lot of money to buy that! Then a lot of time to learn how to use it.

as opposed to the old "as much bass trapping as you can fit",

Great advice, actually! Very true!

"as much mass as you can afford",

For isolation, sure. But that can be calculated reasonably well. Isolation is predictable. Much more so than treatment.

Basically the transition from a journeyman to black belt has been allususive so far, largely due to inability to predict certain results..

I hear you, and yes, that sounds about right! I understand your frustration: It is illusive, and it comes with experience. Or with money: buying those fancy tools and learning them will certainly help.

Its akin to a mucisian playing an E chord, and knowing that they are playing an E chord, then knowing when an E chord could sound right at a certian point in the riff. Or knowing how to play instruments seperately, but then applying that to writing a song.

A great analogy! In that sense, acosutic design is a lot like music composition: How do yo know to use an electric guitar rather than a flute or a trombone to play the bridge? How loud should the keyboard be, in relation to the bass? Should you repeat the chorus twice, or just once in a ballad? Can you teach an inexperienced musician a set of rules like that for composing songs? Or does it come with experience?


- Stuart -

Kpgushue wrote:Isolation, is somewhat more straightforward (to me at least) to give a good estimate of, though technical knowledge and attention to construction detail is necessary, the mass law, budget, and existing conditions or exptations, can get go a long way.

My forte is sound control in construction. I spend my days talking to architects, builders and developers about how to achieve good isolation between sleeping rooms and dwelling rooms. Here in the US there is a code that specifies STC/IIC 50. Architects know this and design for it but many still do not a have real understanding about the science behind their designs. I enjoy talking to them about Mass Law and I enjoy even more talking to them about the various ways we can cheat Mass Law through things like damping, decoupling and the basic mass/air/mass assembly.

I sometimes get plans, however, where an architect has specified STC 50 for a wall assembly without really designing anything that will achieve that. They leave it to the builder to....apply the science.

The tricky part in this is that there are a great many assemblies being designed with no real sound testing to back them up. I am constantly being asked to extrapolate between extant testing to arrive at STC and IIC estimates. I always start with real base data and then look for other test data that is similar to their proposed designs.

I know how important real scientific data is but many projects have been built from my less than purely scientific extrapolations.

I sometimes get plans, however, where an architect has specified STC 50 for a wall assembly without really designing anything that will achieve that. They leave it to the builder to....apply the science.

That sounds just a little like the blind leading the blind! The outcome is probably out there, in the Twilight Zone some place...

I wonder if the architect does that to cover his butt legally? So if the isolation is deficient he can blame it on the builder, saying: "But I DID specify STC50 on the plans! It's his fault that he didn't do that....". Or maybe he just doesn't have a clue how to do that?

I always start with real base data and then look for other test data that is similar to their proposed designs.

I imagine that you are aware of IR-761? That might be a good document to point them at, so they can learn to get it right all by themselves. It's in the document library, here on the forum: viewtopic.php?f=9&t=4 Feel free to point the architects (and builders!) at that.

I know how important real scientific data is but many projects have been built from my less than purely scientific extrapolations.

Right, but you have the background and understanding to do that! It sounds like the architects you are dealing with, don't have much of that at all. If only they universities would give them more acoustic theory and practice during their training.... Sigh!


- Stuart -

Soundman2020 wrote:I imagine that you are aware of IR-761? That might be a good document to point them at, so they can learn to get it right all by themselves.

That is an excellent document and I have passed it on to some who expressed an interest in some of the basic assemblies. The problem is that IR-761 doesn't drill down deep enough into some of the subtle but important nuances that come up in large scale construction.

Things like the difference between Type X and Type C GWB and how different gauges of steel studs perform when you have to start stacking up floors for mid-rise or high-rise construction......those details can make a measurable performance difference.

Younger architects are getting a lot of the theory in school these days I think but for the non-scholastic types like builders and developers all this still feels a lot like a black art, right? That's why I think the premise of this thread.....predictive acoustics...... is such a good one. A lot of folks just believe the acoustic spin that they read on the web from companies pedaling acoustical products that may only work under a certain set of circumstances. When we, as professionals, are asked to provide a performance estimate.......we need to be able to apply the science.

https://www.marshallday.com/innovation/ ... ning-room/
http://www.rpgeurope.com/products/product/tools.html
http://www.rhintek.com/cara/cara21desc.php
https://www.roomeqwizard.com/help/help_ ... alsim.html

DD

DanDan wrote:https://www.marshallday.com/innovation/the-listening-room/
http://www.rpgeurope.com/products/product/tools.html
http://www.rhintek.com/cara/cara21desc.php
https://www.roomeqwizard.com/help/help_ ... alsim.html

DD

I would take CARA with a grain of salt.
It calculates situations in a peculiar way if I remember right.

bert stoltenborg wrote:I would take CARA with a grain of salt.
It calculates situations in a peculiar way if I remember right.

I haven't used it myself, but I heard the same thing. I'm not quite sure how it works, but it doesn't seem to be normal BEM or FEM from what they show on their website.

- Stuart -

Early days perhaps. I use Altiverb all the time. Over time it has included extra tweaks to alter the parameters of the sample in use.
It still sounds utterly 'real' even if one stretches the issue.
I guess that suggests that virtual modelling is probably very possible. Particularly with a new build where all parameters are user defined.
But I would caution that early attempts which focus on say modes, and omit BIR, haven't tied in with reality for me.
DD