John Sayers' Design Forum

John Sayers' Recording Studio Design Forum

A World of Experience
Click Here for Information on John's Services
It is currently Sun Dec 08, 2019 6:04 pm

All times are UTC + 10 hours




Post new topic Reply to topic  [ 6 posts ] 
Author Message
PostPosted: Sun Jul 14, 2019 8:26 pm 
Offline

Joined: Sun Jul 14, 2019 8:06 pm
Posts: 3
Hi,

In my Vocal Booth (2.2 x 2 x 2.1m) I have to use some Diffusers. I've already a drawing of another room with similar Dimensions, so I know where I need which acoustic panel. But they use Diffusers like this: https://www.thomannmusic.ch/hofa_diffusor.htm

That's a 2D Diffuser, a lot of work to build. So I want to use 1D stepper Diffusers like at the att. And turn them 90°, so I should have Diffusion in 2D too. But this Difuser is 420mm Wide. What happens when I scale just the width down to 210mm? Would the upper Freq. limit be 2 times higher? I would like to scale it to 210mm, then cut it into 210x210mm pieces and mount it in a 90° angle.

There are a lot simulations of Diffusers in the internet with different mathematical models, but I've found almost no measurements. Sometimes I think they are afraid of the measurements :-)

Thanks

Tom


You do not have the required permissions to view the files attached to this post.


Report this post
Top
 Profile  
 
PostPosted: Mon Jul 15, 2019 10:53 am 
Offline
Senior Member
User avatar

Joined: Thu Aug 21, 2008 10:17 am
Posts: 11970
Location: Santiago, Chile
Hi. Please read the forum rules for posting (click here). You seem to be missing a couple of things! :)

Quote:
In my Vocal Booth (2.2 x 2 x 2.1m) I have to use some Diffusers.
Why do you "have" to have diffusers in your vocal booth? Who told you that? Whoever it was, you can ignore their advice from now on, because they don't have a clue about acoustics.

Vocal booths do not need diffusion like that. And even if they did, it would not be possible to use a numeric-series diffuser in such a small room. You should never use a diffuser within 10 feet (3m) of your ears or a mic, because of this:
Attachment:
QRD-Diffusion-lobing--pattern-graph-SML-ENH-2.PNG

That shows the lobing patterns formed in the air close to a diffuser. All of those artifacts do not get fully smoothed out into a really diffuse field, until a long distance away from the diffsuer: at least 3m (10 feet), and perhaps more, depending on what frequency you tune your diffuser for.

Which leads to another question: What frequency range WOULD you tune your diffuser for, in a vocal booth? And why would you choose that frequency range?

And another question: Where would you put such a diffuser in a vocal booth? The wall you face? The wall behind you? One of your side walls? The ceiling? The floor?

Quote:
That's a 2D Diffuser, a lot of work to build.
Yes it is a lot of work to build, and yes it is 2D, but it is not a diffuser (despite the name on the website). It claims to be a QRD diffuser, but it very definitely is not. Numeric diffusers (including QRD diffusers) have to be based on prime numbers, which I'm sure you know are all odd numbers, never even. Yet that "diffuser" has 6 wells in each direction! "6" is not a prime number, so that is not a QRD diffuser. Prime numbers are necessary for a very simple reason: it's the only way to get the diffuser to send out the same acoustic power in all directions, equally. If you don't use a prime number when you design it, then you do not get flat power response: there will be more power going out in some directions, and less in others... thus, the sound field around the diffuser will be even more contorted and mangled than in the image above, and it will never smooth out completely... (well, it won't ever smooth out before the diffuse field blends into the background field of the room, and is lost...) That device might scatter a bit, but not evenly, not smoothly, and it is not a true diffuser. So you don't need to build it anyway! :)

The other one is a Tim Perry design, and truly is a proper diffuser: he has some good designs, and I have used some of his designs in studios that I have done for clients. You'll notice that the elements in his diffuser do have an odd number of wells, and are based on prime numbers, so they do work correctly. Here's a link to an example of a control room I designed for a client, that is currently under construction, and uses Tim's LeanFuser: thread about Steve's high-end control room in New Orleans

However, I would still not use such a device in a vocal booth. I doubt that Tim would recommend that either. To start with, it is rather large to fit in a vocal booth! But even if it could fit (using just one module, for example), I still would not do that, for the reasons I mentioned at the start.

Quote:
There are a lot simulations of Diffusers in the internet with different mathematical models, but I've found almost no measurements. Sometimes I think they are afraid of the measurements
They are not afraid of measurements: the problem is that it's very hard to measure "diffusion" in a real room, and even harder to interpret the results. The untrained eye would likely not be able to understand the resulting graphs. For example, here's one of the many tests we did in that very same room you see in the link, just before and just after the diffuser went in, and at the same time we took out some old absorption panels:
Attachment:
STN--REW--FR--20-20k--With-[GREEN]-and-without-[BLUE]-leanfusers.png

Can you look at that frequency response graph, and explain what the diffuser did there? Can you see where it made a difference, and where the absorption panels made a difference, just from looking at that?

Here's the actual impulse response graphs for that (in the form of an ETC):
Attachment:
STN--REW--IR--100ms--With-[GREEN]-and-without-[BLUE]-leanfusers.png

Can you look at that and determine if the diffuser is doing it's job? Which part of the changes in IR, are due to the diffuser, and which parts are from the absorption changes?

Here's a third graph, that also shows the Impulse response, but this time smoothed to 0.2ms to make it easier to see, and including a third curve, which is after additional treatment was added to the room:
Attachment:
STN--REW--IR--100ms--With-[GREEN]-and-without-[BLUE]-leanfusers-and-other-[RED].png

Can you see what the diffuser is doing in there? The ETC graph is the most useful of all, for an acoustician, because it reveals everything about the room: exactly how it responds to a sudden sharp transient impulse. So everything you need to know about the diffuser is in there... but it's pretty hard to see!

In fact, if you look at the blue curve and the red curve, most people would say that they are pretty much the same, with practically no differences, so they would assume that the treatment is not working.... but this is what the studio owner said after he listened to some music in there for the first time after installing the diffuser (excuse the graphic language!):

Quote:
I have no way of listening to reference CDs on my JBL mains right now -- so we actually are listening on a Panasonic "boom box" with the speakers up on music stands, in front of the JBLs at the same height, aimed the same way and a mere 2" from the soffits. We listened to a bunch of commercial reference CDs. The room has to be truly amazing [because] it transforms a shitty boom box into high fidelity, larger than life listening experience with really well mixed material. All I can say is that we never heard those mixes sound that good and that clear, and I had to keep reminding myself this is just a boom box ! And the mixes still sounded big and powerful walking all the way back to the diffuser, euphonic and euphoric and never feeling sucked out or dull.

So basically, Hell yeah! :) That diffuser is making a huge difference to the room. And that comment comes from a mix engineer with many years of experience in all types of studios. He's blown away by what it does for his room. It even makes lousy speakers sound good.

But you'd never know that from looking at acoustic response graphs, without an in-depth understanding of what you are seeing!

So that's why diffuser designers and manufacturers don't usually bother posting test results: It's hard to measure, harder still to present, and REALLY hard to interpret.

Now, before you get all excited about the studio owner's comments, and think that you need such a device in your vocal booth, take into account that this is a heavily treated control room, carefully designed to be as good as possible, and the room is TWELVE TIMES larger than your booth. The air volume of your booth is about 9 cubic meters. The air volume of that control room is about 112 cubic meters....

So that's the sad truth: you don't need a diffuser in a vocal booth. Even worse, your booth is practically a cube! 2.2 x 2 x 2.1m. The only real option you have is very heavy damping, then some reflective surfaces on the wall you are facing and the side walls, such as varying wood slats, angled, and maybe a shallow-curve poly hung from the ceiling to provide a little smooth diffusion in the vertical plane. That's what I would do in such a room. It's still going to sound somewhat "boxy" though, because of the size. You can't make a small room sound large.


- Stuart -


You do not have the required permissions to view the files attached to this post.

_________________
I want this studio to amaze people. "That'll do" doesn't amaze people.


Report this post
Top
 Profile  
 
PostPosted: Thu Jul 18, 2019 5:32 pm 
Offline

Joined: Sun Jul 14, 2019 8:06 pm
Posts: 3
Hi Stuart,

thanks for your detailed answers. Sure it's difficult to measure diffusors, much more difficult than just to take a model, write a simple software & make some simulations. For me, a diffusor is something that diffuses sound, so I would say, this 6x6 panel will be a diffusor too.

Isn't there an university who made measurements of different diffusors angle, frequency & distance dependend in a anecoic room? It's just because I would be interested. If you have models you have to proof them, if you don't do this, it makes no sense to use them in real.

I've studied physics, made my diploma work in nano-optics (optical resonant antennas in combination with fluorescent silicone quantum dots). We made measurents and I've done some E-field simulation with FDTD (finite difference time domain). Reality and models are not always the same. They just show some tendences.

There are a lot of people who recommend diffusors in a small vocal booth. Sure: you never know if they just wanna sell you some stuff or really know what they're talking about.

It's known since a long time, that you cannot say, if a speaker sounds good just because of the measurements, so how then should it be possible, to make some recommendations if a diffuser has a possitive influence in a small room just because of a simple software simulation?

Cheers

Tom


Report this post
Top
 Profile  
 
PostPosted: Thu Jul 18, 2019 6:47 pm 
Offline
Senior Member
User avatar

Joined: Thu Aug 21, 2008 10:17 am
Posts: 11970
Location: Santiago, Chile
Quote:
For me, a diffusor is something that diffuses sound, so I would say, this 6x6 panel will be a diffusor too.
Actually, no it isn't. It's a "scatter". It does not diffuse sound: it scatters sound. The term "diffuse" has a very specific meaning in acoustics: a diffuse field is one where it ie equally probably that the sound will come from any specific direction at the same intensity as any other direction. A 6 x 6 panel of humps and dips won't do that. Sorry. You might want it to do that, but it doesn't. It cannot do that because it does not produce flat power response: in other words, it delivers different power in different directions, thus it is NOT diffusing, and therefore it is not a diffuser.

Quote:
Isn't there an university who made measurements of different diffusors angle, frequency & distance dependend in a anecoic room?
There are many independent acoustic test laboratories around the world, but probably the best known is Riverbank. As I said before, it is certainly possible to test them, and they do offer such testing: the problem is that most people don't understand how to interpret the test results, because it is not intuitive at all.

Unlike absorbers, the test results for diffusers don't actually mean anything in simple terms. An absorber with an absorption coefficient of 0.8 at 250 Hz is easy to understand: it absorbs 80% of randomly incident sound at that frequency. Easy. But if a diffuser has a diffusion coefficient of 0.8 at 250 Hz, what does that mean? How do you visualize that? How does that relate to angles, and intensities, and phase relationships? There is no simple answer. It is certainly possible to measure that, and draw a graph that shows the coefficient for each frequency, but that is meaningless to most people. That's what I was explaining yesterday. Tests can be done, but the vast majority of people do not know what the results mean. So why bother?

Quote:
If you have models you have to proof them, if you don't do this, it makes no sense to use them in real.
OK, here is such a graph: please interpret it for me:
Attachment:
acoustic-diffuser-graph--coefficient-of-diffusion.jpg

That's a real graph for a real product, tested in a real test lab. It compares the diffusion of that device to a flat panel reflector. Please explain what it means. For example, at 1,600 Hz, the device has a coefficient of 0.6. What does that mean? Does it mean that only 60% of the sound was diffused, and 40% was not? Or does it mean that the phase changed by 60%? Or does it mean that the intensity changed by 60%? Or does it mean that the angle changed by 60%? How would I use that graph to determine if the diffuser is any use for my room? How would I use that graph to determine where I should place the device in my room? Can you look at that graph, and tell me if there will be lobing patterns at 1,600 Hz? And if so, in which direction will the lobing happen? How serious is the lobing? Does the 0.6 mean that 60% of the incident sound will be concentrate in lobing patterns, while 40% will not?

Perhaps you can see now why it makes no sense to publish such graphs: they don't actually mean anything, in real world terms. It's easy to understand a graph for porous absorption, or for a resonator, but not for a diffuser.

Quote:
There are a lot of people who recommend diffusors in a small vocal booth.
And they would be wrong! :) There are also lots of people that recommend putting egg cartons on the ceiling of your vocal booth, and lots of other people who recommend putting carpet on the walls. Just because lots of people recommend something, does not mean that it works, or is a good idea.

I'd suggest that you should read the book "Acoustic Absorbers and Diffusers" by Cox and D'Antonio. You can learn all about these concepts, and difficulties.

Quote:
Sure: you never know if they just wanna sell you some stuff or really know what they're talking about.
... true. Or also of they just don't have a clue what they are doing, and thought it looked good.

But take a look at professional vocal booths in high-end studios, designed by leading studio designers: how many of those have diffusers on the walls, near the talent or the mic? :)

Quote:
It's known since a long time, that you cannot say, if a speaker sounds good just because of the measurements,
Actually, that isn't true. As Floyd Toole shows in his book "Sound Reproduction", there is a set of four key parameters that can be measured in an anechoic chamber, and a simple graph that shows all four of them is an extremely good indication of how that speaker will be judged by both expert listeners and causal listeners. Measure the frequency response on-axis, measure again at a set of fixed locations called the "listening window", measure the early reflections, and measure the sound power spectrum. Plot those four on a single graph, and there is very high correlation to the quality of the sound, as judged by all types of people. You should read the book to understand why this is so. It's very logical, and very clear. So your statement is incorrect: it is in fact, possible to say if a speaker sounds good or not based ONLY on the measurements.

Quote:
so how then should it be possible, to make some recommendations if a diffuser has a possitive influence in a small room just because of a simple software simulation?
If the simulation produces meaningful data (not just a graph of coefficients vs. frequencies), then it can indeed be used to determine if the diffuser would be any use in the room. How do you think I decided that a 7-panel LeanFuser with a multiplier of 3, and that height, would be a good choice for the room you see in here?
Attachment:
STVNOUS--room-rear-sofa-riser-diffuer-complete-10.jpg

I analyzed what the room needed, I looked at alternative treatment possibilities, modeled some of them, and determine that this one would do what I wanted. And it does. :) I discarded QRD and PRD skylines, and BAD panels, and polys, because the didn't produce the results I wanted for this room. I have used all of those in other rooms, and obtained the results I wanted in each case, but for this room, the Leanfuser was the best choice.

So I'm not sure why you say that it isn't possible to do that. Just like it is possible to predict how a speaker will sound based on those four measured parameters, so too is is possible to predict what a diffuser will do in a room, based on suitable measurements.

I highly recommend both of the books I mentioned: they should help you to understand that it is possible to predict acoustic outcomes, based on measurements in acoustic test labs, and also based on software simulations.

Here's a simple case of room frequency response:
Attachment:
Timeguy--REW--FR--Predicted-vs-measured.jpg

The top graph is the predicted frequency response curve for a room, and the bottom graph is the actual response measured in the room. Excellent correlation between them, and that's a rather complex thing to do: it takes into account the dimensions of the room, the locations and characteristics of two main speakers and two subwoofers, and the treatment on all six boundary surfaces, and also the location of the mic.
Acoustic software prediction and simulation works a lot better than you give it credit for. I'm really glad it does, or I would be out of a job! :)


- Stuart -


You do not have the required permissions to view the files attached to this post.

_________________
I want this studio to amaze people. "That'll do" doesn't amaze people.


Report this post
Top
 Profile  
 
PostPosted: Thu Jul 18, 2019 8:47 pm 
Offline

Joined: Sun Jul 14, 2019 8:06 pm
Posts: 3
Hi Stuart thanks a lot,

I think one of the "problem" is that a diffuser cannot be characterized by a single parameter. The diagrams you showed me are nice, but there's no interpretation I see. It always nice to show a graph, especially if the customers have no idea about it.

Why don't they measure the dB for different frequencies at different angles for different distances in an anechoic room? Then they should compare the results with a flat surface of the same size & then you should try to make some interpretations. Before you don't have these datas it makes no sense to discuss about it. Perhaps they have to measure at 1000 different mic positions, that's work, but I would be interested in such a result. It wouldn't be possible to show the result in just 1 diagram.

These simple diagrams and parameters remind me to car sellers who talk about PS. But do you really know at which rpm? Do you know the torque because of that? Does it makes sense to talk about PS or torque if you don't know the weight of the car?

Cheers

Tom


Report this post
Top
 Profile  
 
PostPosted: Sat Jul 20, 2019 3:28 am 
Offline
Senior Member
User avatar

Joined: Thu Aug 21, 2008 10:17 am
Posts: 11970
Location: Santiago, Chile
Quote:
Why don't they measure the dB for different frequencies at different angles for different distances in an anechoic room?
You mean like this?
Attachment:
sample-QRD-diffuser-plots.jpg


Attachment:
sample-QRD-diffuser-plots--2.jpg


Attachment:
sample-QRD-diffuser-plots--3.jpg


In other words: they DO make such measurements. It's not a new thing either: way back in 1969 the BBC research department was already developing techniques to do this, including using rotating microphones synchronized to paper-tape recording systems...

You seem to be assuming that this type of testing is not done, or that it is kept secret to hide the defects of diffusers, but that is not the case. The data is there, it does exist. It's just bot published by most manufacturers, because it makes little sense to do so. Software these days is very good at predicting the performance of diffusers, and real-world testing in acoustic labs validates the predictions.

Quote:
Before you don't have these datas it makes no sense to discuss about it.
The data is there: both predicted and measured. The measured data validates the prediction models. It's much, much easier to design diffusers in software and test them with software prediction, rather than build an endless series of physical devices and test them in real labs. You can design and test hundreds of possible layouts in the same time it would take to build and test just one physical model. You should look at the brief paper called "Acoustic Diffusers: The Good, The Bad And The Ugly" put out by Prof. Trevor Cox a few years ago, where he did exactly that, looking at many types of diffuser, including some rather unusual organic-shaped diffusers that work really well. There's a similar paper "Three Limitations of Traditional Number Theory Diffusors" also worth looking at.

Quote:
Perhaps they have to measure at 1000 different mic positions, that's work, but I would be interested in such a result. It wouldn't be possible to show the result in just 1 diagram.
They do, in fact, make such measurements, and it is, in fact, possible to plot the results in a meaningful fashion on one graph: see above. Once again, you seem to be making assumptions that aren't true.

But returning to my original point: while it is possible to assign a single number to an absorber that is somewhat meaningful, such as NRC for overall absorption, or maybe even sabins, it simply is not possible to do that for diffusion. There are such numbers, but they are meaningless out of context. Saying that an absorber provides 17 sabins of absorption, or has an NRC of 0.87 carries some meaning: it tells you useful (albeit incomplete) information about the material, and even hints at how if could be used. But there are no units that you can measure diffusion in, like you can with sabins of absorption, and saying that a device has a coefficient of diffusion of 0.87 tells you nothing at all about the device. That's the issue. Diffusion is far, far more complex than absorption, it can't be used in small rooms anyway (rooms the size of typical home studios), it needs very specialized knowledge and understanding of acoustics to be able to use them correctly, and most home studio builders don't have that knowledge, and don't know how to interpret the full acoustic data for a diffuser, so what's the point in publishing it? Only acousticians need that data: we know where to find it, how to use software to predict it, and how to interpret the results, and how to use the devices. Putting that data out in public doesn't serve any useful purpose, and simply confuses and/or misleads studio builders, such as yourself, into thinking that diffuses are useful for applications where they were never intended, and would not performs as expected anyway!

- Stuart -

- Stuart -


You do not have the required permissions to view the files attached to this post.

_________________
I want this studio to amaze people. "That'll do" doesn't amaze people.


Report this post
Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 6 posts ] 

All times are UTC + 10 hours


Who is online

Users browsing this forum: No registered users and 9 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Jump to:  
Powered by phpBB® Forum Software © phpBB Group