The No.1 Website for Pro Audio
Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
11th October 2012
#61

Quote:
Originally Posted by kasmira
Where did you get the number 12.3 mm? Not to destroy all the work you've done, but that isn't the actual size for any lumber I've seen. Typically, 1/2" lumber nominal is 3/8" actual (9.525mm)

Also, if we were using something like MDF, plywood, hardwood plywood, etc - the actual thickness is the nominal thickness (at least, from what I've seen). Only with boards (1x3, 2x4 etc) is it not. So we could calculate it with either .75" (19.05mm) - the actual thickness of a 1x# board, or by exact measurements using MDF or something of the sort.

Also, found this on Home Depot online: 5.0 mm x 2 ft. x 4 ft. Lauan Moisture Resistant Underlayment Handy Panel-1507004 at The Home Depot

Only thing I can really find that looks useful that is actually in metric units..

Damn the imperial system

It looks like the actual size can vary based on the specific material.

Plywood and MDF
For plywood thickness, this chart suggests 1/2" nominal is actually 31/64 = 12.3 mm thick. Nominal Wood Thicknesses - NewWoodworker.com LLC

I found another chart that gives an actual thickness of 15/32" = 11.9mm thick (again, for plywood).

For medium density fibreboard (MDF), the nominal thickness is usually the actual thickness (1/2" = 12.7 mm). However, I've read that it may vary between 15/32" (11.9 mm) and 1/2" (12.7 mm).

Softwood Lumber
For 1/2" softwood lumber, this chart gives a minimum dressed thichness of 7/16" (dry) to 15/32" (green) = 11.11 mm to 11.9 mm.
Softwood Lumber Standard

Hardwood Lumber
For hardwood lumber dimensions, you're right: Wikipedia gives 1/2" hardwood lumber an actual thickness of 9.525 mm. Lumber - Wikipedia, the free encyclopedia

To estimate performance, round your material thickness to the nearest simulated thickness: 10 mm ('metric') or 12 mm ('imperial')
If hardwood is the most appropriate material to build these, then you're right that 9.525 may be a more typical dimension. But, when picking a size to simulate, I noted that I can't simulate 9.5 mm material using Reflex... I have to round it up to 10 mm (giving it identical simulation results to the metric version).

Reflex only simulates designs within a precision of 1 mm. (I should have mentioned that the imperial design I simulated is actual 12 mm thick.)

Depending on the material you use, the 10 mm ('metric') or 12 mm ('imperial') simulation results will apply to you.

Does that make sense?
11th October 2012
#62
Lives for gear

Good researching Tim. MDF though, I've never seen be anything but actual. IME, it is always the nominal thickness. Same with plywood here in Arizona (if this makes any more sense, they don't sell 1/2" plywood here - only 7/16s, 19/32s, 11/32s, etc so any time I've seen it, the actual thickness is displayed as it's nominal thickness) - for example: Plywood/Sheathing - Plywood - Plywood, Sheathing & Subfloor - Lumber & Composites

And Tim thats a great idea...just choose the one you can get closest too and go for it!
11th October 2012
#63
Lives for gear

Quote:
Originally Posted by Arqen
Reflex only simulates designs within a precision of 1 mm.
Reflex can handle smaller dimensions but only displays full mm increments (check it yourself by adding 10 1,1 mm sections and the total width will be 11 mm). I´ve reported this GUI bug to AFMG a long time ago.
11th October 2012
#64

Quote:
Originally Posted by kasmira
Good researching Tim. MDF though, I've never seen be anything but actual. IME, it is always the nominal thickness. Same with plywood here in Arizona (if this makes any more sense, they don't sell 1/2" plywood here - only 7/16s, 19/32s, 11/32s, etc so any time I've seen it, the actual thickness is displayed as it's nominal thickness) - for example: Plywood/Sheathing - Plywood - Plywood, Sheathing & Subfloor - Lumber & Composites

And Tim thats a great idea...just choose the one you can get closest too and go for it!
Thanks for pointing that out! In that case, 1/2" MDF has a thickness of 12.7 mm, and 3/8" MDF has a thickness of 9.525 mm.

I'm not sure which one will perform better, as I've not tested out a version of the design with 12.7 mm material.
11th October 2012
#65

Quote:
Originally Posted by Jens Eklund
Reflex can handle smaller dimensions but only displays full mm increments (check it yourself by adding 10 1,1 mm sections and the total width will be 11 mm). I´ve reported this GUI bug to AFMG a long time ago.
Thanks for pointing that out, Jens.

When doing simulations, I've been typing in exact dimensions just in case it was a GUI issue. I'm glad that Reflex was indeed simulating them with high precision!
12th October 2012
#66
Lives for gear

Hey Arqen,

Question: what program do you use for your YouTube simulation models? Does Reflex do this or is it another program? (i would like to simulate different configurations for my room)
Thanks for the easy build instructions.
12th October 2012
#67

Quote:
Originally Posted by AwwDeOhh
Hey Arqen,

Question: what program do you use for your YouTube simulation models? Does Reflex do this or is it another program? (i would like to simulate different configurations for my room)
Thanks for the easy build instructions.
I programmed those time domain video simulations using a scientific programming language called Matlab, along with an acoustics toolbox (for Matlab) called k-wave.

But, Reflex is your best bet if you want to do your own performance simulations! Reflex does not let you do time domain animations like my videos show, but it does give you great performance predictions.

Let me know if you have any more questions about my Matlab simulations, or about Reflex.

-Tim
12th October 2012
#68
Lives for gear

Quote:
Originally Posted by Arqen
I programmed those time domain video simulations using a scientific programming language called Matlab, along with an acoustics toolbox (for Matlab) called k-wave.

But, Reflex is your best bet if you want to do your own performance simulations! Reflex does not let you do time domain animations like my videos show, but it does give you great performance predictions.

Let me know if you have any more questions about my Matlab simulations, or about Reflex.

-Tim
Thanks, i appreciate it.
Was mainly interested in modeling how several diffusors interact with each other in multiple points around the room (ie, from back wall, to side wall to ceiling, ect.). though the Reflex software interests me as well.
Is k-wave/Matlab capable of doing 3-dimentional models? if not, any suggestions?
12th October 2012
#69

Quote:
Originally Posted by AwwDeOhh
Thanks, i appreciate it.
Was mainly interested in modeling how several diffusors interact with each other in multiple points around the room (ie, from back wall, to side wall to ceiling, ect.). though the Reflex software interests me as well.
Is k-wave/Matlab capable of doing 3-dimentional models? if not, any suggestions?
No problem!

Is k-wave/Matlab is capable of doing 3-D model of diffusion in a room? Technically, yes. But it's not very practical...

k-Wave was originally designed for medical imaging, and Matlab is a programming language. To build a diffusion simulation environment using these tools, I had to do a lot of programming and pulling my hair out.

I used a simulation method called finite difference time domain (FDTD), which is extremely slow if done in 3D. Reflex, on the other hand, uses the boundary element method (BEM)... which is more time efficient, but does not yield sexy video animations.

Also, k-wave/Matlab is not accessible to most people (Matlab is very expensive and takes time to learn). I have years of experience programming in the Matlab language, and even then it took months of research and over 100 hours coding and testing the simulation environment to produce the videos you see.

If you have a background in scientific programming you can try out Octave, a free Matlab clone. And if you have no interest in scientific programming, I don't know if there is any existing software that will solve the problem you're interested in (well, maybe some \$10k acoustics software ).

Basically, the type of simulation you're talking about would require a lot research and programming, and if you decide to take on such a huge project you may as well turn it into some sort of academic paper. It would be hard work... but it could be worth it!
12th October 2012
#70

Quote:
Originally Posted by AwwDeOhh
Was mainly interested in modeling how several diffusors interact with each other in multiple points around the room (ie, from back wall, to side wall to ceiling, ect.).
That's a complex simulation to do precisely (especially in 3D), but I think you if you're concerned with how multiple diffusers contribute to a room's acoustics, you don't need to be as precise. You may be able to approximate it with ODEON (very expensive software, shown below).

14th October 2012
#71
Lives for gear

Thanks.
Hehe... i think it would be quicker to REW test every single variable location, combination; build, disassemble then build different diffusors, testing each along the way than it would be for me to write the program...all while somehow paying the bills!

I'm not coding illiterate, but i might as well be...
My extent of that world involves a 'C++ for Dummies' book and a couple of poorly written very useless programs that calculate gas mileage and the months/weeks/days/hours/minutes/seconds of life on your birthday...

Oh well, i suppose in another 10-years, (5,256,000 minutes or 315,360,000 seconds ) our cell-phones will be able to crunch out these kinds of things.

Thanks for the info.
...maybe i can convince one of my techie friends that there's a good reason to make such a program...
14th October 2012
#72

Quote:
Originally Posted by AwwDeOhh
Thanks.
Hehe... i think it would be quicker to REW test every single variable location, combination; build, disassemble then build different diffusors, testing each along the way than it would be for me to write the program...all while somehow paying the bills!

I'm not coding illiterate, but i might as well be...
My extent of that world involves a 'C++ for Dummies' book and a couple of poorly written very useless programs that calculate gas mileage and the months/weeks/days/hours/minutes/seconds of life on your birthday...

Oh well, i suppose in another 10-years, (5,256,000 minutes or 315,360,000 seconds ) our cell-phones will be able to crunch out these kinds of things.

Thanks for the info.
...maybe i can convince one of my techie friends that there's a good reason to make such a program...
Haha, yeah... it's only a matter of time until someone makes an acoustics program that can do that, and desktop computers become powerful enough to handle it!
15th October 2012
#73
Gear maniac

Another question I've just thought of:

Seeing the range of these stepped diffusers is very narrow, what happens an octave above the cut of frequency ? Are the diffusers acting then just like a flat wall behind you as far as anything over 3.5khz or so? And if that is correct, that can't be a good thing in terms of reflections can it?
15th October 2012
#74
Lives for gear

Quote:
Another question I've just thought of:

Seeing the range of these stepped diffusers is very narrow, what happens an octave above the cut of frequency ? Are the diffusers acting then just like a flat wall behind you as far as anything over 3.5khz or so? And if that is correct, that can't be a good thing in terms of reflections can it?
I believe (at least for QRD type diffusors... i'm not sure on the step) that the design frequency multiplied by the prime used (number of wells/steps) equals the plate frequency.
plate frequency= top HF cutoff of the design.. so while it may diffuse to a certain Hz, it will still scatter energy up to the plate frequency.

Now, your understanding of the 'flat wall' actually happens below the useful frequencies of a diffusor (ie, over an octave lower then the design, it begins to appear as a flat wall again to low frequencies). this can be somewhat remedied by the 'nesting' of the array, so in effect it becomes a fractal design with the size of each seperate diffusor becoming the well width of the nest.
15th October 2012
#75

Acoustic Diffusion Design Bandwidth

Quote:
Another question I've just thought of:

Seeing the range of these stepped diffusers is very narrow, what happens an octave above the cut of frequency ?
Good question.

The stepped diffusers were optimized for performance in the mid-frequencies, but they will still create diffusion beyond the upper limit of the "diffusion band" (the maximum design frequency). In other words, scattering will occur above 4kHz, but I'm not sure of the diffusion quality.

Consider that the highest frequency of interest for predicting room acoustics is typically about 5500 Hz, and most commercial diffusers don't publish diffusion coefficients above 4kHz. To speed up the simulations and to help with easy comparison against commercial products, I've only tested the stepped diffusers up to 4kHz.

The fractal diffusers, on the other hand, have been tested up to 8kHz and beyond.
16th October 2012
#76

Quote:
Originally Posted by AwwDeOhh
I believe (at least for QRD type diffusors... i'm not sure on the step) that the design frequency multiplied by the prime used (number of wells/steps) equals the plate frequency.
plate frequency= top HF cutoff of the design.. so while it may diffuse to a certain Hz, it will still scatter energy up to the plate frequency.

Now, your understanding of the 'flat wall' actually happens below the useful frequencies of a diffusor (ie, over an octave lower then the design, it begins to appear as a flat wall again to low frequencies). this can be somewhat remedied by the 'nesting' of the array, so in effect it becomes a fractal design with the size of each seperate diffusor becoming the well width of the nest.
Right on with the low frequency stuff...

Now, regarding the high frequency cutoff... stepped diffusers are designed using different principles than QRD diffusers, and while the QRD plate frequency calculation might be useful for speculating the cutoff frequency, I would not expect it to be accurate for stepped diffusers in general.

A better way to assess the cutoff frequency would be to simulate the design up to high frequencies (unfortunately, my trial of Reflex has expired, so I can't do that at the moment). For optimized stepped diffusers, simulation (by physical modelling) is central to the design process because it's the most practical way to predict the performance. At least... so far as I know .

Thanks for pointing out to everyone how the low freq limit works!
16th October 2012
#77
Gear maniac

Tim

A question on behalf of someone else. Would a variance of 1/2 to 1mm in any of the dimensions have an adverse effect on the diffusion behaviour you predicted?
16th October 2012
#78

Quote:
Tim

A question on behalf of someone else. Would a variance of 1/2 to 1mm in any of the dimensions have an adverse effect on the diffusion behaviour you predicted?
It would change the performance, but it would most likely be subtle.

Note that I optimized the stepped diffusers over a design space (grid of possible solutions) where the well depths were restricted to the set of integers between zero and 16. The optimization algorithm navigated through millions of possible solutions on the grid to converge to the designs you see here... but consider this: because the optimization algorithm used a low resolution design space, the designs can be built using low precision. A variance of 1/2 to 1 mm is perfectly acceptable.

Good question!
16th October 2012
#79
Gear interested

Building them

Hey! This is a great thread!

Started building, so I thought I would share my aproach.

Building them out of EPS. Seamed cheap and easy. Built a hot wire cutter in no time and started cutting. Then glued the pieces together and laminating them with paper (the paper thing I found out about in an RC Plane building video, supposedly makes the EPS a lot stronger and harder, we will see about faster )

Then i will paint them, this should give them a hard surface. Anyway I've had fun so far =)
16th October 2012
#80
Lives for gear

.
Hey, THANKS, Arqen!
I have control room small enough to make it problematical to use most of the diffusers I've seen, and this looks like just the ticket.

I had just about decided it wasn't worth it to go to all the time and expense of building diffusers (only to have them make my CR even smaller), but I think I'll give this a try!

...By the way, has anybody tried the fractals on a ceiling over drums?
.
17th October 2012
#81

Quote:
Originally Posted by Onlyone
Hey! This is a great thread!

Started building, so I thought I would share my aproach.

Building them out of EPS. Seamed cheap and easy. Built a hot wire cutter in no time and started cutting. Then glued the pieces together and laminating them with paper (the paper thing I found out about in an RC Plane building video, supposedly makes the EPS a lot stronger and harder, we will see about faster )

Then i will paint them, this should give them a hard surface. Anyway I've had fun so far =)
Cool. Glad you found them so cheap and easy to built, and thanks for posting photos!
17th October 2012
#82
Gear maniac

Only one ,

Make sure you let us know how you think they effect the sound in your room. Haven't seen any feedback yet from anyone who has them actually installed and operating.
17th October 2012
#83

Quote:
Only one ,

Make sure you let us know how you think they effect the sound in your room. Haven't seen any feedback yet from anyone who has them actually installed and operating.
No detailed feedback yet, but it sounds like Schaap has done a basic listening test with the stepped diffuser A1-LF and the fractal version of it, A1-Frac:

Quote:
Originally Posted by Schaap
Today I 'compared' the fractals(1") - flat - with QRD's of 1" , so called 'Flutterfree" and by hearing I couldn't experienced much difference.
When the fractals were modulated by 1,2,0,2,1 it took the diffusion to a higher level and so very noticeable difference.
In the meantime I've finished the fourth module

Henk
17th October 2012
#84
Lives for gear
Thanks OnlyOne! That got me thinking about wooden diffusors and how I might build some of them, too, like what you are doing. I am sure your idea will add improvement in your room(s), and mine, too, now that you have shared your good idea. I really appreciate this thread and I also keep in mind that walking around listening for where to put "which" mic is the best policy, in any room....it is also the most fun for me. My mics and pres keep getting better, and my rooms will as I work on them, too. It's the best take that will do it for me, trusting that I have done the preliminaries -my home work. Michael
17th October 2012
#85

Quote:
Originally Posted by mcintalker
Thanks OnlyOne! That got me thinking about wooden diffusors and how I might build some of them, too, like what you are doing. I am sure your idea will add improvement in your room(s), and mine, too, now that you have shared your good idea. I really appreciate this thread and I also keep in mind that walking around listening for where to put "which" mic is the best policy, in any room....it is also the most fun for me. My mics and pres keep getting better, and my rooms will as I work on them, too. It's the best take that will do it for me, trusting that I have done the preliminaries -my home work. Michael

On the topic of listening tests, it would be great if someone did an A-B test. E.g., if someone recorded the sound at various locations in a room with and without the diffuser.
18th October 2012
#86

Quote:
Originally Posted by 12ax7
.
Hey, THANKS, Arqen!
I have control room small enough to make it problematical to use most of the diffusers I've seen, and this looks like just the ticket.

I had just about decided it wasn't worth it to go to all the time and expense of building diffusers (only to have them make my CR even smaller), but I think I'll give this a try!

...By the way, has anybody tried the fractals on a ceiling over drums?
.
Great!

These diffusers were actually designed so they can be used in smaller rooms (I optimized them using a listener distance of 2.5 m, where typically, diffusers are rated for use at 5 m). Of course, these diffusers perform at larger distances as well, but they're quite efficient at spreading sound out in small spaces.

Glad to hear you're thinking of building these.

Tim
19th October 2012
#87

FDTD Room Acoustics Model by Jonathan Sheaffer

Quote:
Originally Posted by AwwDeOhh
Thanks, i appreciate it.
Was mainly interested in modeling how several diffusors interact with each other in multiple points around the room (ie, from back wall, to side wall to ceiling, ect.). though the Reflex software interests me as well.
Is k-wave/Matlab capable of doing 3-dimentional models? if not, any suggestions?

You might want to check out this new sketchup plugin and Matlab toolbox by Jonathan Sheaffer (vid below). Looks like a great tool that makes finite difference time domain (FDTD) simulation more accessible!

20th October 2012
#88
Lives for gear

Quote:
Originally Posted by Arqen
You might want to check out this new sketchup plugin and Matlab toolbox by Jonathan Sheaffer (vid below). Looks like a great tool that makes finite difference time domain (FDTD) simulation more accessible!

Hey thanks. that looks promising!
Good find!
20th October 2012
#89

Quote:
Originally Posted by AwwDeOhh
Hey thanks. that looks promising!
Good find!
No problem! Happy to pass on useful tools.
24th October 2012
#90
Lives for gear

Quote:
Originally Posted by Arqen
On the topic of listening tests, it would be great if someone did an A-B test. E.g., if someone recorded the sound at various locations in a room with and without the diffuser.
Hi Tim et all,

Today I made a recording with a Olympus LS-11 mobile recorder in a room of about 85m3 at the same spot and conditions. I had 5 metric fractal modules laying on the floor next to each other and the music arrived to the diffuser in a indirect way, so not direct from the speakers. Not optimal circumstances to judge(maybe later) but I think it gives the 'essences' op diffusion and the working of the part of the optimized diffuser if you're have decent monitoring.
Music is used for educational purpose from a CD and is not mine. Hope that's not a problem.

The modulation sequence of the modules were 0,4,5,3,5 Total: 1.2 m2
Attached Files

Arqen Without - a1.wav (3.60 MB, 1993 views)

Arqen Modulated 5 a1.wav (3.57 MB, 1996 views)

Mentioned Products

Topic:

## Welcome to the Gearslutz Pro Audio Community!

###### Registration benefits include:
• The ability to reply to and create new discussions
• Interact with VIP industry experts in our guest Q&As
• Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD \$20/year
• Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.