for home made diffuser....

[SouthSIDE Glen]

Not to poo poo on a pretty good post, but I was under the impression that it was WAY more complex than that. I thought there was some kinda crazy algorithm that determined the length of each block. Or am I mis-remembering?

I got very little sleep last night, and it's all your fault, NL5 . I decided to do some research and reading up on skyline diffuser panel design.Here's what I get out of it thus far:

Yeah, many companies have filed patents on fancy prime number-based algorithms for determining block size. Mostly these algorithms are used as pseudo random number generators to help ensure a fairly natural randomness in block size and to help ensure uniqueness in block size.

While this is very nice, it's not absolutely necessary to achieve serviceable diffusion. It's kind of a measure of smoothness of the diffusion. For the best in smooth diffusion, you can spend $150 for a 2x2 panel of stiff foam that follows a patented randomness design. For effective enough diffusion that may not have the absolute ultimate in smoothness, but for most applications will be effective enough to effectively reduce or eliminate the actual first reflection and comb filtering problems that such diffusers are meant to address, you can build your own without having to use a fancy patented algorithm.

This board, besides depriving me of sleep, is constantly giving me ideas for projects to work on. It seems to be that writing a little program/applet that generated a quite effective cutting plan for building a simple random skyline diffuser panel without violating anybody's patents should be pretty easy to develop. I'll add that to my to do list and hopefully get to it someday...

G.

 

[Sonixx]

for a small room and specifically with similar dimensions or dimensions that are divisible by a common divisor, the Bottom is very problematic and depending on locations, you can have a lot or very little bottom reinforcement. i.e. you can't trust what you're hearing. the room just does not want to support the longer wave length frequencies evenly. modes are not sufficiently close and in enough quantity.

overall, you need a combination of Bass treatment, Absorption and Diffusion. The Bottom will be the biggest challenge.

back to diffusion, I have built Quadratic Residue Diffusers. below is a small one. the depth determines the bottom frequency diffused.

you really can't expect to get much from just putting up irregularly shaped patterns on the wall.

Here's some larger QRD in action next to 4" deep absorbers.

 

[SouthSIDE Glen]

OK, here's a specific plan for making a skyline diffuser using nine standard (American) 2'x2s and a piece of plywood using the same "primitive root" algorithm used by the Big Boys.

The effective low frequency handled by such diffusers is determined by well depth (block height). The effective high frequency is determined by well/block width.

Since the effective actual width of each side of a 2x2 in the States is 1.5", this makes the effective upper frequency diffused by such block sizes to be about 4.5kHz. Not too bad. This wont diffuse sibilance, but it covers most of the harsh upper mids below 5kHz.

For the low frequency, you can adjust the lengths of the blocks - the longer they are overall, the lower the lowest diffused frequency. The following numbers cover all the way down to 500Hz. Most of what lies below that should be trapped instead of diffused anyway.

To make a diffusion panel about 19.5" x 18" (roughly) that covers this low of a frequency will require a total of just over 87 feet of 2x2, or nine 8ft lengths. Here's how to cut those 2x2s into the size chunks you need for the diffuser. The first number is the length of the chunk in centimeters, the second number is the number of chunks of that length you'll need:

0 cm: 2
1 cm: 4
2 cm: 5
3 cm: 5
4 cm: 4
5 cm: 5
6 cm: 4
7 cm: 5
8 cm: 5
9 cm: 4
10 cm: 5
11 cm: 4
12 cm: 5
13 cm: 5
14 cm: 4
15 cm: 5
16 cm: 4
17 cm: 5
18 cm: 5
19 cm: 4
20 cm: 5
21 cm: 4
22 cm: 5
23 cm: 5
24 cm: 4
25 cm: 5
26 cm: 5
27 cm: 4
28 cm: 5
29 cm: 4
30 cm: 5
31 cm: 5
32 cm: 4
33 cm: 5
34 cm: 2

Note that there are a couple of wells of 0 centimeters. Unless you want to leave unsightly holes in your skyline where the 0 cm wells go, you can mount all these blocks on a backing piece of plywood or hardwood sheet.

Finally, we need to know the randomized order in which these blocks are mounted together into the final checkerboard. Here's the pattern determined by the prime root algorithm (each number representing the length in centimeters of each block):

1 33 15 16 8 17 13 5 15 30 5 7 5
24 5 28 8 11 7 19 32 23 7 12 26 3
14 17 27 2 5 21 2 25 24 12 34 28 28
4 2 18 33 8 25 4 9 24 19 25 32 3
13 18 12 20 31 4 23 21 9 16 28 20 23
14 31 24 26 31 18 20 11 1 13 12 6 34
33 1 19 18 26 17 21 30 19 4 29 27 29
10 29 7 26 23 27 15 2 11 27 22 8 31
20 17 7 33 29 13 32 9 10 22 0 6 6
30 32 17 1 26 9 30 25 10 15 10 2 31
21 16 22 14 3 30 12 13 25 18 6 14 11
20 3 10 8 3 16 15 23 33 22 22 28 0

If you want to play with the numbers some more, you can find an online calculator here.

G.

 

[Ethan Winer]

I thought there was some kinda crazy algorithm that determined the length of each block.

Not crazy at all, and definitely not random. Download the Acoustic Calculator here:

Acoustic Calculator

Much more about diffusion here:

All About Diffusion

--Ethan

 

[SouthSIDE Glen]

Not crazy at all, and definitely not random.

Well, certainly pseudo-random (pure randmoness cannot be synthesized). There's importance to the limiting numbers, they are not just random, but the whole point as to the distribution of the numbers in a skyline diffuser is to provide (close enough to) random distribution of the well depths and angles of incidence.

G.

 

[Myriad_Rocker]

Fantastic! What's a skyline diffuser, though? Does it go on the rear reflection wall from the mix position?

OK, here's a specific plan for making a skyline diffuser using nine standard (American) 2'x2s and a piece of plywood using the same "primitive root" algorithm used by the Big Boys.

The effective low frequency handled by such diffusers is determined by well depth (block height). The effective high frequency is determined by well/block width.

Since the effective actual width of each side of a 2x2 in the States is 1.5", this makes the effective upper frequency diffused by such block sizes to be about 4.5kHz. Not too bad. This wont diffuse sibilance, but it covers most of the harsh upper mids below 5kHz.

For the low frequency, you can adjust the lengths of the blocks - the longer they are overall, the lower the lowest diffused frequency. The following numbers cover all the way down to 500Hz. Most of what lies below that should be trapped instead of diffused anyway.

To make a diffusion panel about 19.5" x 18" (roughly) that covers this low of a frequency will require a total of just over 87 feet of 2x2, or nine 8ft lengths. Here's how to cut those 2x2s into the size chunks you need for the diffuser. The first number is the length of the chunk in centimeters, the second number is the number of chunks of that length you'll need:

0 cm: 2
1 cm: 4
2 cm: 5
3 cm: 5
4 cm: 4
5 cm: 5
6 cm: 4
7 cm: 5
8 cm: 5
9 cm: 4
10 cm: 5
11 cm: 4
12 cm: 5
13 cm: 5
14 cm: 4
15 cm: 5
16 cm: 4
17 cm: 5
18 cm: 5
19 cm: 4
20 cm: 5
21 cm: 4
22 cm: 5
23 cm: 5
24 cm: 4
25 cm: 5
26 cm: 5
27 cm: 4
28 cm: 5
29 cm: 4
30 cm: 5
31 cm: 5
32 cm: 4
33 cm: 5
34 cm: 2

Note that there are a couple of wells of 0 centimeters. Unless you want to leave unsightly holes in your skyline where the 0 cm wells go, you can mount all these blocks on a backing piece of plywood or hardwood sheet.

Finally, we need to know the randomized order in which these blocks are mounted together into the final checkerboard. Here's the pattern determined by the prime root algorithm (each number representing the length in centimeters of each block):

1 33 15 16 8 17 13 5 15 30 5 7 5
24 5 28 8 11 7 19 32 23 7 12 26 3
14 17 27 2 5 21 2 25 24 12 34 28 28
4 2 18 33 8 25 4 9 24 19 25 32 3
13 18 12 20 31 4 23 21 9 16 28 20 23
14 31 24 26 31 18 20 11 1 13 12 6 34
33 1 19 18 26 17 21 30 19 4 29 27 29
10 29 7 26 23 27 15 2 11 27 22 8 31
20 17 7 33 29 13 32 9 10 22 0 6 6
30 32 17 1 26 9 30 25 10 15 10 2 31
21 16 22 14 3 30 12 13 25 18 6 14 11
20 3 10 8 3 16 15 23 33 22 22 28 0

If you want to play with the numbers some more, you can find an online calculator here.

G.

 

[SouthSIDE Glen]

Fantastic! What's a skyline diffuser, though? Does it go on the rear reflection wall from the mix position?

A skyline diffuser is simply a diffusion panel made in the manner I describe; a checkerboard pattern of blocks of varying depth. It's called "skyline" because if you lay it flat, it kind of resembles a big city downtown skyline:

Diffusers - and there are more types than just the skyline; there's the "Quadratic" diffusers like the kind Sonixx shows in his post, and parabolic diffusers like the kind I describe in an earlier post as well (Ethan has a picture of one of those on his site) - are probably found most often on the rear walls, but they can be used any place where you want to break up or reduce direct reflections but don't necessarily want to deaden the room even more by using an absorption panel there.

G.

 

[Myriad_Rocker]

A skyline diffuser is simply a diffusion panel made in the manner I describe; a checkerboard pattern of blocks of varying depth. It's called "skyline" because if you lay it flat, it kind of resembles a big city downtown skyline:

Diffusers - and there are more types than just the skyline; there's the "Quadratic" diffusers like the kind Sonixx shows in his post, and parabolic diffusers like the kind I describe in an earlier post as well (Ethan has a picture of one of those on his site) - are probably found most often on the rear walls, but they can be used any place where you want to break up or reduce direct reflections but don't necessarily want to deaden the room even more by using an absorption panel there.

G.

Gotcha. Thanks.