[bassboy]

Due to my fairly large room, small diffusors won't help much.  I will need lots of diffusion area.  I'm thinking about 60 sq feet per unit, with a minimum of 5 units in the room.

My research leads me to believe that 1 large diffusor will cover a given area better than an array of smaller diffusors.  Lots of small diffusors cause lobing problems, apparently.  This is easily overcome by designing a single diffusor based on a very large prime number and using a well width which correspond to the desired finished dimensions.

Ok, so that's easy enough when dealing with a 1d diffusor, but I have not found a calculator that will design a very large 2d (skyline) diffusor, the option being to make an array of smaller units to cover the desired area.  I assume that lobing is still an issue with arrays of 2d diffusors.  I've seen examples of patterns that can be used for placement of units within an array of 1d diffusors to help with the lobing issue, but not sure how these patterns would apply to an array of 2d units.

So, am I better off making a single (very large) 1d type of diffusor, or using many units of skyline diffusors to cover the same area?  And if so, how should they be patterned?  And is there a way to design a single large area 2d diffusor, like the large prime number 1d example?

[Jason Jones]

I'm not sure how large your room is, but the fact that you are thinking of putting 300 square feet or more of diffusion does give me some idea. In a room that large, I would probably not put a 2D diffuser. 2-D diffusers have an advantage when you have to be close to the reflecting surface, because by diffusing into more directions they will attenuate the reflected energy more. However, 2-D diffusers are not as effective at low frequencies as an equivalent 1-D diffuser. The only reason to use a large 2-D diffusor in a room that large is because you think they look cool - and if you are spending much time in the room, looks do matter.

No you will not find a recipe for a 1-D diffuser that large on the internet. I did actually design a pretty big one for a guy on a forum a month or two ago. It involves the integration of a few abstract mathematical methods. It isn't really hard for me to do, but is fairly time consuming.

If you go the 1-D route, using several smaller 1-D diffusors next to each other will cause lobing issues. However, this is actually easily solved. You can alternate the diffusors so that one is horizontal and the one next to it is vertical in a kind of checkerboard pattern. That will actually work very well. If you want a real no compromise set-up, you can add one more step. This would be making two different diffusers based upon different prime numbers. People commonly use 5 and 7, but you could certainly go higher - no real point going beyond 11 or 13, though. Then you would use a combination of the different prime number diffusers in the checkerboard pattern above. Which prime for which spot would be chosen by a simple modulation scheme which I would be more than glad to help you with - sounds complicated but really I could walk you through it no problem.

Jason

[bassboy]

Since 1d is easier to make, I'm happy that you think it would work best, or at least not worse than 2d.

I had seen some commercial models that were very large (ie the one in the multipage diffusor thread currently active in this forum) and they use single large prime number sequences when possible instead of smaller ones, so I was thinking I could do a prime 89 or larger as a single panel (from the instructions on the mlsoft page) BUT

if you have ideas on something that would work better I'm certainly more than a bit interested.  All the theory I`ve read is from a few articles I could find, and some of them are contradictory.  This is going to be a big job and I'd like to do it properly the first time.  I am open to the original Schroder (sp?) with fins, or the optomized design without fins (although I haven't found the paper to figure out how to design one yet, and I've read evidence that the optomized design has little benefit over the original when large prime numbers are used).

The room is only about 27 x 19.  I was thinking about cutting it down significantly with a dividing wall, but I'm not going to.  This may be a bit much diffusion for the room size, but I wanted to use the diffusors to completely cover an existing large window and double door, so about 8 feet x 8 feet per unit should be about right.  I want one for each wall and ceiling, to cover the first reflection points as well as possible.

Any ideas you can share will be greatly appreciated.  And if you feel like explaining why there is no benefit to going further than a prime 13 design, I'd like to know what goes into these things, as I'm severely math challenged and quadratic theory is kind of lost on me when viewed through equations.  Is it just because there`s no lobing when properly patterned?

Thanks for your time, it`s very much appreciated.  I`m open to anything you have to suggest.

[bpape]

Just FYI - Jason is the person who designed our new D1 diffusor.  It's been an excellent product.  He's very well versed in the science but not afraid to think outside the box.

Bryan

[bassboy]

No doubt in my mind - that's why I'm asking here!  Everyone seems to have a long list of credentials behind them, except me, lol.

I'm open to anything he has to suggest.

[Jason Jones]

Is this a 2-channel room or 5 or home theater or a performance room?

Jason

[Jason Jones]

Hmmmm......when you are designing a diffusor array for a room there are trade-offs for every individual product you use and they change, sometimes dramatically in combination and depending on where in the room and size and etc. etc. etc. I think I will have to respond in bit-size chunks to avoid writing something that is confusing and causes me carpal tunnel syndrome. 

First, what we are trying to do with the diffusor is to keep some liveliness to the room and give a sense of spaciousness without changing the tonal balance (i.e. comb-filtering). The way to do this is to control your reflections in such a way that you can tell where the reflections are coming from generally, i.e. behind you or to your right and behind, etc. But, where you cannot pick out any particular reflection's distance or specific location.

A concert hall does this well, because you have so many different reflections coming from many, slightly different directions and from many different distances. This gives us exactly what we want in a smaller room, except that we don't want quite so long an RT60 (reverberation time) and we don't have the luxury of that much real estate.

Let me know if you are with me so far - and any questions - and then from there we can get to the specifics and why this or that particular choice.

Jason

[bassboy]

Jason, this will be mainly a 2 channel room, but 5 channel is fun sometimes too.  But I would prefer to design around a 2 channel system.

I am definitely with you so far.

I am not set on any particular style of diffusor or size.  The ONLY restrictions I would like to place at this time are that I would like to cover the existing door and window.  The door is centered on the short wall (19 feet) and the window is somewhat centered on the long wall (I haven't done exact measurements yet but it looks centered, and the long wall is 27 feet).  The window is only about 4 x 4 feet, so the side and ceiling diffusors don't necessarily have to be quite as large as I mentioned previously.  The ceiling is 10.5 feet tall.

I'll do some accurate measurements and post the results tomorrow.

Thanks again.

[Jason Jones]

I'll do some accurate measurements and post the results tomorrow.

Thanks again.

Sounds good - I always have fun working on these custom designs!

Jason

[bassboy]

Ok, this is the best I can do with my too short, broken tape measure, all by myself.  (If you need more accurate measurements than this, let me know and I will get a guy that does measurements and drawings for a living, but it may take awhile).

First, see my crude drawing.



The short hall to the side is not to scale, (actually nothing is in this crappy ms paint depiction).  The doorway opening is about 4 feet wide.  This hallway is 7 feet long and needs to remain open and unblocked.  The notch in the back corner of the room is not square, and it protrudes into the corner about 6 feet x 3 feet.

The red areas show where the window and door are located.  These are the ones that need to be covered.  The door frame is 5.75 feet wide by 8 feet tall, and is not quite dead center, it's center is offset by about 8 inches between the 2 walls.  The window is about 56 x 56 inches and it's center is offset about 5 inches between the front and back wall (a bit closer to the front wall).

The green lines show where the ceiling begins to slope down.  The big flat area in the center of the ceiling is 10.5 feet and it slopes down to about 8.5 feet at the edges of the sidewalls.

The room is finished (badly) in drywall, the ceiling is 1/4 or 1/8 inch OSB (cheap chipboard), the floor is concrete.

I apologize about the quality of the picture, I don't have a scanner and ms paint is the only drawing program I know how to use.  Is this enough info to start?  I can get even more accurate measurements and a much better drawing, but that may take awhile, as I mentioned.

[bassboy]

EDIT - this is my second consecutive post, if you missed the last one on page one, it contains measurements of my room.

Additional details - I've only been in this room for a few weeks now.  It's currently being insulated.  It is a garage, and as such, there is an insulated garage door on the opposite wall to the door I want to cover.  Not sure what I'm going to do about that yet, but it will probably be filled in as a solid wall in the future.  For now, I just want to stick a big diffusor in front of it.

I have about 450 sq feet of nice carpet.  It's not enough to cover the entire floor, but I should get almost 2/3 coverage.  I was thinking about covering the back and leaving a bit of the front of the room uncarpeted, with a wood or tile finish.  I will need to have a corn stove (similar to a wood stove, if you are not familiar) as the only source of heat for at least a few months of the year.  It will be located somewhere in the back half of the room.

The speakers are dipole hybrid - dipole woofer with compression tweeter (Hawthorne 15 inch coaxials).  Any woofers to be added will either be dipole, dipole hybrid or IB.  No boxes in this system, although I am going to be working with single driver fullrange horns at some point as well.

Someday I will have a projector and an IB subwoofer system, but not anytime soon.  At this point it still looks very much like a garage, since I've done nothing with it yet, but I have plans to make this the nicest room in the house, although it may take years to address the frivolous visual appeal, I want to take care of the sound first.

I plan on using extensive bass trapping as well, many varieties and sizes but overall huge sq footage.  Some the high bass traps will address absorbtion of higher frequencies as well, if they are placed properly.  That's the plan anyway, I don't plan on needing any stand alone high frequency absorbtion other than what is intrinsically built into the bass and diffusion treatments.

I thought diffusion would be a good place to start wrt treatments for this room.

[Jason Jones]

Your drawing is fine. I'll try to get something up later today. But, I have a pretty busy Saturday ending with a halloween party where I am supposed to dress up in some obnoxious costume - maybe I'll go as an acoustics geek? 

Jason

[bassboy]

There's certainly no rush, I'm busy with seasonal issues as well.  It's a busy month.

No costume party for me this year, but my neice (2.5 yrs) is going to be a tiny little lion.  It's SO cute.

[Glenn K]

Your drawing is fine. I'll try to get something up later today. But, I have a pretty busy Saturday ending with a halloween party where I am supposed to dress up in some obnoxious costume - maybe I'll go as an acoustics geek? 

Jason

Put on a white shirt, glasses and carry a CAT. 

[Jason Jones]

Picking a diffuser prime number:

OK, so why 11 or 13? First, the biggest problem in diffuser design is to make the product work the same at all frequencies. For example, we don't want it to be great at 1 khz, but specular (like a bare wall) at 10 khz and completely absorptive at 500 hz. The standard QRD with 7 wells has a big "flate-plate" frequency problem.

The Flat Plate Frequency Problem:

Lets imagine that we have a diffuser with three wells. One well is 1" deep, the second well is 2" deep and the third well is 3" deep. Now lets imagine that sound is a 1" deep block of wood. Lets put one block in the first well (notice it is full to exactly the top). Let's put two blocks in the second well (notice it is full to exactly the top). Finally, let's put three blocks in the third well (notice it is full exactly to the top).

So, we now have a diffuser in which the blocks (representing reflected sound) are all at the same height coming out of the diffuser. With the blocks (sound waves) matched up, they won't interfere with each other. This interference is what we are counting on - it is what makes the diffuser work.

To find the "flat plate frequency" you multiply the number of wells by the lowest frequency that the diffuser is designed to work at. For a common QRD with 7 wells that would be: 7*500 hz = 3500 hz. So, at 3500 hz (around the frequency of chalk board scratching and babies crying), this 7 well diffuser is behaving similar to a 'flat plate', or bare wall. If we move up to 11 it would be 5500 hz and at 13 it would be 6500 hz.

Now I know you are thinking, "isn't that just more reason to use a bigger prime like I said the first time!" I'll get to that next - in the mean time I actually do hear a baby crying.

Jason

[Ethan Winer]

Put on a white shirt, glasses and carry a CAT. 

Hey, I resemble that - literally!

[bassboy]

Let me begin by expessing how much I appreciate your time and (easy to understand) answer.  My research so far has led me to the papers at the top of the google search list for diffusors, and it would seem at this point that they are overly simplified, as your answer is a bit different than the ones I've been reading about.

For example, from "Diffuser design, Trevor Cox 2004-5, (c) University of Salford, www.acoustics.salford.ac.uk"
"Optimum diffusion only occurs at the design frequency (or a multiple of the design frequency).
At frequencies in between the scattering is not optimum."

And "we want a QRD with the narrowest wells possible to get the widest frequency range."

Unless I am missing something, Cox, never mentions that it is possible to optomize the diffusor over it's usable bandwidth, and there is no mention that the high frequencies have anything at all to do with the number of wells, only the width of the wells.

Also, one of the articles I can't find right now was claiming that prime 7 was some type of "magical" number, based on physical vs apparent well depth, but no mention at all of plate frequency issues.

So again, thanks for this, as you have explained a few things I haven't found in ANY of the papers I've read so far (which probably means I just need to read a lot more).

[bassboy]

Ok, nevermind, I found it, or at least a reference to it.  The BBC diffusor paper states that a 860 hz prime 17 diffusor has an upper limit of 13.76 khz, so I have seen this before but didn`t recognize it since they either didn`t tell how they came up with that or it`s hidden in a bunch of math I don`t understand.  At least I didn`t recognize the (incredibly easy to use) formula you posted.

Also, by simply using an online diffusor calculator that only asks for width, depth and prime number and does the rest for you, I wasn`t learning much.  I'm having a "dumb" day I guess.

Either way, regardless of whether or not I`ve seen it before you explained it very well, thanks for that.

[Jason Jones]

D'Antonio and Cox do mention the "flat plate" problems in their research, I think they usually call it a "critical frequency" - "flat plate" seems more descriptive to me.

I was going to go into a long explanation about well widths and absorption problems, etc. But, I'll just boil it down and feel free to ask for any clarifications. Here are the rules:

1) No well widths under 1".
2) Make absolute, 100", complete, unwavering surety that your wells are sealed. i.e. if you put your diffuser on it's back on the ground and you poured water into the wells, it would not leak behind the well or into any other well.
3) Don't cover the diffuser in cloth.

These are straight from the D'Antonio and Cox book. If you do any of the above, you will end up with an expensive, peaky ABSORBER, with a little bit of diffusion.

Jason

[Jason Jones]

Where are your speakers and where is your seating position?

Jason