64 Cell (8x8) PRD (Skyline Diffuser) question

kareface · Aug 13, 2010

I prefer the scattering on the PRD's over the QRD's, but the dimensions I have to work with won't factor into a traditional PRD. For one reason or another it has to be 8 rows & 8 columns. I was thinking of generating one that's a little larger and just clipping off the extra cells. I know this will have an overall negative impact, but considering how random the patterns of the PRD's tend to be will it make any audible difference?

One other question, I was planning on incorporating the diffuser into a larger structure, so I would actually make 4 of the diffusers and place them in a pattern like so:

(O's = Diffusers, X's = No Diffusers)
XOX
OXO
XOX

Would it be better to generate a huge skyline pattern that would encompass the entire structure and just segment out the 4 diffuser patterns from the larger design? I doubt that it'll make any difference, but I'm curious none the less to hear peoples opinion.

Bulldogger · Aug 13, 2010

I have used PRD diffusers from two different sources, RPG and Waveguide Acoustics which appears to be out of business. The Waveguide stuff uses larger protrusions and is less complicated than the Skylines. However of the two, I preferred it. The Skylines are very effective but can tend to brighten the sound. In fact I purchased 10 of mine, I have about 30 from a seller that did not like this effect. QRD diffusers do not appear to be nearly as effective in my room. I would not be surprised if even though your "pattern" may be more simple, it would still be superior to a QRD diffuser. The placement of the any acoustical treatments should be directed at first reflection points primarily. That is where I would start. In my room I had most of the Skylines on the ceiling. The Waveguide diffusers were on the front wall between the speakers and on the back wall covering the upper half. I used absorption/bass traps in the corners and 2 inch OC 703 at sidewall first reflections. A combination of absorption and diffusion worked best for me. My room was almost the exact dimensions as the the person from whom I purchased some of my diffusers. He had Rives Audio model his room. I used that model as a starting point from my own room.

audioguy · Aug 13, 2010

Bulldogger said:
The Skylines are very effective but can tend to brighten the sound.

My back wall is Skylines and I have an upper midrange glare the cause of which we have not been able to locate (either Rives, who designed the room, or another company). Last evening I covered all of the Skylines with 2 inch 703 and, unfortunately, it did not solve my problem even a bit.

kareface · Aug 13, 2010

I've noticed this problem as well, this is why the X's in the matrix above are dampening. I was fortunate enough to hear a room with small diffusers and absorption alternating every foot and it sounded much better than strict rear diffusion in the same space. The pattern wouldn't be dramatically different, as 9x8 is a PR capable dimension, so I'd just be removing 1 column.

terryj · Aug 14, 2010

kareface said:
I prefer the scattering on the PRD's over the QRD's, but the dimensions I have to work with won't factor into a traditional PRD. For one reason or another it has to be 8 rows & 8 columns. I was thinking of generating one that's a little larger and just clipping off the extra cells. I know this will have an overall negative impact, but considering how random the patterns of the PRD's tend to be will it make any audible difference?

One other question, I was planning on incorporating the diffuser into a larger structure, so I would actually make 4 of the diffusers and place them in a pattern like so:

(O's = Diffusers, X's = No Diffusers)
XOX
OXO
XOX

Would it be better to generate a huge skyline pattern that would encompass the entire structure and just segment out the 4 diffuser patterns from the larger design? I doubt that it'll make any difference, but I'm curious none the less to hear peoples opinion.

The maths of a PRD are what they are.

change them, well you are out on a limb. It will probably scatter, but it will not be a PRD.

If you want to navigate uncharted waters, then make an 8*8 'something'. It won't be a PRD tho.

Maybe tell us why it has to be an eight by eight??

You have made no mention of frequencies you wish to diffuse (which makes an arbitrary 8*8 kinda, well, arbitrary)

Do you mean you have a certain area you need to fill?? I don't understand why you ays it must be 8*8.

You need to supply more data to get better information.

kareface · Aug 14, 2010

I'm not sure why but your post seems oddly aggressive. It might just my interpretation but I don't see any reason to be.

The math for the PRD's wouldn't be changed, I would just be removing a column. I don't think the reason why the structure has to be 8x8 has any impact on the question so I left out unnecessary information. The way I'm building it limits its dimensions and I'm designing it based on another setup I've experienced that I know turned out well. 8x8 happens to be the dimensions I have to work with. I asked this question in a few different forums to see if I could get a reply from someone who has more experience with designing diffusers, I did manage to get a PM from an acoustical engineer who said it'll be largely inaudible.

terryj · Aug 14, 2010

well then I agree with you. Odd, cause it was no way meant to be aggressive!

Again I agree, no need for an aggressive answer (not that I meant to do one)

A PRD is based on prime numbers. A common one is prime 157. I have just built four, and intend to build a few more.

The PRD uses 'prime minus one' number of cells. Hence, in my prime 157 build there are 156 cells/blocks whatever you want to call them. Therefore there will always be a different number of columns than rows, just the nature of the beast. In my case I use twelve rows and thirteen columns (12*13=156)

You want to use 8*8,which equals 64. Add the one and we get 65, which is not a prime number. There (for starters) will not be a calculator on the planet that you could use to calculate the depths, simply because it is an invalid number to plug into those calculators.

So the maths will not be correct, it will not be a PRD.

You CAN go a 2d QRD, maybe that is what you were thinking of?? They are square, but again not would not be an 8*8 (as again they too are based on primes, so you bould go 7*7 or 11*11 etc etc)

The reason I asked about 'why?' 8*8 is that you may have been after an 'exact square' shape externally. There is nothing to stop you adjusting one dimension to compensate (ie slightly rectangular cells rather than square cells) and end up with a finished square design.

That is why I asked for 'more data'.

Have a look at QRDUDE. there is a lot of explanation behind the theory etc, tho it concentrates on QRDs rather than PRDs. I use that calculator in conjunction with a normal PRD calculator to work mine out.

But, it may be worth your while to get the background of it all (rather than reading the 'bible'!! cox and d'antonio).

An eight by eight simply is not a PRD, and can never be. The maths won't work and does not work. As I said, you will have a nice looking scatterer but it will not diffuse as such. I don't see the sense (myself) in going to all that work to make something that might not be all that effective, and you will never know unless you measured it etc.

The nearest prime to 65 (as you are after 8*8) is 67, so the closest you could get that will fit the maths is 66 cells...but the only way to use that really is six by eleven!!

There is a reason that most PRDs follow a 'common theme', as squareness does seem to be a desired trait! Hence the majority are prime 157 (which means, in theory at least, every depth will be a unique depth, ie 156 DIFFERENT depths...but most cheat and only use a very small number of that)

Check out qrdude, maybe see if a 2dQRD will do the job for you. They are not as sophisticated in performance as a PRD, no free lunch in this world.

Why it has to be 8*8, for all I knew, could be based on some aspect that could be worked around...if I knew what it was. I did ask for more data so I could better answer and help you. It was not meant to be aggressive. Sorry.

Eight by eight, in itself, means nothing in the 'diffusor world' without dimensions, both the width of the 'block' and the depth. I'm sure you know (but I don't know if you know, you know?) that the width gives the upper cutoff frequency, the depth the lower.

That was another reason the eight by eight stipulation kind of stood out as odd, if you follow.

"" Does he have space limitations he needs to fit it into which is why he chose an arbitrary eight by eight??" kind of thing. That is why we need more data to help.

Steve Williams · Aug 14, 2010

OK i will ask the dumb question.

Could you define QRD & PRD

DonH50 · Aug 14, 2010

QRD = quadratic-residue filter
PRD = primitive root diffuser

Different mathematical expressions used to derive the design of diffusers. I believe Manfred Schroeder was the first, or among the first, to suggest diffusers based upon mathematical models. There are various other models and types of diffusers, e.g. reflection-phase grating, diffraction-grating, etc. PRD is called pseudo-random diffuser by some folk, but the original PRD term was derived from root equations (thus the name) and a pseudo-random diffuser is something a little different. The goal of all of them is to diffuse the sound so that no single tones "build up" in the listening area.

HTH - Don

terryj · Aug 14, 2010

http://www.subwoofer-builder.com/qrd.htm

as I said, mainly about QRDs, but as a first step in understanding it may help.

In the calculator, you can model 2dQRDs (a normal QRD are the ones with fins running down them, a 2d qrd looks more like the skyline diffusors we have seen) which are naturally square in dimension, the PRDs (unless you account for it) will have a different number of rows than columns, so not naturally square.

Only mentioning square as it may be part of the reason for 8*8.

here is the QRD calculator

http://www.subwoofer-builder.com/qrdude.htm

kareface · Aug 15, 2010

Sorry if I was mistaken, it just seems very 'in your face' to me for some reason.

terryj said:
The PRD uses 'prime minus one' number of cells. Hence, in my prime 157 build there are 156 cells/blocks whatever you want to call them. Therefore there will always be a different number of columns than rows, just the nature of the beast. In my case I use twelve rows and thirteen columns (12*13=156)

You want to use 8*8,which equals 64. Add the one and we get 65, which is not a prime number. There (for starters) will not be a calculator on the planet that you could use to calculate the depths, simply because it is an invalid number to plug into those calculators.

So the maths will not be correct, it will not be a PRD.

Yes, but that's only if I base the design strictly on the 8x8 pattern. If I use a pattern that produces a near match and only removed one of the columns it'll produce almost identical results (or so I'm told). I agree it won't be identical, I was only asking if the impact of removing a column would be audible, and the general answer seems to be it's unlikely.

On the subject of QRD's, I'm well aware of QRD's and I have no misconceptions as to which I am building.

terryj said:
An eight by eight simply is not a PRD, and can never be. The maths won't work and does not work. As I said, you will have a nice looking scatterer but it will not diffuse as such. I don't see the sense (myself) in going to all that work to make something that might not be all that effective, and you will never know unless you measured it etc.

I think you're operating under the assumption that I'm using the formula to create a 8x8 design, which is not the case. Instead I'm using a PRD pattern that produced a similar result and removing a column. I don't see how the removal of that column will prevent the majority of the frequencies and surface area that from sill being effective, I would love your insight on the matter. I will loose effective surface area and the conjoining column will likely not be as effective, but to say it won't diffuse seems a bit odd. If you are implying that it won't diffuse as evenly across the spectrum I would argue that no diffusers will, no matter how they're structured. Do you think there will be an audible difference between an 9x8 and a 8x8 diffuser using the same pattern only short a column? As I've mentioned I haven't see anyone besides yourself that seems to believe this to be the case, but I would love the input either way.

The nearest prime to 65 (as you are after 8*8) is 67, so the closest you could get that will fit the maths is 66 cells...but the only way to use that really is six by eleven!!

There is a reason that most PRDs follow a 'common theme', as squareness does seem to be a desired trait! Hence the majority are prime 157 (which means, in theory at least, every depth will be a unique depth, ie 156 DIFFERENT depths...but most cheat and only use a very small number of that)

Again, I understand the math. Instead of using a similar cell count I looked for one that produced a pattern closest to the intended design. As I mentioned I was only removing one of the columns, so the design I had in mind would contain 8 rows and 9 columns. From there I'm sure you can figure out the prime I used was 73. I wasn't trying to warp the equation to meet the criteria, that would be impossible. I was simply asking if A) Removing rows would create an audible impact and B) if it would be more effective to create a larger pattern and divide it between the 4 diffusers. So far the people I've talked with think both are no.

Check out qrdude, maybe see if a 2dQRD will do the job for you. They are not as sophisticated in performance as a PRD, no free lunch in this world.

As had been mentioned here and echo'ed else where, a truncated PRD would likely still produce better results than a QRD, which I would agree with not being the greatest fan of QRD results.

Why it has to be 8*8, for all I knew, could be based on some aspect that could be worked around...if I knew what it was. I did ask for more data so I could better answer and help you. It was not meant to be aggressive. Sorry.

No reason to be sorry, it's just something that I don't want to change. I liked the acoustic results of the product I'm modeling the design after and I think the overall benefit of having the design this way will outweigh the negative.

That was another reason the eight by eight stipulation kind of stood out as odd, if you follow.

I wasn't worried about the effective frequency range for the same reasons as you described. I'm well aware that the cell count doesn't effect the same elements as cell depth. I know what range and depth I plan to use, which is why I excluded it from the discussion.

I think this discourse can be largely written off as not seeing eye to eye. I'll be frank for the sake of mutual understanding. I don't need to know anything in relation to the design other than the impact of removing a column from the design on it's audibility. I'm studying physics, I'm familiar with acoustics, design sound rooms (some quite costly), have worked with and listened to many diffusers. I understand the difference between PRD and QRD, I understand the underlying math, I understand other elements of diffusion like well depth, listening distance and positioning. I'm only interested in peoples opinions on the subject of removing a column from the design and it's impact on it's acoustics. If anyone had first hand experience (unlikely but welcomed regardless) or if people just have an opinion on the matter.

You seem, unless I'm mistaken, to believe that it'll have a largely audible impact. I'm glad you mentioned it and I'm curious as to why you feel that will be the case. I honestly want to know what your thoughts on that matter are. From my perspective it would seem like a reduction in the effective surface area and a shift in the dispersion from the previously conjoined column. I can't imagine given the size of most wave lengths that the entire diffuser would be rendered moot, especially considering the random nature of the diffuser pattern. However, I would love to see others perspective on the issue. This is something I'm trying to gauge to decide if this it the route I plan to take.

Edit:

(which means, in theory at least, every depth will be a unique depth, ie 156 DIFFERENT depths...but most cheat and only use a very small number of that)

I thought I would add the measurements I've made so far have been accurate within +/- 1mm of the intended results. I'm not lazy about this stuff, lol.

terryj · Aug 15, 2010

kareface said:
it just seems very 'in your face' to me for some reason. .

Yes. I know exactly how you feel.

Yes, but that's only if I base the design strictly on the 8x8 pattern. If I use a pattern that produces a near match and only removed one of the columns it'll produce almost identical results (or so I'm told). I agree it won't be identical, I was only asking if the impact of removing a column would be audible, and the general answer seems to be it's unlikely.

On the subject of QRD's, I'm well aware of QRD's and I have no misconceptions as to which I am building.

To be fair to you, I 'missed' the very first sentence of your post, which does imply an understanding of the two. In any case, my 'long winded' post (based on missing that sentence) was only made because I wanted to help.

At the very least, even if redundant for you,it is there for others who may come across this at a later date. As are the links.

I think you're operating under the assumption that I'm using the formula to create a 8x8 design, which is not the case. Instead I'm using a PRD pattern that produced a similar result and removing a column. I don't see how the removal of that column will prevent the majority of the frequencies and surface area that from sill being effective, I would love your insight on the matter. I will loose effective surface area and the conjoining column will likely not be as effective, but to say it won't diffuse seems a bit odd.

It was not that clear (to an external reader) exactly what you were doing..we often type things and WE know exactly what we meant, after all we did type it, but sometimes it is not immediately clear to the reader. Which is why I (me, personally) asked for more data. Now that I know you will be making a 'proper' PRD and losing a column I would not have used the words 'won't diffuse'. I did get the idea you were making a 'prd-like' diffusor that will be 8*8. Any misunderstanding is probably completely due to me.

I don't have an insight, I am not an expert (and I hope I didn't sound like I thought I was), but MY view, as I hinted, would be something along the lines of 'if I'm going to do all that work I may as well do it properly...because I am NOT an expert'! But that is just me, and my gut feel is similar to yours...not optimal but probably no great harm done.

At the risk of offending you again, have you asked over at gearslutz??? http://www.gearslutz.com/board/bass-traps-acoustic-panels-foam-etc/ It is inhabited by the usual suspects haha. I imagine they would be conservative in their answer, something along the lines of 'we don't know, you'd have to measure it', but that is just a guess.

Do you think there will be an audible difference between an 9x8 and a 8x8 diffuser using the same pattern only short a column? As I've mentioned I haven't see anyone besides yourself that seems to believe this to be the case, but I would love the input either way.

As I said, I was under the impression you were building a lookalike. It will most definately be better than a bare wall, and I retract the 'won't diffuse' and replace it with (as we both know) 'not optimal, but way better than nothing'. (Especially if you are completely wedded to 8*8...which if so kinda makes the asking of the question pointless??)

As had been mentioned here and echo'ed else where, a truncated PRD would likely still produce better results than a QRD, which I would agree with not being the greatest fan of QRD results.

Can you expand a bit on this?? Sounds like you have been able to do a direct comparison, would love to hear your thoughts. You seem to echo that again in your next bit

No reason to be sorry, it's just something that I don't want to change. I liked the acoustic results of the product I'm modeling the design after and I think the overall benefit of having the design this way will outweigh the negative.

Maybe after you have done the truncating, you could somehow compare to the original and come back with a report??

I think this discourse can be largely written off as not seeing eye to eye. I'll be frank for the sake of mutual understanding. I don't need to know anything in relation to the design other than the impact of removing a column from the design on it's audibility. I'm studying physics, I'm familiar with acoustics, design sound rooms (some quite costly), have worked with and listened to many diffusers. I understand the difference between PRD and QRD, I understand the underlying math, I understand other elements of diffusion like well depth, listening distance and positioning. I'm only interested in peoples opinions on the subject of removing a column from the design and it's impact on it's acoustics. If anyone had first hand experience (unlikely but welcomed regardless) or if people just have an opinion on the matter.

Oh well, I may think twice in the future about making a long post in order to help. Being a slow typist and all.

But happy to take full responsibility if that helps.

You seem, unless I'm mistaken, to believe that it'll have a largely audible impact. I'm glad you mentioned it and I'm curious as to why you feel that will be the case. I honestly want to know what your thoughts on that matter are. From my perspective it would seem like a reduction in the effective surface area and a shift in the dispersion from the previously conjoined column. I can't imagine given the size of most wave lengths that the entire diffuser would be rendered moot, especially considering the random nature of the diffuser pattern. However, I would love to see others perspective on the issue. This is something I'm trying to gauge to decide if this it the route I plan to take.

I think this has been covered?? hope so.

Now, cmon, spill the beans!! How are you planning to build it??

Oh, back to your first post (the gaps). I presume these will be placed directly on the wall? (sounds a silly question, but just checking...the ones I have built are in open space but that's another story).

To me that 'drawing' is kinda misleading! I'll try and draw it's analogue.... O
O X O
O
unless the crosses are needed as it fits into a cavity or something? (and kinda feels right with having a restriction on the matrix)

[something happened to my drawing!! oh well, not important]

What I would do is have the diffusors the same. If I were making a *HUGE* diffusor of a very high prime, maybe then for ease of handling etc I'd break it into four, but they would then butt up against each other in the reconstruction. Because there are gaps, and it's ''compromised'' anyway, I think it would lessen the effectiveness if it were broken up and spaced apart?? At least if each one were 'whole' it will diffuse properly on it's own. There is a diagram somewhere from RPG which shows diffusors spaced apart, with absorption between them, and the combination itself helping the diffusion. Don't know if your gaps are bare wall or absorption.

that is all just gut feel, YMMV and IMHO type stuff, because I don't actually know.

I think aesthetically it would look better too??

So, fill us all in ok?

kareface · Aug 15, 2010

Sorry about any confusion, I felt like the other people who read the thread (or clones of it elsewhere) understood what I was getting at so I didn't realize there was any misconceptions generated by the way I presented the information. I'll try to parse my words better in the future. From my perspective, I came in asking a question, only to have someone answer something unrelated and begin explaining fundamentals of the subject. I was just confused as to why you were responding this way, it came off as patronizing initially but I understand the genesis of the discussion now.

Can you expand a bit on this?? Sounds like you have been able to do a direct comparison, would love to hear your thoughts. You seem to echo that again in your next bit

I find the QRD's don't do it for me. I can't easily identify the quality of the sound I don't like (not at least with out resorting to the classic audiophile cliches which I try to avoid). If I was pressed I'd say it doesn't impact the tonality as much. However I've never taken the time to identify the problem with the microphone and I hesitate talking about acoustics with out data. Too many people hear things that can't exist and I don't want to be the source for or involved in some psychoacoustic debate.

Now, cmon, spill the beans!! How are you planning to build it??

It wasn't some huge secret, I was just omitting unnecessary information that didn't factor into the question at hand. As I mentioned in on of my earlier responses the gaps will be dampening, which I've seen first hand produces very good results when alternated frequently. The pattern is simple, something that'll cover the majority of the primary listening position from the rear reflection points. The limitation of the dimensions comes from the size of the blocks and the space between the dampening, it's 1 foot of room to work with and the easiest solution based on materials is a 8x8 design. If I was just creating a diffuser I would certainly be willing to stick with the common dimensions. My goal is to get some benefits of both dampening and diffusion with out the deadness of one or the brightness of the other.

The pattern:
..XOX
XOXOX
..XOX

that is all just gut feel, YMMV and IMHO type stuff, because I don't actually know.

None of the questions I was asking I was expecting someone to unequivocally know. I was just looking for peoples opinion on the matter to see if I could justify taking that route. Otherwise I would of just stuck with the QRD's, which I'm again not a huge fan of. At least this route I'll be able to measure the results, I'll take it to a few different rooms to test the potential benefits.

terryj · Aug 15, 2010

kareface said:
Sorry about any confusion, I felt like the other people who read the thread (or clones of it elsewhere) understood what I was getting at so I didn't realize there was any misconceptions generated by the way I presented the information. I'll try to parse my words better in the future. From my perspective, I came in asking a question, only to have someone answer something unrelated and begin explaining fundamentals of the subject. I was just confused as to why you were responding this way, it came off as patronizing initially but I understand the genesis of the discussion now.

all cool. and I DO apologise if it sounded patronizing, prob the last thing I'd ever think of myself as. Opinionated? sure. arsehole? sure.

patronizing...nahh

I find the QRD's don't do it for me. I can't easily identify the quality of the sound I don't like (not at least with out resorting to the classic audiophile cliches which I try to avoid).

you've just gone up a few rungs in my estimation haha.

If I was pressed I'd say it doesn't impact the tonality as much. However I've never taken the time to identify the problem with the microphone and I hesitate talking about acoustics with out data. Too many people hear things that can't exist and I don't want to be the source for or involved in some psychoacoustic debate.

How did you do this?? One of the 'problems' with diffusors is, well, a) they're expensive or b) if diy, a lot of work. So, at best, a great leap of faith to go and build these bloody things...and yet not know if they will be worth it! Follow me?? I mean, I just threw up raw f/glass batts to try absorption and could hear a difference. If I didn't hear a difference, $40 down the drain, but then you can throw them in the ceiling.

Build a few diffusors...and not hear a difference?? grrr.

Also (mine are about 600*600), would I hear any difference with ONE? Ok, so don't bring them in till I have X minimum number...but what if I don't hear a difference? should I have tried with one (not hear a difference) so not build anymore...but should have tried with four when I would hear a difference??

leap of faith eh.

So, what and how did you compare the two?? Was it simply one, and that was enough to form an opinion, or more.

It wasn't some huge secret, I was just omitting unnecessary information that didn't factor into the question at hand. As I mentioned in on of my earlier responses the gaps will be dampening, which I've seen first hand produces very good results when alternated frequently.

Well there ya go, missed that one too. Reckon I should read things first? haha. Ok, so you have probably seen that reference I referred to.

The pattern is simple, something that'll cover the majority of the primary listening position from the rear reflection points. The limitation of the dimensions comes from the size of the blocks and the space between the dampening, it's 1 foot of room to work with and the easiest solution based on materials is a 8x8 design. If I was just creating a diffuser I would certainly be willing to stick with the common dimensions. My goal is to get some benefits of both dampening and diffusion with out the deadness of one or the brightness of the other.

The pattern:
..XOX
XOXOX
..XOX

None of the questions I was asking I was expecting someone to unequivocally know. I was just looking for peoples opinion on the matter to see if I could justify taking that route. Otherwise I would of just stuck with the QRD's, which I'm again not a huge fan of. At least this route I'll be able to measure the results, I'll take it to a few different rooms to test the potential benefits.

Looking forward to the measurements. If you don't post them here, where will they be? (you hinted at threads in other places).

I had decided (arbitrarily, based on reasoning like above) to not bring the PRDs in till I had six made. But a few guys were here yesterday, so I brought in the four I had finished.

BIG room yada yada, but by crikey there was a big noticeable difference in the in room measurement from the LP.

haha, puts cables and the like into perspective....if I'm going to make a change in my system, you'd better beleive it had better make a measurable difference.

kareface · Aug 15, 2010

terryj said:
all cool. and I DO apologise if it sounded patronizing, prob the last thing I'd ever think of myself as. Opinionated? sure. arsehole? sure.

patronizing...nahh

Glad we're on the same page

How did you do this?? One of the 'problems' with diffusors is, well, a) they're expensive or b) if diy, a lot of work. So, at best, a great leap of faith to go and build these bloody things...and yet not know if they will be worth it! Follow me?? I mean, I just threw up raw f/glass batts to try absorption and could hear a difference. If I didn't hear a difference, $40 down the drain, but then you can throw them in the ceiling.

Build a few diffusors...and not hear a difference?? grrr.

I've worked with them over the last year quite a bit and had a chance to hear them in various situations. When you do anything DIY you must accept the possibility you'll do it wrong. It's one of the reasons people who build their own hardware really need calibration equipment to verify the results. You'll be sure the first thing I do after finishing the design is test the polar response, MLS and autocorrelation, lol. This is one of many acoustic tools I'm building for the room so I can't be squeamish about wasting time on something that might not pan out. I'm looking forward to constructing a combination MPD & Hermholtz panel for the ceiling.

Also (mine are about 600*600), would I hear any difference with ONE? Ok, so don't bring them in till I have X minimum number...but what if I don't hear a difference? should I have tried with one (not hear a difference) so not build anymore...but should have tried with four when I would hear a difference??

leap of faith eh.

Even if I heard a difference, if it was my room I'd want to see something on paper. Faith would imply there's no evidence, just take the time to verify the difference. It would suck if all that time produced no results, but it's no reason no to try.

So, what and how did you compare the two?? Was it simply one, and that was enough to form an opinion, or more.

Unfortunately I didn't have a direct comparison. I setup a sound room with some QRDs, and we listened both with and without them. I didn't like the emphasis that existed with the QRD's in the room. We brought the rear ones back and replaced them with PRD's that were mounted on stands (and weighed more than the average American) and tested it with and without them in the room. It sounded better in my opinion, but there's no way to tell if it was simply expectation bias. I would say it was fairly unanimous among the other listeners, but that doesn't amount to proof. I will note, that after the room was near completion during the acoustic calibration we did adjust the amount of absorption in the rear of the room to tone down the PRDs. That's just a byproduct of that style of diffuser.

Looking forward to the measurements. If you don't post them here, where will they be? (you hinted at threads in other places).

I'll post them here for sure. I've got a lot of work before they'll be done but I'll make sure to post an update. I also have to measure some LF results from low density fiber for Ethan when I get a chance.

I had decided (arbitrarily, based on reasoning like above) to not bring the PRDs in till I had six made. But a few guys were here yesterday, so I brought in the four I had finished.

BIG room yada yada, but by crikey there was a big noticeable difference in the in room measurement from the LP.

Good to hear, it's always nice to verify the fruits of your labor.

haha, puts cables and the like into perspective....if I'm going to make a change in my system, you'd better beleive it had better make a measurable difference.

This, this, this. I can't tell you how many people I've talked with or worked with who are willing to drop everything on the cost of the cables and leave the room (something which has a clearly negative impact on the reproduction) completely under treated. From a physics standpoint the whole cable thing is laughable. There maybe some legitimacy for IC's due to return current paths but that doesn't necessarily correlate to expensive being better, it just depends on the resistance of the amps line cord ground in relation to the IC's shield. Even in that instance the coupling is frequently inaudible (or at least unmeasurable). People get so focused on things that have so little impact on their systems that they leave unaddressed much larger problems. Room acoustics is the last 50% of what you hear, and while digital correction is getting better it can't replace the results of proper acoustical treatment. It warms my heart to see more people focus on the important elements of the room. Maybe I can get some more business out of it as well, lol.

Bulldogger · Aug 16, 2010

audioguy said:
My back wall is Skylines and I have an upper midrange glare the cause of which we have not been able to locate (either Rives, who designed the room, or another company). Last evening I covered all of the Skylines with 2 inch 703 and, unfortunately, it did not solve my problem even a bit.

I would have put the OC703 on the side walls at the first reflection points. I am no expert at this. This is only the stuff that worked in my room with the speakers I was using. The OC703 2 inch stuff will absorb the midrange. It is just a question of how much too use and location. Enough OC703 will make the room extremely dead. If the designer of your speakers has designed them to have additional energy in the midrange, then treating reflected sound is only going to help so much. Are you using EQ? That would seem to be the only way to address the direct sound if it is a design factor.

audioguy · Aug 16, 2010

The side walls use RPG BAD ARCS. (Diffusors and absorbers). The orignial design of the rear wall was Skylines spaced about 4 inches off the wall spaced about 6 inches apart with 2 inch 703 behind each skyline. I (arbitraritly) decided that was too much work so just stuck the skylines directly on the wall. We shall see if his idea of replacing every other Skyline with 2 inch 703 on the wall works?

Sure do hope so!!

kareface · Aug 16, 2010

Ah, if you have Bad Arcs on the side walls that might explain some of the problem. Starting at around 1000hz the absorption of those panels becomes less effective. You'll still get some scattering from the diffusing surface but above that you are sorely lacking absorption. The other problem is the design of the arc isn't idea for the first reflection point. You generally want the progression of sound to start at the front of the room and always progress back where you have something to break up or absorb the excess energy. With a curved design, some of the sound will be reflected to the front of the room again, which wouldn't be a problem with walls that have a gradient of 10% of greater, but a room with parallel walls might not produce ideal acoustics. This panel in my opinion would be better served in a position behind the main listening position, but this is my opinion from working with several other geometric diffusers.

In your situation I would try and identify the problem frequency. Run a sweep to see if any mid or high frequencies stand out on the waterfall. The problem could also be from the "brightening" that commonly occurs with diffusion that lacks dampening. In you case you have the BAD ACR which has dampening, but is designed to dampen more effectively below the diffusion cutoff, which doesn't help you because the mids or highs depending on the room are the ones who's tonality are negatively impacted by balances between diffusion and dampening. If you alternate the diffusers and the dampening in the rear of the room it'll help, but you should test to see if you get better results with a normal broadband absorber at the first reflection points in place of a diffuser. Good luck with your room.

kareface · Aug 17, 2010

Here's a question for the AE's out there. I've been reading through the patent on the 2D primitive root diffuser. Unless I'm mistaken the formula does not require the least primitive root and seems to work with any primitive root. I suspect the creator of the calculator used the least primitive root as a simple starting point and because it was the example used in the patent. However from what I've read any primitive root of the prime of (C*R)+1 will work.

The algorithm looks like fortran, but maybe someone can correct me on this:

dimension idif(200,200),id(13,12), idd(13,12),ip(30),idis(30) dimension ipp(30) dimension idc(156) open(unit=20,file='out.dat' ,form='formatted', status='unknown') C ipr=157 irt=5 ni=13 nj=12 c ii=0 jj=0 mmod=1 do 20 n=1,ipr-1 mmod=mmod*irt mmod=mod(mmod,ipr) iii=mod(ii,ni)+1 jjj=mod(jj,nj)+1 id(iii,jjj)=n idd(iii,jjj)=mmod idc(mmod)=idc(mmod)+1 ii = ii+1 jj=jj+1 20 continue c 40 continue do 300 j=1,nj write(20,310) (id(i,j),i=1,ni) 310 format(2x,13i4) 300 continue write(20,330) 330 format (//) do 320 j=1,nj write(20,310) (idd(i,j),i=1,ni) 320 continue do 857 i=1,ipr-1 857 write(20,310)i,idc(i) close(20) end

It seems the algorithm just puts root^bin# in each bin and uses a pattern to number the bins so hypothetically any primitive root would generate a functional permutation. Any thoughts?

Here's a link to the patent:
http://patft.uspto.gov/netacgi/nph-...&f=G&l=50&d=PALL&RefSrch=yes&Query=PN/5401921

I like the line at the end:

Of course, various changes, modifications and alterations in the teachings of the present invention may be contemplated by those skilled in the art without departing from the intended spirit and scope thereof.

Steve Williams · Aug 17, 2010

Wow

Reading this makes me think we need to start a mensa forum

64 Cell (8x8) PRD (Skyline Diffuser) question

New Member

Well-Known Member

WBF Founding Member

New Member

New Member

New Member

New Member

Site Founder, Site Owner, Administrator

Member Sponsor & WBF Technical Expert

New Member

New Member

New Member

New Member

New Member

New Member

Well-Known Member

WBF Founding Member

New Member

New Member

Site Founder, Site Owner, Administrator

Similar threads