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
No temporal diffusion with polys? Reverb & Delay Plugins
Old 15th September 2011
  #1
Gear Maniac
 
cborg's Avatar
 

Question No temporal diffusion with polys?

Hi folks!

Got a couple of questions I'm not really getting my head around. Maybe someone will be able to help me.

First:
What's the definition on temporal diffusion?

From what I've gathered so far it is related to the reflections having different path lengths and by that also a difference in time.
Maybe the path lengths needs to be distributed in a specific way, I don't know.


Second:
I've read here and on other places that polys, a (semicirkular) convex arch, doesn't really sport any temporal diffusion.

Is this true? and if so, how?


Using what I wrote above polys has theoretically a nice polar dispersion (not sure if I'm using the right words here). Shouldn't that also give different path lengths when reflected back to the listening point?
What am I missing here?
Old 15th September 2011
  #2
Gear Addict
 

Quote:
Using what I wrote above polys has theoretically a nice polar dispersion
theoretically and measured, very good spatial diffusion.


Quote:
I've read here and on other places that polys, a (semicirkular) convex arch, doesn't really sport any temporal diffusion.

Is this true? and if so, how?

I think the reason is that often single polys are discussed, normally one would have several diffusors, and it will introduce temporal diffusion, it is best to use a randomniced pattern, with diffusors of different size.

Another method is to to have a surface with both poly diffusors and absoprtion (randomniced)

temporal diffusion is important in order to avoid negative combfilter effects.
Quote:

Consequently, a good
diffuser needs to generate
spatial AND temporal
dispersion. While a Schroeder diffuser may be designed purely for its spatial
scattering properties, it naturally generates temporal dispersion because of its
complex geometry (provided the period width is wide enough). In the case of semicylinders,
arranging them in a periodic array may not be sufficient to disperse the
temporal response, and some form of randomised arrangement is probably
needed.
from http://www.rpginc.com/news/library/tyndall_paper.pdf


I have not seen a standard method for testing temporal diffusion, and I cannot find any coefficient table for temporal diffusion in Cox/D`Antonio "acoustic absorbers and diffusors" (first edition).
Old 15th September 2011
  #3
Gear Addict
 

Another way of avoiding combfiltering effects is to use narrow diffusors.

a semicircle/elipse with a width of 40 cm will have a much sharper curve than a semicircle/elipse with a width of 80 cm.
Old 15th September 2011
  #4
Gear Maniac
 

Quote:
Originally Posted by hsal View Post

I think the reason is that often single polys are discussed, normally one would have several diffusors, and it will introduce temporal diffusion, it is best to use a randomniced pattern, with diffusors of different size.

Another method is to to have a surface with both poly diffusors and absoprtion (randomniced)

temporal diffusion is important in order to avoid negative comb filter effects.
Glad I wasn't the first to point this out. Properly done, polys will be different sizes, different spacings, and even different rotations. When modeled, this should be taken into account along with the wall behind them as part of the system.

As previously stated, this isn't usually done. Some have even deliberately crammed multiple polys of the same size edge to edge and then crept right up on them to "prove" how bad they were.....of course, they were trying to sell their own product.
Old 15th September 2011
  #5
Lives for gear
 
avare's Avatar
 

Quote:
Originally Posted by FLZapped View Post
Glad I wasn't the first to point this out. Properly done, polys will be different sizes, different spacings, and even different rotations. When modeled, this should be taken into account along with the wall behind them as part of the system.
+1.

With no contextual diffusion,
Andre
Old 15th September 2011
  #6
Gear Guru
 
Glenn Kuras's Avatar
We actually where going to put something like that on the market, but the price would be pretty much out of this world. Here are few pictures of what it looked like.


Old 15th September 2011
  #8
Lives for gear
 

Quote:
Originally Posted by Glenn Kuras View Post
We actually where going to put something like that on the market, but the price would be pretty much out of this world. Here are few pictures of what it looked like.



I've seen these two pictures in several threads in a poly search today. These things are gorgeous! Glenn, why do I have a feeling that every prototype you made is in use in your room? hogging the poly-QRDs!!!
Old 15th September 2011
  #9
Gear Maniac
 
cborg's Avatar
 

Quote:
Originally Posted by hsal View Post
theoretically and measured, very good spatial diffusion.
Ok, I can see that. And we're talking single poly now, aren't we?

Quote:
temporal diffusion is important in order to avoid negative combfilter effects.
Understood.

Well, the thing I don't get is how the response is formed for the poly. And for simplicitys sake lets look at the single, semicircular poly.

It seems that the combfilter created from the interfering response is almost as that of a plane surface.
(I don't know how the test was done so I cannot be sure how it applies though.)

Is this because of the continuous surface perhaps?
Compared to a 1D QRD for example.

I mean, even the poly has different pathlengths depending on where the wavefront strikes.


On another note, there are a couple of "may"s and "probably"s in the quote and the report which makes me a bit hesitant to draw any conclusions from those parts.

Quote:
I have not seen a standard method for testing temporal diffusion, and I cannot find any coefficient table for temporal diffusion in Cox/D`Antonio "acoustic absorbers and diffusors" (first edition).
And neither have I ...

(Thanks so much for taking the time!)
Old 15th September 2011
  #10
Gear Addict
 

Just to have this out in the open, I make and sell both qrd and poly diffusors (primerarly for the Norwegian market).

From Cox/D`Antonio "acoustic absorbers and diffusors" vol 1 page 295:

" Figures 10.28 and 10.29 show the total field time and frequensy responses for a complicated arrangement of cylinder sizes and shapes has further broken up the frequensy responses for a complicated arrangement of many different-siced semicylinders.The use of a random arrangement of cylinder sizes and shapes has further broken up the frequensy response, making it much less likely that coloration will be heard ".

Acoustic elements with circular shape was first used by vitruvius in the greek amphi theaters of ancient greece.

Live rooms e.t.c where polys have been used : RCA studios NYC, several rooms, including main. http://www.scottymoore.net/studio_radiorecorders.html

Walt Disney, RCA film studios in 40`s 50`s and on, also listening rooms:

http://forums.klipsch.com/forums/storage/3/653028/Volkman%20on%20Polycylindrical%20Diffusers.pdf

Universal studios, RCA NBC plus "many indiependent studios"from mid fifties and on:

Quote:
Polycylindrical
diffusers (1) were more prominent in the design of studios fbr RCA and NBC than
elsewhere, but soon became common-place in other independent studios
Quote:
We constructed a "roll-around" band shell with interior polycylindrioal diffusers to prevent focusing, shown in figure #3 . This quickly gained approval from the string players
Quote:
[/QUOTE]http://www.aes.org/aeshc/pdf/putnam_history-of-recording-studios.pdf

Last edited by hsal; 15th September 2011 at 09:09 PM.. Reason: spellcheck, correct quotes e.t.c
Old 15th September 2011
  #11
Gear Addict
 

Quote:
Well, the thing I don't get is how the response is formed for the poly. And for simplicitys sake lets look at the single, semicircular poly.

It seems that the combfilter created from the interfering response is almost as that of a plane surface.
(I don't know how the test was done so I cannot be sure how it applies though.)
looking in Cox/D`Antonio "acoustic abdorbers and diffusors" it seems that it is the same as in figure 10.18, large poly diffusor 1 meter wide.

"Concequently, large cylinders and semicylinders produce combfiltering similar to that of a plane surface."

Looking at figure showing the combfiltering effect from a single 1meter wide panel and a single 1meter wide semicircle, it is similar, but the dips of the combfiltering is more pronounced with the flat panel.
Old 15th September 2011
  #12
Gear Maniac
 
cborg's Avatar
 

Quote:
Originally Posted by FLZapped View Post
Glad I wasn't the first to point this out. Properly done, polys will be different sizes, different spacings, and even different rotations. When modeled, this should be taken into account along with the wall behind them as part of the system.
I've actually tested to model both polys and 1D QRDs, with and without the wall in my studio, with BEM.
So I can definitely see how it affects the system as a total.
It doesn't say anything about the temporal diffusion though so I'm trying to find a way to get some understanding for that. To make the best compromise for my room.

Thanks!
Old 15th September 2011
  #13
Gear Maniac
 
cborg's Avatar
 

hsal, thanks again!
I'll look into those links on this specific subject to see if I get any wiser.
Old 15th September 2011
  #14
Gear Addict
 

My pleassure, I am very interested in the history of acoustic treatment, just glad to share some links
Old 15th September 2011
  #15
Gear Maniac
 
cborg's Avatar
 

Quote:
Originally Posted by Jens Eklund View Post
Thanks for publishing prediction for those models Jens!
It seems there is no way to safely predict the temporal bit unless using another mathematical model if I understand your conversation with Nordenstam?
Old 15th September 2011
  #16
Lives for gear
 
Jens Eklund's Avatar
Quote:
Originally Posted by cborg View Post
Thanks for publishing prediction for those models Jens!
It seems there is no way to safely predict the temporal bit unless using another mathematical model if I understand your conversation with Nordenstam?
Correct, FDTD (Finite Difference Time Domain) or other models are needed (or actual full-scale measurements).
Old 15th September 2011
  #17
Gear Guru
 
Glenn Kuras's Avatar
Quote:
Originally Posted by Torea View Post
I've seen these two pictures in several threads in a poly search today. These things are gorgeous! Glenn, why do I have a feeling that every prototype you made is in use in your room? hogging the poly-QRDs!!!


And my personal art panels!heh



Old 16th September 2011
  #18
Lives for gear
 

Quote:
Originally Posted by Glenn Kuras View Post


And my personal art panels!heh



Old 16th September 2011
  #19
Gear Addict
 

Regarding polys, QRDs and comb-filtering. This is what I think happens...

Consider a single loudspeaker in a room and a single listener. Consider the case of sound travelling as rays. Comb filtering occurs when the direct sound from a source is combined with a reflected sound ray from another surface (be it wall, floor or ceiling). The reflected sound arrives at the listener with a time delay but much of the same audio information as the direct sound. When these combine, the effect can be audibly undesirable.

With QRDs and PRDs, the sound ray that is geometrically reflected from the panel towards the listener is at least a few dB lower than the purely specular case. Therefore, this combines more weakly with the direct sound than a purely specular reflection would, hence weaker comb filter effect.

In the case of a poly, it also would have a somewhat reduced geometric reflection, depending on its size, but I presume not as significantly lower compared to the case for effective QRDs/PRDs.
Old 16th September 2011
  #20
Gear Addict
 

Quote:
In the case of a poly, it also would have a somewhat reduced geometric reflection, depending on its size
I presume that with a smaller width the reflection has lower level than with larger (assuming same shape), are there any measurements of the effect of size on this parameter?

Another way to lower the level of the reflections is to place the polys so that the incident angle is benificial.

For instance, with an incident angle of 70 degrees or less (appr), it is possible to make the level of the reflections a lot lower, if the listening position is not in the specular direction ,listening position can be perpendicular to the polys (or similar) on the sidewall, polys starting from a position directly beside listening position, and continuing backwards on the sidewall.

Figure 10.23 in "acoustic absorbers and diffusors" Cox/D`Antonio show a measurements with different incident angles

This works when the panel is not a semicircle, but a bit "flattened" (like most polys I`ve seen constructed.)

Another aspect is that with many polys there will be many combfilter effects (unless placement/sice avoids it), and with good variation they will overlap due to the timing difference of reflections causing different frequensies for the dips and peaks of the combfiltering effects

Last edited by hsal; 16th September 2011 at 11:20 AM.. Reason: added example from "acoustic absorbers and diffusors", minor changes nr 2
Old 16th September 2011
  #21
Lives for gear
 
avare's Avatar
 

Quote:
Originally Posted by Sebg View Post
With QRDs and PRDs, the sound ray that is geometrically reflected from the panel towards the listener is at least a few dB lower than the purely specular case. Therefore, this combines more weakly with the direct sound than a purely specular reflection would, hence weaker comb filter effect.

In the case of a poly, it also would have a somewhat reduced geometric reflection, depending on its size, but I presume not as significantly lower compared to the case for effective QRDs/PRDs.
First, a thank you to Glenn. Several years ago he asked at what point in terms of distance the reflection from a cylindrical diffuser would no longer be significant. I tried to think what criteria to use, and then how to calculate levels from the diffuser. Using the EBU standard of reflections within the ISD being 10 dB down, the reflection from a poly would be 10 dB down at distance of twice the radius of the poly.

A degree of temporal diffusion would be exhibited by multiple cylindrical diffusers on a wall or ceiling, because each diffuser would radiate reflections throughout an arc twice the angle subtended by the diffuser.

Cylindrically,
Andre
Old 16th September 2011
  #22
Gear Addict
 

Excellent info Andre, thanks alot.
Old 16th September 2011
  #23
Gear Addict
 

Andre,

I just want to clarify that when you say "10dB down", this means at 2x the radius (ie at a distance = diameter of curvature), compared to the level of the incident sound on the poly. (The actual radius of curvature of a poly is usually much larger than the size of the poly itself).

I will check if I saved any of the AFMG Reflex calcs I did using polys of different sizes; as one of the calcs the software spits out is a polar plot of the reflected sound level (this can be done for any angle of incidence). In terms of comb filtering, we are interested in comparing the level of the reflected sound to the level of the direct sound at the listener. This will depend on the relative differences in distance from the source and the width of the room. I imagine that a level that is say at least 5dB down compared to the direct sound would significantly help reduce comb filtering effects (from that room surface). The graph of the effects of lateral reflections in Everest MHOA does not indicate the region of comb filtering (level vs time delay) only regions of "spaciousness" and "echo" and "inaudible effects".

There is a neat You-Tube video that shows simulated comb filtering by using audio software. I presume one could experiment with reducing the level (amplitude) of the delayed sound to see when the audible comb filtering goes away.

Hearing Comb Filtering - YouTube

At a listening distance of 4m, with a 2m side distance, the delayed (reflected) sound is at approximately 5ms.
Old 17th September 2011
  #24
Gear Addict
 

Quote:
There is a neat You-Tube video that shows simulated comb filtering by using audio software. I presume one could experiment with reducing the level (amplitude) of the delayed sound to see when the audible comb filtering goes away.
I don`t think this method is transferable to real world application, due to the auditory systems ability to suppress coloration of timbre (depending on incident angle of sound).

For instance when reflections come from the side , both the timing difference of arrival time at each ear, and the difference in tonality of the two signals due to one being diffracted around the head is important cues that the brain use to decolorize the sound.

The effect of the demonstration in the video is different due to the combfiltering being created by several identical audio tracks being played with different timing, and played back thru the loudspeakers, hence no difference in angle of incidence for direct signal and delayed signal.
Old 17th September 2011
  #25
Lives for gear
 
avare's Avatar
 

Quote:
Originally Posted by Sebg View Post
Andre,

I just want to clarify that when you say "10dB down", this means at 2x the radius (ie at a distance = diameter of curvature), compared to the level of the incident sound on the poly. (The actual radius of curvature of a poly is usually much larger than the size of the poly itself).
You clarified it very well. Of course there is the proviso that the surface must be significant compared to the wavelength of the sound. If there were not any qualifiers to obscure things, it would not be acoustics.

Well rounded like a poly,
Andre
Old 19th September 2011
  #26
Lives for gear
 
jhbrandt's Avatar
Quote:
Originally Posted by avare View Post
.... If there were not any qualifiers to obscure things, it would not be acoustics.
HAHA! +1

and to add that 'The actual radius of curvature of a poly is usually much larger than the size of the poly itself'... we usually add slats to the surface of the poly to increase the effective diffusion at close range.

A poly that is made with one sheet of plywood bowed between verticals that are 46 1/2" (97% of 48") apart will have a radius of 5' 11" and therefore a usable distance of 12'.

Cheers,
John
Old 20th September 2011
  #27
Gear Maniac
 

Quote:
Originally Posted by jhbrandt View Post
HAHA! +1

and to add that 'The actual radius of curvature of a poly is usually much larger than the size of the poly itself'... we usually add slats to the surface of the poly to increase the effective diffusion at close range.

A poly that is made with one sheet of plywood bowed between verticals that are 46 1/2" (97% of 48") apart will have a radius of 5' 11" and therefore a usable distance of 12'.

Cheers,
John

While that size would be fine for a large room, like the sanctuary in a church, it wouldn't be very practical for a small room like a control room in a recording studio. So your example may be rather extreme.

-Bruce
Old 20th September 2011
  #28
Lives for gear
 
avare's Avatar
 

Quote:
Originally Posted by FLZapped View Post
While that size would be fine for a large room, like the sanctuary in a church, it wouldn't be very practical for a small room like a control room in a recording studio. So your example may be rather extreme.
With all due respect, what is your point? The units would designed for the room they would be used in. Study my signature line.

Studiously,
Andre
Old 20th September 2011
  #29
Lives for gear
 
jhbrandt's Avatar
Quote:
Originally Posted by FLZapped View Post
While that size would be fine for a large room, like the sanctuary in a church, it wouldn't be very practical for a small room like a control room in a recording studio. So your example may be rather extreme.

-Bruce
Bruce,
It's an example of the formula... make a smaller poly or squish the sides in more and you can work closer... Exactly half the size of the example poly will give a nice working distance of 6 feet. Vary the width and/or radius by 5% (at least) for an array. Add slats on it.. and you can even work much closer.

Cheers,
John
Old 20th September 2011
  #30
Lives for gear
 
avare's Avatar
 

Quote:
Originally Posted by jhbrandt View Post
It's an example of the formula... make a smaller poly or squish the sides in more and you can work closer... Exactly half the size of the example poly will give a nice working distance of 6 feet. Vary the width and/or radius by 5% (at least) for an array. Add slats on it.. and you can even work much closer.
We appear to be posting together.

A point that may overlooked by many reading this thead, is that the distance figure was worked out provide some analogous relationship with the QRD guideline of 3 times the wavelength of the design frequency. Closer than that distance and the effect of the unit will become more obvious. If that is desired, of course it is good.

Hmm, seems like good studio building has a lot of designing in it.

Cunningly designed,
Andre
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • 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.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Forum Jump
Forum Jump