The No.1 Website for Pro Audio
 All  This Thread  Reviews  Gear Database  Gear for sale     Latest  Trending
2 ch listening room, aiming for best possible.
Old 11th September 2009
  #151
Lives for gear
 
Nordenstam's Avatar
 

QRD Rotation = bad?

Another set of Falstad simulations. Once again adressing the basic issue of sequence rotation. The common knowledge is that diffusors can start and stop at any place. Have suspected for a while that it doesn't work that way unless it's repeating a lot of times and the response at the edges don't matter. It is with careful concern that I now speak up against that established truth. Just can't see how an even dispersion is possible unless the small numbered prime diffusers are symmetrical. The Huygens drawings are harder to interpret than the Falstad pics. The tendency is still the same: A symmetrical array is needed to give a symmetrical response! The simulations illustrates the point.


Asymmetrical rotation 4-2-2-4-1-0-1:
2 ch listening room, aiming for best possible.-rotation.png

Asymmetrical again, a bit later:
2 ch listening room, aiming for best possible.-rotation2.png



Symmetrical rotation 2-4-1-0-1-4-2:
2 ch listening room, aiming for best possible.-rotation3.png

Symmetrical again, a bit later:
2 ch listening room, aiming for best possible.-rotation4.png


(original sequence is 0-1-4-2-2-4-1)



The rotated one still generates seven lobes, but the energy is spread unevenly. It also seems the symmetrical +3 offset have more energy in the specular direction than a normal QRD. The change from having the zero in the middle to having it on one end is minimal when it comes to building these things. So practically speaking, I don't see any reason to rotate them, unless it's desired to send the energy off in a certain direction. Or, perhaps, I'm all wrong about this? Input on these musings are much appreciated!


Cheers,

Andreas Nordenstam
Old 12th September 2009
  #152
Gear Maniac
 

sneaky bugger, been wondering how your snaps always looked so 'clean'....use single source and then turn it off eh?

It's a pity that whenever you click back to 'no source' that it resets itself. That is fine with single source, but with two sources etc you can position them where you want and turn it off, but then next time you need to re-position them again.

Why am I mentioning this?? Real world situation.

I always think that you gotta be careful just how much you let yourself get bogged down in the minutiae. Audiotwits do it all the time...."my binding posts have 0.08% more rhodium than yours, THAT is why I have 'less veils' than you. I recommend patented Acme binding posts for that reason, and an absolute steal at only $700 each".

Bogged down in things that don't matter one whit.

As soon as, for example, you add a second source to the ripple tank, it becomes absolutely chaotic in there.

Imagine the real world of our room, multiple sources, reflections from every conceivable place and from every conceivable angle......ultimately are we arguing about angels on the head of a pin?

It IS good to sort the wheat from the chaff of course, but we do not want to run the risk of giving the wrong message that it is 'all too hard' or whatever.

Just a thought on this, that's all.

On a side note, watch a ripple tank simulation for a little while before listening....then when listening with your eyes closed you 'see' the waves expanding everywhere in your room!!

way cool maaan.

I just wish the 'clear waves' button was not so close to the 'clear walls' button, very annoying when you've had a few beers.
Old 12th September 2009
  #153
Gear Maniac
 
collo's Avatar
 

Lupo,
more food for thought from you, as usual.....


Proposed rules for auto-well width

How do these sound?

- Small enough so that HF cutoff excedes design frequency times four. By how much??

- Should not be smaller than 25mm

- Large enough for period width >= 1 wavelength of design frequency
ie Panel width >= 2 * design depth
This is generally met for panels with N>7
If it's close, it will vary with fin width

These rules mean that the recommended width will no longer be a fixed proportion of the design wavelength, but float around to give the best design.

It's worth getting the standard N7 as good as we can, because that is what most people will want to build.

If you like to run with auto turned off, you will be happy to know that the final version will remember your preferences.

Also on the to-do list are:
- Save / Load project
- Support for inches - more than half my visitors are from the US, and they just love it when software supports their preferred units!



Referring to the book
- Whilst I don't have the book, I have no problem supporting their work, it's just that it is out of my price range. I'll certainly give it a plug at the top of the links section.
- Most people who use the program won't have the book either.
- I would like to be able to justify any functions in the calculator without having to refer people to the book.
- There is a limit to how many smarts from the book can be included in the calculator, and thus explained on the webpage, before RPG decides we have plagiarised too much of their work. We may be there already.
- As Terry pointed out, we're starting to split hairs a bit.

Sideways shifting of panels
Good info about symmetry of wells and diffusion.
Not an issue with standard panels. It's only when doing an advanced panel that serious side shift would be contemplated. Would be nice to add some screenshots to the webpage to show the issue.

Speaking of screenshots, I can't see your latest pictures.

Terry - the idea of including some ripple-tank screenshots to illustrate different points is a good one. I've been a bit busy lately, so still can't say if the clipboard interface to the ripple tank is a go'er or not...
I'm still writing the bit on the Barker sequence - will make any "panel rotation" bits clearer...
Old 12th September 2009
  #154
Gear Addict
 

I thought I'd take a quick plunge into the discussion here. The ripple tank applet is a nice little programm, but I'm afraid you guys are putting too much faith into it. I've spent the last week researching acoustic simulation algorithms, and I really doubt a simple reflection like the applet seems to do gives accurate results. Good for a quick demonstration, but no good for judging how different diffusor configurations compare.
Old 13th September 2009
  #155
Gear Maniac
 
collo's Avatar
 

hmm, that interface seems to be getting further away...

The new rules for auto well width have been implimented and documented. Lupo - have a play and let me know what you think!
Old 13th September 2009
  #156
Gear Maniac
 

Quote:
Originally Posted by spm_gl View Post
I thought I'd take a quick plunge into the discussion here. The ripple tank applet is a nice little programm, but I'm afraid you guys are putting too much faith into it. I've spent the last week researching acoustic simulation algorithms, and I really doubt a simple reflection like the applet seems to do gives accurate results. Good for a quick demonstration, but no good for judging how different diffusor configurations compare.
Perhaps we are putting too much faith in it.

But for me, it's value lies not so much in the accuracy of comparing one diffusor to another, tho I do think that IS useful to a degree, but more the graphic illustration of the underlying concepts.

Eg which is more useful to a guy investigating diffusors.

" A qrd breaks up and disperses the incident wave into many directions, with a commensurate lowering of the spl, and has an added advantage of delaying the signal as well"

OR, that idea accompanied by the following picture



which clearly shows a few things....off the flat wall (the source was above and directly between the flat wall and the qrd) the reflected wave is clearly a duplicate of the original, and obeys the principle of 'angle of reflection is equal to the angle of incidence'.

From the qrd it is clear that it is much less a duplicate, and the temporal delay is also evident.

It seems it may be too difficult for collo to incorporate the animation into his calculator, which is unfortunate, but never the less the inclusion of 'stills' like this will go a long way to make a verbal descriprion much more real.

On a side note, being the egocentric bugger that I am haha, I note that no-one is talking about me and my room anymore. Well, that got me dander up (not really, just joking)...

I could not but help myself, just had to try the look of using pvc pipe of varying lengths in the space coupler.

It looks pretty good, but might just be that tad bit too funky for a house and room like mine.

Loosely based on the PRD grid, mainly to get a random look, just throwing them in anywhere I'm sure would lead to regularity.





So, will go back to the wood look for the coupler I think, will reserve the use of pvc pipe to make a PRD.

Rather than square 'wells' they will of course be round, but hey (as I understand it anyway...we don't seem to have quite gotten to the bottom of this one yet) it is a compromise to make a prd without wells, so the use of round ones dos not phase me too much. There WILL be diffusion and scattering even if it is not mathematically optimum.

Strangely enough, looking at what happens with the ripple tank once you have put a few walls etc in has made me a little blase about sweating the 'minor' stuff.
Attached Thumbnails
2 ch listening room, aiming for best possible.-flat-wall-vs-diffusor.jpg   2 ch listening room, aiming for best possible.-100_6162.jpg   2 ch listening room, aiming for best possible.-100_6163.jpg   2 ch listening room, aiming for best possible.-100_6164.jpg  
Old 13th September 2009
  #157
Gear Addict
 

If you put the pvc coupler with the flat side in front, it doesn't look quite as spacey. You could also cover it with acoustically transparent fabric. Please let us know how it sounds though.
Old 17th September 2009
  #158
Gear Maniac
 

found a few packets of spare time this week, built the next panel of the space coupler

just propped for the photo, whaddya reckon, find some nice fabric to hide the yellow f/glass??!!





The build should pick up a little after this weekend, things ease up for a brief period.



inspired, I have been playing with The Calculator (note the capital letters, a mark of respect reserved for things that are worthy, as in 'The Book' heh) to see what it now tells us about the diffractal proposed by Lupo way back in post 64 (p3).

A couple of obvious things have come up.

First off, the 'restriction' of min 25 mm well width is obviously not applicable to all circumstances.

An example that does not meet that is this one

http://www.kineticsnoise.com/interio...true%20QRD.pdf which I recall is exactly the same as in Alton Everests handbook.

So we need the ratio between well width and depth that must apply. A depth of only two mm (to exaggerate) will not suffer viscous losses with a well width of 24 (to exaggerate) millimetres.

There is obviously a trade off with how the calc is set up, we can't have infinite adjustment of each parameter that affects everything else. In this calc, the total width of the panel is set by the design frequency, the other choice would have been to be able to change the design width and have the frequency change. (useful in this specific circumstance, as the small qrd fits into, and so is set by, the width of the larger qrd).

No big deal, just change the field until you get what you want (same applies to the lower frequency of the small one meeting the upper frequency of the larger one)

Guess what I am saying is that it would be good (dunno how do-able it is) to select ONE constant (frequency, well width, total panel width or whatever) and that choice then ripples thru everything else (even say if we select a maximum depth, and that stays constant if we switch to the inverse, or frequency or whatever)

gee, hope that does not sound like criticism!

Will fine tune it over the next few days, currently looking at for the large one prime 17, well width of 105 (inc 3mm wall) depth of 200. Flow 430 Fh 1686. That was basically based on the space available to me, about 1800 wide.

Small qrd prime 7, depth (actual) 30...Flow 1622 Fh 14333

that of course ignores the warning of less than 25 mm well width.

So, do we need to match exactly the upper limit of the big one to the lower limit of the small one? Guess for those of us not in RPG this is all uncharted territory.

Anyway, just throwing up a few points, thoughts on a final diffractal design more than welcome.

Jan, have decided to go with the original look, the pvc probably does not suit my room, the wood (as you can see) fits in nicely with the existing woodwork in the room.

Too late now, but as I was putting the vertical dividers in, what looked really nice was rectangles (miss every second vertical piece) rather than squares.

food for thought for others perhaps.
Old 17th September 2009
  #159
Gear Maniac
 
collo's Avatar
 

Nice photos - good progress!
Definitely need some decent fabric in there. Joint effort with the missus I would expect...

Funny you should mention the 25mm limit. I've just spent several hours re-writing the section on the webpage about fine tuning the well widths.
This is in advance of making the analyser in the calculator a bit smarter.
It makes provision for one of the rules to base well width on build depth rather than design frequency. Have a squiz and let me know what you reckon.

A change to qrdude is a new menu option for Scaling the picture. This allows you to dump the panel pic to the clipboard, preserving the relative heights for panels with different build depths. Designed to help stop people drawing up their panel arrangements and forgetting that their alternate panel might be a different depth to the main ones.

I'll have a think about the further changes you suggested. (It would be silly to see free feedback as criticism)

Been thinking about diffractals too. It's easy to fire up two copies of qrdude - one for the panel itself, and one for the well inserts.

I found a nice match between the two using:

Panel
-----
Data for One Dimensional QRD (Quadratic Residue) Diffuser
Standard N7 Panel
Design frequency 702 hz
Number of wells: 7
Deepest well in mm 140
Fin width in mm: 3
Width of wells in mm: 70
Ratios 0 1 4 2 2 4 1
Well depths in mm 0 35 140 70 70 140 35
Phase shift in degrees 0 51 205 102 102 205 51
LF cutoff frequency 351 hz HF cutoff 2457 hz
Panel width 511 mm
Minimum distance to seating position 2940mm (3 times wavelength of LF cutoff frequency)

Inserts
-------
Data for One Dimensional QRD (Quadratic Residue) Diffuser
Standard N7 Panel
Design frequency 4914 hz
Number of wells: 7
Deepest well in mm 20
Fin width in mm: 1
Width of wells in mm: 9
Ratios 0 1 4 2 2 4 1
Well depths in mm 0 5 20 10 10 20 5
Phase shift in degrees 0 51 205 102 102 205 51
LF cutoff frequency 2457 hz HF cutoff 19111 hz
Panel width 70 mm
Minimum distance to seating position 420mm (3 times wavelength of LF cutoff frequency)


With the LF cutoff frequency equaling the point where the diffuser stops working completely, I would be tempted to overlap the ranges by a bit - make the wells in the inserts 30mm deep instead of 20.

thusly...
Inserts
-------
Data for One Dimensional QRD (Quadratic Residue) Diffuser
Standard N7 Panel
Design frequency 3245 hz
Number of wells: 7
Deepest well in mm30
Fin width in mm: 1
Width of wells in mm: 9
Ratios 0 1 4 2 2 4 1
Well depths in mm 0 7 30 15 15 30 7
Phase shift in degrees 0 51 205 102 102 205 51
LF cutoff frequency 1622 hz HF cutoff 19111 hz
Panel width 70 mm
Minimum distance to seating position 636mm (3 times wavelength of LF cutoff frequency)

Should give coverage from 351 hz to 19111 hz
(Being from Bathurst, you would realise the magic in the 351...) heh
Old 17th September 2009
  #160
Gear Maniac
 

Quote:
Originally Posted by collo View Post
Nice photos - good progress!
Thanks for that, been hectic. Just hope it is not raining on sat (raining as I type)...imagine a couple of hundred wet pairs of feet thru the house!

Might be able to make a bit of progress in the next few weeks, is only three weeks till bathurst (been a year already?) and the gang then arrives.

BUT, I really should do a video shoot...maybe the diffractal if we get one down.

(thanks for the design suggestions...you fell into my cunningly laid trap!)

Quote:
Funny you should mention the 25mm limit. I've just spent several hours re-writing the section on the webpage about fine tuning the well widths.
This is in advance of making the analyser in the calculator a bit smarter.
It makes provision for one of the rules to base well width on build depth rather than design frequency. Have a squiz and let me know what you reckon.
Ahh, thanks for reminding me. I did note the new additions and explanations you had included, forgot to mention that.

It is even more comprehensive now!? wow, gotta say excellent work.


Quote:
Been thinking about diffractals too. It's easy to fire up two copies of qrdude - one for the panel itself, and one for the well inserts.
haha, you too after xenons thread eh? that is basically how I did it, have two running simultaneously.

It is a good point, do we exactly match the cutoffs from one to the other, or overlap as you suggested.

lupo had a wholistic type of theory going regarding all that, don't know if it was true but it made some sort of sense, in an 'om' kind of way!

Pity I eat pies and hamburgers eh, not that healthy organic mush heh heh



Quote:
Should give coverage from 351 hz to 19111 hz
(Being from Bathurst, you would realise the magic in the 351...) heh
hah, the boys go to the races, I stay here. It's always the way aint it, I lived in the blue mountains and never even visited the three sisters.

be buggered if I am going to deafen myself at a six hour race. (bikes? then I could be tempted, but even then rather watch on tv and see it all)

Will re-run your figures, thanks for that.

What about your build, is the design starting to firm up yet?

Got space for a 1.8 m diffractal? Then work the figures up and down till you have the best you can get, I'll steal them for myself and we can build both our qrd's.

I can help you with the build.
Old 17th September 2009
  #161
Gear Maniac
 

argh, where did you get a well divider of 1mm from?

I will admit I am tending toward a higher prime to fill the space (around 1.8), rather than repeats of a lower size.
Old 25th September 2009
  #162
Gear Maniac
 

Update.

Ok, the National Trust visit went well, we were blessed by warm dry weather, so that was good with a couple of hundred pairs of feet thru the house!!

Made another coupler in addition to the last photo.

The last pic I showed with the second delicately balanced on the first (held up by a little pole only)...well the cat decided to jump on it didn't he.

Yes, you guessed it, the whole lot came crashing down, and with major damage to it! arrgghh.

Man, the simple expedient of taking it down after the photo......

So, no time to get the coupler finished before the bathurst gtg, so on the back burner for yet another few weeks.

Way back in the thread we designed a possible tri fractal, with collo's qrdude I have settled on that final design. Will start that build on the weekend, give data then.

The main reason for this post is to assemble the 'confusions' (prob only held by me) surrounding the barker code. Then with glenn's and jasons indulgence, hopefully we can get it all sorted.

Ha, glad glenn asked me top get the question together, I'd thought it was all in this thread, but a lot of it came from xenons (QRD Prime + Period Question (Live Room) )thread...he would not have been able to find it all!

So to recap, here is Jasons explanation of how to implement the barker code.



We need two different diffusers for this to work. It can be the diffuser and
the same diffuser "upside down"...or it could be a 7 well diffuser and a 5
well diffuser. Pick one of these (or one configuration "right side up" or
"upside down"). This will be diffuser 1. The other diffuser or configuration
will be diffuser 2. Set them up like this:

1112212 or 2122111

You have now modulated your sequence. Stand up with pride and make that
"Star Trek V thingy" with your right hand.

Jason

P.S.

To modulate a row of 7 diffusers, you need a diffuser that is asymmetrical.
Most are by nature since the number theory sequences used are not symmetric.
One exception however, is the RPG QRD's. They take the last well that has a
depth of 0 and put half on the far right and half on the far left. This
makes the diffuser symmetric. If you line them up all the same way this
makes no difference. But, if you want to use a modulation sequence (which
you definitely should) you need that asymmetry.



Trouble is that has been interpreted in different ways, as follows.



Quote:
Originally Posted by terry j View Post
I think we are (again???!!) mixing terms? I mean by rotating 180 the wells go 1234567 to 7654321, rather than 'turn it over' or upside down.

Think pinwheel.

At least, that is how I understand it.


Quote:
Originally Posted by collo View Post
Ah, there lies the confusion.

Because the panels are basically symetrical, your "flipped" panel would give the same diffusion pattern - no use as the alternate panel in a Barker sequence.


I think the more general usage of the term "flipped" is as I referred to above.


Quote:
Originally Posted by terry j View Post
As I said, I could have completely misunderstood it, but to me that says 1234567 goes to 7654321.


But the trouble is, I can't interpret Jasons comments and get your answer. Not saying you or he are wrong, only that I just can't see any other way of understanding what he said, which I take to mean rotating the existing qrd as described before.
Quote:
Originally Posted by collo View Post
Terry,

the way I see it, the statement from Jason, and the comment from Ethan's forum are talking about two different things.

Jason's comment:
or you can use a diffuser and for your "second" diffuser you flip the diffuser so it is oriented the opposite way of the first

I interpret this to mean turning through 90 degrees, so that the fins that are normally vertical are now horizontal. (the key word being "oriented")
Now, I interpret collos version as meaning the following pic



which clearly does not work, would be worse if you'd made the diffusor longer than 500mm as here!!

This is having the fins horizontal (or vertical) ie 90 degrees in applying the barker code.


Quote:
From Ethan's forum
We need two different diffusers for this to work. It can be the diffuser and the same diffuser "upside down"...

I interpret this to mean using the back of the panel for the inverse as I described above with the drawings.

The statement only makes sense if you imagine the panel lying flat on the floor. Of course once you put it on the wall, "upside down" infers what you took it to be.

From Ethan's forum
Pick one of these (or one configuration "right side up" or "upside down"). This will be diffuser 1. The other diffuser or configuration will be diffuser 2. Set them up like this: 1112212 or 2122111

The number sequence described is the modulation sequence for the panels, not an indication that the order of the wells of one panel is a mirror image to a normal one.

Maybe everyone should adopt a convention such as....



I wonder what terms NASA uses to describe the motion of the space shuttle.



The drawings done a couple of posts back were done by hand to illustrate the point.

Here is how I interpreted jasons post



it won't let me make it larger, but hopefully you can see that for the -1 all I did was rotate it 180 degrees, meaning the wells went 1234567 to 7654321.

If collos version was correct, then for example there would be no need for an asymmetrical diffusor vs a symmetrical one. Also, if you need to use the inverse rather than my interpretation, that means you have an awful lot of wasted depth (the -1 panel is deeper than the +1 panels)...you may as well go slightly wider, enough to no longer need the barker code, make all of them as deep as the inverse would have been and diffuse lower.

Jason, can you see enough here that you can understand the confusion, and then be able to clear it up please???
Old 25th September 2009
  #163
Gear Nut
 

Quote:
Originally Posted by terry j View Post

Here is how I interpreted jasons post



it won't let me make it larger, but hopefully you can see that for the -1 all I did was rotate it 180 degrees, meaning the wells went 1234567 to 7654321.

If collos version was correct, then for example there would be no need for an asymmetrical diffusor vs a symmetrical one. Also, if you need to use the inverse rather than my interpretation, that means you have an awful lot of wasted depth (the -1 panel is deeper than the +1 panels)...you may as well go slightly wider, enough to no longer need the barker code, make all of them as deep as the inverse would have been and diffuse lower.

Jason, can you see enough here that you can understand the confusion, and then be able to clear it up please???
Yes, that is what I meant. You got it!

Jason
Old 26th September 2009
  #164
Gear Maniac
 

thanks jason, I can see from your post you are a busy boy!! and leave the net chitchat to the boss glenn...been meaning to ask, I assume the G and the K stand for Glenn Kuras??, if so what is the I??

Anyway, tail wagging the dog?? hahaha. Or does the boss just put his feet up and let the workers do the work hehheh (Go Glenn, that's the way to organise it!)

Can I ask you to look at this please??

I think this goes a long way in explaining why some (most?) thought that by rotating the same panel 180 may not be the best solution, the phase angle scattering does not seem as complex as a true inverse

2 ch listening room, aiming for best possible.-qrd-rotate180-vs-inverse-2.gif

Maybe the answer is somewhere in the definition of 'what is good enough' or something.

Whilst I can see the point being made with that diagram, I am still left personally with the idea of 'why waste all that space behind the +1's'???

Engineering is always swings and roundabouts, here is one example I spose. Save the space behind them and go as deep as possible would be my direction, and accept that the breaking of symmetry by rotating 180 is 'good enough'.

To all of us neophytes out here in qrd land, I guess the ultimate authority is The Bible, is there a reference in there that you are aware of?

Or are we, as has been suggested before, simply getting too complicated for our own good.

(it is 'important' to clear this up...by tomorrow I will have five diffusors that I need to organise correctly for the barker code, happy to go with what has been established here tho!! on the basis of the reasoning already given)




[IMG]file:///C:/DOCUME%7E1/General/LOCALS%7E1/Temp/moz-screenshot.png[/IMG][IMG]file:///C:/DOCUME%7E1/General/LOCALS%7E1/Temp/moz-screenshot-1.png[/IMG]
Attached Thumbnails
2 ch listening room, aiming for best possible.-qrd-rotate180-vs-inverse.gif  
Old 29th September 2009
  #165
Lives for gear
 
Nordenstam's Avatar
 

Hi!

Quote:
Originally Posted by collo View Post
Proposed rules for auto-well width

How do these sound?

- Small enough so that HF cutoff excedes design frequency times four. By how much??

- Should not be smaller than 25mm

- Large enough for period width >= 1 wavelength of design frequency
ie Panel width >= 2 * design depth
This is generally met for panels with N>7
If it's close, it will vary with fin width
Sounds good! I don't know how much the HF cutoff should exceed frequency times four. It's a guesswork.. It should be proportional to base frequency. (ie 2000Hz with 200Hz overshot and 20000Hz with 2000 Hz overshot) - or similar.

1 or 2 (music log scale) tones seems like a minimum to me, but that's a pure guess.
Old 29th September 2009
  #166
Lives for gear
 
Nordenstam's Avatar
 

Quote:
Originally Posted by spm_gl View Post
I thought I'd take a quick plunge into the discussion here. The ripple tank applet is a nice little programm, but I'm afraid you guys are putting too much faith into it. I've spent the last week researching acoustic simulation algorithms, and I really doubt a simple reflection like the applet seems to do gives accurate results. Good for a quick demonstration, but no good for judging how different diffusor configurations compare.
Excellent point. It's not exactly BEM. :D Any recommendations for such software at a reasonable prise?
Old 29th September 2009
  #167
Lives for gear
 
Nordenstam's Avatar
 

Quote:
Originally Posted by terry j View Post
So to recap, here is Jasons explanation of how to implement the barker code.



[B][I]We need two different diffusers for this to work. It can be the diffuser and
the same diffuser "upside down"...or it could be a 7 well diffuser and a 5
well diffuser. Pick one of these (or one configuration "right side up" or
"upside down"). This will be diffuser 1. The other diffuser or configuration
will be diffuser 2. Set them up like this:

1112212 or 2122111

You have now modulated your sequence. Stand up with pride and make that
"Star Trek V thingy" with your right hand.
So what's the idea with this?

Any QRD sequence (symmetrical) becomes the exact same sequence when flipped backwards, except that the zero well shifts from left to right. What good is that going to do? It's only an itsy bit different than the standard panel. I just can't see how that can produce much of a real world difference. Please explain!
Old 29th September 2009
  #168
Gear Addict
 

Quote:
Originally Posted by Lupo View Post
Excellent point. It's not exactly BEM. :D Any recommendations for such software at a reasonable prise?
The software is not the problem. The open-source Scilab and Octave (matlab equivalents) can do BEM, and there are some more obscure progs available too. The problem is setting up the equations. It's not a question of point-and-click.
I suppose with enough manpower, one could set up a BEM simulator especially for diffusors, using open source tools, that would then be easy to use.
Old 29th September 2009
  #169
Gear Maniac
 

Quote:
Originally Posted by Lupo View Post
So what's the idea with this?

Any QRD sequence (symmetrical) becomes the exact same sequence when flipped backwards, except that the zero well shifts from left to right. What good is that going to do? It's only an itsy bit different than the standard panel. I just can't see how that can produce much of a real world difference. Please explain!
You askin me???

Me, I am just a dumb schmuck trying to find out!!

Collo has very much the same doubts, so he too would like to get a bit more data on this whole question.

arrgghh, barker code, asymmetric symmetric, modffusor and bodffusors....

c/mon, give a poor anchor a break will you guys???
Old 29th September 2009
  #170
Lives for gear
 
Nordenstam's Avatar
 

Hehe.. Was asking the author of the quote; Jason!
Old 29th September 2009
  #171
Lives for gear
 
Nordenstam's Avatar
 

Quote:
Originally Posted by spm_gl View Post
The software is not the problem. The open-source Scilab and Octave (matlab equivalents) can do BEM, and there are some more obscure progs available too. The problem is setting up the equations. It's not a question of point-and-click.
I suppose with enough manpower, one could set up a BEM simulator especially for diffusors, using open source tools, that would then be easy to use.
Interesting! Though I'm not good enough in math to dabble in that yet.. :( Will hopefully be able to have a go at it when I've done the EE degree starting next year.
Old 30th September 2009
  #172
Lives for gear
 
PaulP's Avatar
 

Quote:
Originally Posted by Lupo View Post
It's so F'ing neat!

Suddenly made acoustic prediction a lot more interesting!

Am done with the styrofoam 2D diffuser project now and have a few bits of scrap that I'm planning to turn into QRD's. Been pondering for a long time on the need for well dividers. They're of course better if its' practical to do so. With the styrofoam, the easiest way is to just glue it up as is, without dividers. This made me feel better about dropping the dividers:

Video clip at: Acoustic diffuser simulation - wells vs well-less on Vimeo

It's startling how little difference there is!
Thanks Lupo for the link to the ripple tank simulator. I've been playing
with it and agree that it's great. It might help if we knew what frequency
was being simulated and the relative dimensions of the diffusors you sketched
before considering how well or not they work. I imagine that they
behave quite differently at different frequencies. Perhaps you have
already done this in which case I'd be interested in your thoughts.

I've been playing with polys. While one seems to work well enough, putting
two side by side with a space between causes some weird effects.

Paul P
Old 9th October 2009
  #173
Gear Maniac
 

for a bit of a muckaround, I replaced the diffusors I had behind me (the 'kinetic noise' ones) with a new design.





A bit unclear, sorry, hope you can make it out.

They are the difractal part of a trifractal I will make later in the room build...unless I change the design again hah hah, what you will see hopefully is that it is a seven prime base, with a five prime nested with the seven.

(and I hope it resizes!)







At a later stage, each of these will be inserted within yet a larger qrd, from memory a thirteen prime??, will have to look it up.

But anyway just an excuse to use QRDUDE from collo, (who did the hard work and developed it with a LOT of help and advice from lupo I would like to point out), and go back and forth matching frequencies etc with available wall width to come up with a finished product.

Just as an interesting observation, really all you need to cover the whole range of frequencies is a difractal.

It was only the available width that I had (which I wanted to fill fully) that meant I had to go trifractal, else with a difractal it was either 'too narrow' for the space, or 'too wide' for the space....if that made sense.

Anyway, BIG weekend this weekend, all the guys from over aus are arriving for a few days of music, chats, meeting old friends and of course, LOT'S of music.

As the cowboys say, YEE HAH!!heh
Old 9th October 2009
  #174
Gear Maniac
 
collo's Avatar
 

Nicely done mate!!
Old 9th October 2009
  #175
Lives for gear
 
Nordenstam's Avatar
 

Quote:
Originally Posted by PaulP View Post
It might help if we knew what frequency
was being simulated and the relative dimensions of the diffusors you sketched
before considering how well or not they work. I imagine that they
behave quite differently at different frequencies. Perhaps you have
already done this in which case I'd be interested in your thoughts.
I just tuned the wavelength til I got the desired effect.

The pictures I made can be seen to have a wavelength that is very roughly 2/3's of the deepest prime 7 depth. The full wavelength of the diffuser is 7/4*2=3.5 times the deepest well. So if the wavelength is 2/3 of that well, it's about 0.66/3.5=0.18 times the diffuser effective depth. Given that these numbers are from quick dirty eye sight "measurement" and very rough estimates, and the QRD does give the expected effect, it's not unrealistic to assume that it's tuned to five times the base frequency.

Changing frequency does indeed give a very different effect. It's clear that it's tuned device that only works as intended on certain frequencies.


Quote:
Originally Posted by PaulP View Post
I've been playing with polys. While one seems to work well enough, putting two side by side with a space between causes some weird effects.
The curse of repetition! In addition to the parapolic effect between polys, they are also lacking in temporal dispersion like one gets from diffusion.
Old 9th October 2009
  #176
Lives for gear
 
Nordenstam's Avatar
 

Quote:
Originally Posted by terry j View Post
They are the difractal part of a trifractal I will make later in the room build...unless I change the design again hah hah, what you will see hopefully is that it is a seven prime base, with a five prime nested with the seven.
This looks superb! Very impressive build.
Old 9th October 2009
  #177
Lives for gear
 
Schaap's Avatar
Quote:
Originally Posted by terry j View Post
At a later stage, each of these will be inserted within yet a larger qrd, from memory a thirteen prime??, will have to look it up.

Diffusors looks great! Don't forget the small QRD's!! I've built a lot of 'antiflutters' heh lately and I'm very excited about how they works. It takes away the 'hazy' , edge etc of e.g vocals, mixes. Basicly it is just a QRD with a deepest well depth of 2.6 cm and in my case well width of 1.2 cm.
Making is pretty easy because I use wood with 2.6 cm height and the widths is standard available in the DIY store-> that's why I use 1.2 cm.
Glue the wood on a MDF panel with 1.2 cm space and you have automatically the deepest wells. The other depth I do with a handheld machine on .65cm and 1.3 cm.

Henk
Attached Thumbnails
2 ch listening room, aiming for best possible.-antiflutters.jpg  
Old 9th October 2009
  #178
Lives for gear
 
PaulP's Avatar
 

Quote:
Originally Posted by Lupo View Post
The pictures I made can be seen to have a
wavelength that is very roughly 2/3's of the deepest prime 7 depth....
Thanks for the additional info. This thread is fantastic.

Paul P
Old 29th October 2009
  #179
Lives for gear
 
Nordenstam's Avatar
 

For posterity, here's a link to another very interesting discussion on QRD's, this time in 2D: Diffusor baffles at Skywalker Sound
Old 9th November 2009
  #180
Lives for gear
 
Nordenstam's Avatar
 

Another Falstad ripple tank simulation video: 3x prime7 - various frequencies on Vimeo - showing how different frequencies give different response. Low frequencies gives near specular reflection, acting as a flat surface, with higher frequencies having more complex radiation. Notice that the higher frequencies does not all act the same, some of them are closer to the design frequencies while others miss the mark more. Use the pause button to have a closer look!
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