Measuring Diffusivity [moved]

[bert stoltenborg]

viewtopic.php?f=8&t=3831

Now read this and tell if I"m funny.

[Bob]

http://forum.studiotips.com/viewtopic.php?f=8&t=3831

Now read this and tell if I"m funny.

What's the difference between curved and corrugated.
I've seen corrogated cardboard, and it's curved.

I presume curved is poly or RPG's WaveForm ( http://www.rpginc.com/products/splineg/index.htm ) kind of thing.
(I just read the abstract thus far -- but I didn't see any pictures to explain the difference)

I hope corrogated is QRD ( http://www.rpginc.com/products/qrd734/index.htm http://www.decware.com/achowto.htm ) and Skyline ( http://www.rpginc.com/products/skyline/index.htm and that studio with five walls that are super skylines ) and perhaps even Harmonix ( http://www.rpginc.com/products/harmonixg/index.htm ) kind of stuff.

Personally, I've always thought that uneven rocks making a curved wall were very pretty (there's a studio I'm thinking of, but for now I've got http://www.bemidjistate.edu/students/re ... l_kids.jpg ).

BTW, why is there no date on the PDF on page 1?
Where is Figure 3?

Presumably the {"0 degree position" page 7 to "60 degree position" page 9} vs {-40 to +40 degrees on each graph} represent perpendicular axis,
i.e. the -40 to +40 are measurements in the room across the floor, and
{"0 degree position" page 7 to "60 degree position" page 9} are angling the diffuser in an up or down axis.

I presume lots of even red is good. Blue is bad, blue stripes are worse.

[Scott R. Foster]

Thanks for the comments Bob.

corrugated = a surface shaped with parallel ridges and valleys - often regularly/symmetrically spaced and sized.

http://en.wiktionary.org/wiki/corrugated

for example:

http://www.google.com/search?q=corrugat ... 20&bih=645


=====

I think humming birds are pretty

=====

date added [good catch!]

several figures of DUT's and the measurement process were deleted to reduce file size but obviously not all references to same [another good catch]!

=====

If I understand your question regarding axis of rotation, then the answer is "no". The DUT was not rotated [see page 3].

"Each of the three diffusors was in turn subjected to a correlated test sound signal at normal incidence [zero degrees as shown in Figure 2] and discreet measurements were taken at eleven of microphone positions ranging from +60 to -60 degrees. The reflected sound for each observation position was recorded and then the speaker moved along the 4m arc to a different source location ranging from zero to +60 degrees whereupon the measurements were then repeated for each of the eleven observation positions."

===

A perfect "anti-specular" diffuser with efficacy across the measurement band of frequencies would produce a sonogram of a single color from edge to edge [both vertically and horizontally]. Please re-read section 4.1 and let me know if this is not clearly stated.

[Bob]

"Each of the three diffusors was in turn subjected to a correlated test sound signal at normal incidence [zero degrees as shown in Figure 2] and discreet measurements were taken at eleven of microphone positions ranging from +60 to -60 degrees. The reflected sound for each observation position was recorded and then the speaker moved along the 4m arc to a different source location ranging from zero to +60 degrees whereupon the measurements were then repeated for each of the eleven observation positions."

In retrospect that makes sense. QRDs and Polys tend to be effective in a 2D sort of way, rather than a 3D sort of way,
so moving the speaker in the same axes as everything else makes sense,
whereas rotating the diffusion device up/down wouldn't.

A perfect "anti-specular" diffuser with efficacy across the measurement band of frequencies would produce a sonogram of a single color from edge to edge [both vertically and horizontally]. Please re-read section 4.1 and let me know if this is not clearly stated.

I've just re-read section 4.1,
and frankly the way I figured it out by looking at the graphs, and more importantly the indexes on the graphs (vertical 'angle', horizontal 'frequency', and colour 'dB' indexes), was much easier to comprehend.
Hence my simplistic: "lots of even red is good. Blue is bad, blue stripes are worse" two posts ago.
Lots of even is good because that means diffuse.
Lots of red is good because that's the lowest dB reading and the opposite of a massive specular reflection which is blue -- at least from a 'diffusion' and not a 'scattering' point of view. I didn't work out the sum of the reflected energy to see if it was the same as the speaker energy (i.e. if the first device is really a super absorber and cheating the whole way).
Lots of even red is similar to your "sonogram of a single color from edge to edge" (although yours is better since it could theoretically include a sea of blue although that woud be some sort of magical amplifier diffuser device)
Blue stripes are bad because that's lobing -- either by angle or by frequency or both.


I presume the ~300hz to 10khz is a room or device restriction. (as opposed to say 50hz or 100hz).

I also assume that if you'd tested more than one poly, as would be used in a room, you'd end up with lobing effects. That you're not yet ready to make a statement about if small or large polys should be at the front of a home theatre sidewalls (i.e. front to back: {large, medium, small} vs {small, medium, large} vs {medium, large, medium}), or if one poly should span the entire back wall vs three different sized ones vs ten thousand really tiny ones.

[bert stoltenborg]

Thanks for your comments, Bob.

These sonograms consider the 0 degree measurement as a reference, so speaker inaccuracy is eliminated, the speaker is corrected like it is lineair.
The off axis measurements should, when a perfect diffuser is at work, have of course the same frequency response as the reference.
Hence a perfect diffusor would have a monochrome color plot.
The more color differences, the more deviation from the ideal lineair frequency response (or call it lobing or unwanted phase shit or comb filtering or absorbtion or whatever).


http://www.alconsaudio.com/site/product ... st_eng.pdf

In page 6 of this test you see a sonogram of a line array; maybe this clarifies stuff.

[Scott R. Foster]

I also assume that if you'd tested more than one poly, as would be used in a room, you'd end up with lobing effects. That you're not yet ready to make a statement about if small or large polys should be at the front of a home theatre sidewalls (i.e. front to back: {large, medium, small} vs {small, medium, large} vs {medium, large, medium}), or if one poly should span the entire back wall vs three different sized ones vs ten thousand really tiny ones.

I beg to differ.

[Bob]

maybe this clarifies stuff.

It was already clear the moment I saw the sonograms in yours and Scott's paper -- that's a great and obvious way to explain it.
But thanks for the additional words.

You play with speakers a lot, so sonograms are probably your favourite graph.

[Bob]

I beg to differ.

Do you mean multiple similar sized polys would have no lobing? (contrary to Master Handbook Of Accoustics)

Or do you merely mean you have a statement/opinion, perhaps with or without a correspondingly similar study you haven't yet shared.

[Scott R. Foster]

yes

[Ido]

bravo!
it's nigh time this whole diffuser thing was torn down to bits in order to be better understood, and to enable creation of better diffusers, whatever they may be.
this is just the beginning.
The acoustic users, pros included, have IMO mostly been force fed with existing diffusing tools (even "good" ones), instead of outlining the diffusion goals, and then trying to accomplish them.
so to this I add my constant nagging question, what is the ideal diffuser? what would we want it to do? and in what color?
thanks Bert & Scott

[Scott R. Foster]

The goal Bert and I share is not promoting a particular design for a particular circumstance but rather that folks recognize that measuring diffusion devices is within the grasp of most folks operating at or above the "advanced user" level and therefore should become a practice just as common as the measurement of loudspeakers. We believe modest success in that goal will lead to a common understanding of how to optimize.

[Bob]

yes

Ha !


________________
you remind me of the ancient skeptic who demanded of the wise old sage to be taught all the world's wisdom while standing on one foot
-- Dr. Tristan Adams

[Ido]

ah, Bert has a goal..

[Ido]

this reminds me of one of sarge Foster's quotes, which went along the lines that "..some of the best engineering lies rusting away in junkyards..."

[Ido]

looking into it again, some mild remarks:
I'd change the subject from "Using Traditional Methods" to something that emphasizes the un-traditional-ness of this work.
I'd include pics of the DUT, especially the Geek.
I'd try and make the intro easier to understand.
next step: correlation between these results and real life-human subjective evaluation (whatever that is).

[Tubamark]

looking into it again, some mild remarks:
I'd change the subject from "Using Traditional Methods" to something that emphasizes the un-traditional-ness of this work.
I'd include pics of the DUT, especially the Geek.
I'd try and make the intro easier to understand.
next step: correlation between these results and real life-human subjective evaluation (whatever that is).

I'd also prefer a different term than "corrugated". I initially expected (and eagerly hoped to see) measurements of various corrugated sheet materials (industrial siding, roofing, etc), not a specific device that already has a well-understood name(s) like "QRD", "2-D Schroeder diffuser", "quadratic phase grating" etc.

Aside from that, it was a great paper that demonstrated what many never considered, and a few have knowingly avoided: 1) The QRD isn't ideal, 2) other methods can work as well or better, and 3) we should measure (and perhaps demand) meaningful performance data on commercial acoustic devices.

'Second the vote for statistically evaluating real-world perception of diffusion in-situ. Not everything measurable matters, and perhaps not everything that matters has been measured.

-- Mark

[Scott R. Foster]

Thank you for your comments TM... next time Bert and I chat I'll ponder with him about a more generally descriptive and accurate word than corrugated.

[Ido]

... next time Bert and I chat I'll ponder with him about a more generally descriptive and accurate word than corrugated.

koerrugated?