Before we get started with diffusion, I wanted to post this information concerning corner bass traps.
So here is a basic layout of how your corner bass traps should be built. These are full length from floor to ceiling and are built directly into the corner walls. The panel is diaphragmatic, and must be unobstructed in order to operate properly through resonance of frequency. That resonance factor is determined by the thickness, and rigidity of the face panel. High grade plywood is a very good material to use. The panel edges are sealed in place, but are free to vibrate within the mounting channels. Make sure that the fiberglass behind the panel maintains an airspace between the two or the performance will suffer.

Okay, so that should just about cover the subject of absorption control.
Now we are ready to begin covering diffusion concepts.PLANAR BINARY AMPLITUDE DIFFUSION: B.A.D. Panel
Sound is attenuated by absorption, re-directed by reflection, and uniformly dispersed by diffusion.
By varying the phase of the scattered waves, we can effectively diffuse the sound waves. The binary amplitude diff-sorber maintains a well controlled combination of absorption and diffusion, using a variable impedance scheme comprised of reflective and absorptive areas. These areas provide pressure reflection coefficients by a
factor set of 0 and 1 per a Maximum Length Sequence, or M.L.S. The M.L.S. Is based upon shift register theory, rendering it an ideal candidate for successful implementation.
This will achieve a flat power spectrum using values of easy computation. Most importantly, this technique provides an important and valuable asset to your audio room.
Your listening space will take on a vast new sense of spaciousness while maintaining natural ambience. The ambiance of your listening space is
delicate, and extremely
vulnerable to acoustic influence or interference.
There is a drawback to where the
standard B.A.D. panel falls short of optimal performance, and there is a solution. In the upper frequency range, as there becomes a reduction of the binary patch below ½ wavelength in proportion, specular reflection is restricted to an attenuation of approximately 6dB within a surface area that is fifty percent absorptive. This situation is corrected by a new “
ternary” sequence which based upon
interference, has the ability to reflect waves in an “
out-of-phase” condition, which in turn, forms a dramatic reduction upon specular energy.
This modified sequence is comprised of
-1's, 0's, and +1's. The major change here of course is the addition of the final well designed with a depth of
¼ wavelength at the design frequency, the Fo, and odd multiples of this frequency. This “
end well” now produces a
reflection coefficient of -1 to the incoming wave. Therefore, you now have a well that transforms these waves “
out-of-phase” with the primary influence responsible for the specular energy.
This infiltration upon areas
R = +1 results into substantial reduction of that specular energy, and better diffusion is the gain. There is even one more advancement that can be initiated to provide even further gain in performance, that being another design approach, the “
Quaternary” number theory sequence. I am not going into that technique any time soon due to complexity. For now, I am satisfied with the advanced ternary sequence in order to get the job done. But wait!...there's more on where this design can go. I shall touch on this later in this writing.
If your space is
large enough to allow
full projection of bloom between a diffuser and you, then I would definitely suggest the Quadratic Residue Diffuser approach. It would still be beneficial to take advantage of the diff-sorber binary amplitude concept, just as an auxiliary acoustical treatment along with the higher prime element 1d diffusers, BUT,
only when placed to avoid repetitive lobing derived from periodic sequences which causes destructive interference on it's own part. The way to avoid this problem is by using
a-periodic placement with properly inserted panels of
“inverted” wells.
The effective coefficient of the diffusion will increase substantially.
So here is the finished
TERNARY design which I expanded in form by the implementation of folding, and inversion to create the higher order sequence.

For the
Binary Amplitude Diffuser, the auto-correlation function of the surface reflection is responsible for even dispersion well into the distant domain. For the
uni-polar model (standard B.A.D. panel), cancellation within the side lobes are due to the resulting reflection coefficients being either
zero or
one.
But for this improved ternary design, the factor of optical sequences are justified. For the improved ternary sequence, the
auto-correlation function allows cancellation within the side lobes resulting in optimal total side lobe energy.Diffuse reflections are those dispersed in space with both spatial and
temporal factors
intact. An “ideal” diffuser would be expected to generate the spatial and temporal spread evenly across the entire audio spectrum within the human auditory range. Ideal sequences are those with strong auto-correlation qualities.
Strong auto-correlation properties will result into a flat power spectrum, in turn resolving the scattering cycle consistently throughout with great diffusion as a result.
An important factor to incorporate into the diffusion process is that of
A-periodicity. This is achieved by using a long number sequence with good auto-correlation properties. One workable solution is to use a modulation concept. What really works well is to
establish an array with the diffuser, and its own inverse, however, these must interchange to form an “A-periodic” formation.
I suggest the use of Binary-Amplitude-Diff-sorbers (B.A.D. Panel) in the
curved version to suppress sidewall reflections as an alternative method to simply using panel absorbers. This in basic form started out as a flat hybrid surface utilizing a “zero” depth perforated mask over a porous absorbent material.
These panels are diffusive from about 1k on up with a transition to pure absorption below that frequency.These were developed utilizing the Chinese Remainder Theorem in which to fold one dimensional sequences into two dimensional arrays with preservation of the desirable auto-correlation and Fourier properties responsible for their performance properties.
More recently, this concept found improvement by use of smaller variable size square panels. I have the pattern and design layout for these panels as well. I will cover this design process further on in this writing.
Through a logical grouping manner, good diffusion remains far less affected by
lobing, which normally results from periodic repetition. Through variable placement of various depths of these panels in which to form an
a-periodic sequence array, considerable improvement in performance comes into play when grouping a large array of these panels together.
The “flat” version of the B.A.D. panel Inherently produced specular reflections that were not exactly optimum for a great diffuser. The answer to correcting this problem was through incorporating
curvature to the binary encoded perforated panel. This bending of the flat surface resulted in redirection of the binary wells to
now operate out-of-phase, therefore drastically reducing the specular reflection, while maintaining the absorption coefficient, but with far greater dispersion patterns.
The curved B.A.D. panel enriches the sweet spot for greater spatial interpretation, and overall clarity.
The design of the binary amplitude diff-sorber panel is to achieve
diffusion at or near one octave over the frequency for which the wavelength is two times the binary element in surface area. This creates a uniform diffusion pattern which is spread over a wide frequency range.
So additionally,
I propose yet another alternative to standard Q.R.D. diffusers. This model is derived from simpler two dimensional sequencing, based on the Binary-Amplitude-Diffuser/Absorber concept.
The design layout as you see in three stages of development here results from folding, and inversion, until a full, high order grid evolves into a very powerful tool for your room acoustics arsenal. This device can work
similar to the ternary version, in that you can
fine tune your room like you fine tune a speaker's crossover network in order to achieve the sound quality that you desire.
You can easily see the degree of built-in diffusion compared to the amount of absorption built into these panels. The
blue segments represent the hard diffusive areas or
1, and the
white areas represent absorption or
0.
Now this model is quite a bit different from the Ternary diffusion concept, as the Ternary version is much more complex, and far more difficult to build. This version is built directly onto the wall with interconnecting panels in a proper sequence which allows it to evolve into the higher order grid. This grid is built in smaller sections as dictated by the full grid sequence. You start with a .75” plywood backing with multiple layers of semi-rigid 1/2” fiberboard sheathing (
Celetex High Density Fiberboard) built up over the plywood. This board material is very dense and fibrous, while still manageable, and much safer for your health to use.
Multiple sheets layered with air gaps sealed in-between, create a very good absorption coefficient, while providing an excellent base to build the diffusion segments upon. The
reflective segments must be a very hard material, and non-absorbant to the working frequencies of this design.
At the very least, you need hardwood, or a thin metal covering to insure a surface which can't be penetrated. Do NOT use MDF or soft wood, unless you glue a metal face to the segment. You can buy sheet metal to cut this from. Thick HVAC aluminum tape
could work on a wood base, but something thicker might be in order. Several coats of hard urethane applied to hardwood should do well by itself. The pieces that make up the segments are all identical and measure
.05” square by .05 “ depth (1/2” cubes), glued directly to the top sheet of Celetex fiberboard.
A grid pattern template will be necessary to install the segments in proper and uniform sequence. This will be used repeatedly for quick, easy building of the unit. Your grid is already built for you and waiting to be picked up at the local big box store. You get a ceiling light grid for about $15 which is plastic, and just happens to be a bit over ½ inch by ½ inch square wells, and the perfect depth to work with. These lighting grids are big enough for the design layout of one section which is a 35 x 36 well grid.
If you want to step up the design a notch, I recommend actually using the
plastic grid panels permanently as the top layer. You simply cut the wood cubes precisely to fit snugly within the plastic grid wells and if they are snug enough, they won't require glue to keep them in place. I have already done this technique and it works very well. You need to be absolutely precise and double check the fit of your cubes as you cut them. It sucks to have to throw out a bunch of hardwood cubes that not only are expensive, but you just spent half a day cutting them. Always use quality control measures as you do batch work. An alternative which is easier, and will guarantee success, is that of
pouring liquid plastic into the wells requiring reflection, instead of cutting all of those wood cubes. It is your choice as to what works best for you.
You will make a reusable template with this lighting grid by painting the appropriate
absorption wells red, and leaving the
diffusive wells white. DO
NOT screw up the grid sequence, as the mistake will be repeated
over and
over again until it becomes a
major flaw within the sequence. Take pride in your work, and absolutely
double check your layout.
You only have to lay out one section to be used as your production
template. This allows easy, and fast placement of the diffusion segments as you glue them in place to the fiberboard underneath. All you have to do is
flip the grid and
turn it as the full grid sequence dictates. Very simple to do, and allows rapid development of large scale units.
I would suggest screwing the plywood onto your wall
directly to the studs which should be 16 inches on center.
Secure this panel to your wall
well, as it is going to be heavy when finished. Then you can build onto the face of the plywood backing. Since the celetex sheets and the plywood are about the same size, go ahead and mount the material full size to the plywood.
Do not mount face to face flush. You will build a wood frame around the perimeter and also a center strip for which the next layer attaches. Do this with every layer. A
half inch airspace should suffice between layers. The last layer of celetex could be mounted to the wall after you glue the diffusive segments on, it is your choice. It is easier to glue these segments in place on a bench. I suggest three layers of celetex total over the plywood, which gives you
3 inches of absorption including the airspace allotted between the layers. The airspace provides a much greater absorption coefficient, therefore
do not omit this important factor.
The one modification that I strongly recommend is the addition of a
full facia which covers the entire face side of the absorption material (a very
thin hard layer), just as with the Ternary Diffuser/Absorbers, Why is that, and what purpose does this serve?
The reason for this is a very logical approach to fine tuning the degree of diffusion and absorption ratios in a controlled manner, by order of
frequency transition modeling. The thickness/hardness of the facia becomes the crossover point where you want absorption coefficients to take affect,
leaving frequencies above that range untouched and un-absorbed. This
prevents over absorption of
delicate higher frequencies in order to
preserve ambient spatial bloom within the room. At that determined transition point, full absorption begins to take affect to extend well below that point.
More to follow soon, please check back as we continue.