EnABL Processes

[BudP]

JoninCR

Well, this is good, you are forcing me to think, excellent.

The last thought, the one concerning the multiple vent "mesh" with Mamboni treatment, was not intended to dissipate the back wave. Instead my thought was to control the back wave front, on the box inside plates, with Mamboni's infinite edge termination process, applied as if the box faces were those of a driver and between the actual driver and the mesh of ports. I am certain there will be some loss involved, not certain how much, as this is a thought experiment with only a small amount of empirical activity to back it up with.

What I am aiming at is a large area vent, used to create additive sub frequencies, in phase with the front wave emissions, physically stepped back by perhaps 18 inches, to maintain some semblance of emitted phase coherency.

If the Mamboni process can be used to control reflections within the box by controlling the boundary layers of the box interior, we might get that bass EQ without added electronics. I have experimented with multiple small diameter ports for bass reinforcement with considerable success. I have not experimented with more than four, but, I see no reason to assume that many more than four would be less effective. These four holes, and Bass Box Pro will model them quite effectively, rather than one big port. have some added benefits. There is a port flow mass that acts much as a shock absorber. When sized correctly the bass is linearized and without the "boing" character so offensive from more normal ported, closed, boxes.

Oddly these small ports do not chuff until a very high SPL is achieved, 103 dB and higher in my environment and this at 35 Hz, from a single 9 inch Dynavox woofer in an 0.8 cu ft box. Admittedly the woofer has been treated with the EnABL process so that the back wave is quite a bit more orderly than is the usual case. This set up will not charge a large volume, but two boxes will perform very competently in a 20 X15 X 8 volume.

So, what I am after here is bass in phase with the emitted front wave , of lower octave content than the front wave, that has the same dynamic characteristic of the front wave and uses the back wave, through the mesh port scheme, with terminated reflections in the box, to achieve this goal. That the phase coherent energy coming off of an EnABL treated woofer will aid in this is certain, I do not believe the scheme would work otherwise.

So this does fall into one of your two categories that are not monopoles but with a 180 degree, hopefully, twist to the bass that just might eliminate that EQ situation. But, again, there is the Gary Pimm factor and the staggeringly powerful and dynamic bass he has obtained. I will attempt to get him to post here, but he is on sabbatical from audio, specifically due to the lack of intelligent response he encountered on another forum, where that sort of response too often seems to be the norm.

I like your Helmholtz low frequency scheme and am going to put that idea to simmer in the back of my brain, I see considerable parallel to the intent of mesh port idea. Except of course that your idea is working in reality rather than imaginality.

Bud

[BudP]

AJ,

For those of us on the side lines of what appears to be a long running fray, would you be so kind as to lay out test types, methods and results that would, in your qualitative judgment, either disprove JoninCR's allegations or allow him to move on in his actual physical experiments? For all I know you have already done this in some other thread and pointing us there would be sufficient I am sure.

It does sound as though you have a rather strong bias against qualitative findings. May I suggest that you have a deficiency in training only. The obvious fact that you have ancestors indicates that you have all of the physical equipment needed to make effective evaluations, of the nature of both static and changing sonic signatures, but may have become determined not to use what a couple of million years of threat assessment of a holistic, hostile sound field have selected for.

Please do not take this to mean I am adverse to quantitative measurement, I find it invaluable in my profession. The point I am pushing towards is that our hearing is a very specialized, highly correlated, self protection device.

The semi autonomous correlator that sits on top of the raw sense data, which is not materially different from that received by a test microphone, is scanning for relative data, what is different about the constructed data field that has been rendered, in a 360 degree assessment of threats and non threats. This correlator can, with relative ease block a known benign sound to enable new, unknown and possibly threatening sounds to be addressed. This goes on below the typical level of consciousness 24 hours a day.

The level of discrimination is superior to that of a spectrum analyzer, but it is not an objective discrimination, without training. When I say superior to a spectrum analyzer I am referring to the "information in the grass" and the ability of the correlator to create an understandable and actionable data set from what just looks like noise on a display screen.

Even with training there is not any sort of hard copy printout available. This does not in any way prejudice the results derived from this extremely fine tuned discriminator of all of the values required, to make a correlatable sound field that is understandable to our semi autonomous threat assessment system. Neither you nor I can make any sort of "objective, quantitative " determination of JoninCR's sound field projectors. We have zero quantitative data. And, since we have not heard them, we have no qualitative data either.

Without both, no judgments can be made on the merits of his creations, except those he makes about the recreation of an understandable sound field and qualitative measurements have reliably proven to be a useless guide to what is and what is not a correlatable recreation of an understandable and "pleasing" recreation of a distant sonic event. So, I am skeptical of the value of objective tests without also accepting Jon's qualitative data.

We can, however, listen closely to his qualitative comments to get some notion of his sound field projectors value in recreating a sound field that is intelligible to our correlator based "hearing".

He does after all have the same lineage of ancestors, ones who were able to assess and so left us capable of assessing a very complex threat field, over time, distance and vector, that allowed them to remain uneaten and us to enjoy a mock up of reality, portrayed with enough clues for our correlator to reconstruct a believable event from what would otherwise be meaningless data.

Not useless data, just meaningless. We are, after all of the measurement and analysis is done, the only judge of what is good and what is bad in sound reproduction and that can only be a qualitative judgment.

I also am really tired of the excuse commonly used for tossing qualitative evaluations, "that we all hear differently". This is not correct. We all hear the same, after a few million years of truly excellent predator selection of our species, we must all hear very close to the same. It is only our consciously trained, conscious listening skill set, that varies.

Bud

[Lynn Olson]

JohninCR's split-path design is most interesting - a TQWT in acoustic parallel with a short open-back box. The backwave goes in two paths, with quite different delay times, in addition to the nominal back-to-front delay distance. Since the TQWT is folded, there's a bit of low-passing going on in that path.

It would be most interesting to have the spectra of the rearwave escaping from the back of the short box as well as the wave escaping from the exit of the TQWT. When I measured the spectra of the wave emerging from the exit of the Ariel TL, I was quite surprised to find it was flat (1 dB) from 30 Hz to 100 Hz, and dropped off quite rapidly above and below those frequencies. I surmise the complex labyrinth was doing some low-passing for the TL, since TL's are usually a bit more resonant than what I saw. The drivers (measured nearfield) dropped like a stone below 80 Hz, so the TL was definitely working as advertised.

As for the ideal performance of a simple flat-baffle dipole, yes, it's free of cabinet coloration, but most certainly not free of standing waves on the face of the baffle. When you measure a driver on a standard IEC baffle (80 by 115 cm), it is essential to mount the driver somewhat off-center, otherwise the measurement will be contaminated by standing waves on the baffle. This is audible and measurable, yet it occurs with the simplest possible flat baffle. So flat baffles are not free of standing-wave resonances.

Returning to what I think of as a Progressive-Loss Mesh, it can be applied to the full gamut of shapes, flat baffles, horns, and pipes, progressively dissipating the wavefront as it travels towards the edge of the surface. Unlike damping felt, the frequency characteristics and loss-with-distance can be controlled by the density (pitch) of the drilled holes, which are ideally much smaller than the smallest wavelength of the driver. Drilling hundreds or thousands of holes by hand would be extremely tedious, but this can be automated by a NC-controlled drill press - think of all the holes in a circuit board, for example.

As mentioned earlier, and what seems to escape most high-end speaker designers, it is the edge of the surface (regardless of shape) that creates (acoustic) standing waves. Edges, just as much as hard reflecting surfaces, reflect energy. That's how I look at any enclosure - AR3a, Altec A7, Bose 901, Wilson WATT, you name it - as a collection of edges. The more edges, the worse the sound. House of the Seven Gables might be a good way to sell a $500,000 house, or make a cuckoo clock, but it isn't the way to make a good-sounding speaker.

Just before I left Audionics I built a prototype ultralow diffraction speaker, which ended up looking like a giant vitamin pill supported by a very narrow wood stand behind the speaker. The prototype had lots of issues - resonances in the cardboard tube, mediocre performance from the Audax 2" mid dome - but it delivered on the low diffraction. When the lights in the room were out, the speakers disappeared so completely you could walk right into them while they playing - there was no "point-source" effect at any listening distance. The images simply hung in space, entirely free of the speakers, a most uncanny effect - and this was true of every recording, even old mono LPs.

The commercial MBL system comes closest to what those prototypes did back in 1979. As a result of that experiment, I've become sensitized to the sound of diffraction - which in the simplest terms is what creates the apparent source size when you walk towards the speaker (it is less noticeable when you don't move). When diffraction drops below the threshold value, it becomes very difficult to tell the size of the speaker as you walk towards it. The soundstage loses its hard edge on the left and right sides, and extends well outside the left and right speakers.

Now I want to create a low-diffraction dipole - and I suspect the Progressive-Loss Mesh is a good way to do it, provided the holes are small enough and there are enough of them (hundreds or thousands). The PLM can be applied both to the flat baffle and the length of the short-box/TL, dissipating energy along the length of the structure.

Oh yes - before I forget - Bud's EnABL pattern could also be realized as a set of cutouts, perhaps as very small cutouts closest to the driver, getting bigger towards the edge of the cabinet (highest frequencies are dissipated first). This would be another interesting variation of the PLM.

[BudP]

Hi Lynn,

Thanks for the intro.....

The EnABL pattern could indeed be used as a set of cutouts. Lot of work though. I usually just apply the pattern blocks at every edge termination on the front and back of a driver, it's mounting bezel, the surface it mounts on and the surfaces connected to and incident to that mounting surface. Also a lot of work, but easy and tedious to perform. Not even particularly exacting work either, as the pattern allows quite a bit of flexibility in size and position mistakes.

However, the results are worth it. Invisible speakers. Absolutely no artifacts emitted from the surfaces that are not part of the information packet describing the sonic event being portrayed. And all this with no reduction in SPL or transient color, or micro dynamic detail. It is not a damping process. Just a one way gate, at every edge, that leaves the energy in question with ony complete emission into the air as a remaining path.

You have seen my speakers, they are not simple in shape nor in distribution of masses. I have worked to reduce the number of edges on the outside of the box but for every mounting plate terminus, every driver emitting surface and for the periphery of the front face, there is a full, two row pattern of blocks. I also have a ring of pattern blocks midway back, around the outer cabinet surface. This rings location controls where the front of the sonic recreation occurs, in real space, according to the correlator, which is quite expert at these calculations.

All of the blocks are applied with a pen and some acrylic paint, thickness of about 3 mils max when dry. You can paint over them, shellac over them or leave them exposed, without any alteration in their performance of the task of eliminating standing waves, caused by surfaces with an edge. Anyone on this forum can accomplish this with their speakers, a little time and some easily learned skills.

Truly you do have the correct, important point, singled out here. Once the standing waves caused by transient signal and terminus edges have been controlled, everything else falls into place very quickly and near perfect speakers become not only possible but actually easy to build. Just a lot of tedious handwork.

I have been sent a pair of Lowther's to treat. I could make a training guide, with pictures as I proceed, but, Lowther's are very touchy beasts due to the whizzer cone and a mistake could easily turn them into ultra clear ear wax removers. I will do this in another thread, rather than pollute this one, but I am more than willing to help anyone here learn how to free their speaker designs from a limitation that almost no one realizes is present and addressable, because they have never experienced a speaker without edge defined sonic corruptions. The only place you get that is right next to the orchestra, playing in a field, far from buildings. Usually, they are also marching.

[Lynn Olson]

Originally posted by johninCR

Can I ask a question about the standing waves on a flat baffle? Is that the result of the secondary source coming from the edge (diffraction artifacts)? If I have it wrong please explain in visual terms.

For me it took listening to a magnet mounted driver with no baffle at all to start hearing edge diffraction from baffles. I'm sure many of the technical will chuckle, but now I can often hear the outline of the baffle from diffraction occuring at the edges.

I hope you're onto something with the idea of a gradual pressure release to eliminate diffraction. The problem I keep running into is, if I address diffraction on the front side, diffraction on the back becomes more difficult to cure. I hoped my dipole waveguides would solve the dilema, but based on vibrations in my roundovers, I doubt it. That means horns, especially round ones, must have tremendous amounts of edge diffraction. Maybe I'll have to go with a front and rear driver dipole as the only solution.

I visualize the expanding wavefront meeting a characteristic impedance, the same as radio waves seeing the characteristic impedance of free space. This impedance stays the same as long as the wavefront experiences no change in the expansion characteristics, but changes abruptly when the wavefront (at the speed of sound) encounters the edge of a boundary surface.

The edge behaves similarly to a kink in an RF cable or a fracture in an optical surface - some of the energy continues to move in the same direction, some diffracts in all directions with associated frequency-dependent prismatic effects, and some even returns in the direction it came from.

All it takes is a discontinuity in the wave expansion surface - this is very well known in the RF field, where getting rid of reflections in cables is troublesome. (When I was at Tek they made a gizmo called a Time Domain Reflectometer, or TDS, or more simply, cable radar. Even hard-to-see kinks were plenty visible on the TDS display.)

In optics, light moving across a hard edge causes loss of resolution in a stopped-down camera lens, and also causes those interesting cross pattern in astrophotographs. (The diffraction is caused by the vertical and horizontal supports of the 2nd mirror in the telescope.) The star-like artifacts can be greatly reduced by using a variable-density filter to "soften" the hard edges of the lens aperture or the mirror supports.

Since the edges of the loudspeaker enclosures are much smaller than the wavelengths passing across them, the wavefront diffracts in all directions, and there are frequency-dependent prismatic effects as well. In addition, the reverse-reflected energy is free to travel all the way across the cabinet face and encounter the opposite edge. Since there are almost no losses moving across the face of the cabinet, this succession of reflections sets up a standing wave, very similar to the standing waves set up inside conventional box cabinets.

(Note: an acoustic standing wave is NOT the same as the walls of the cabinet flexing. You can have quite strong standing waves with a rigid cabinet - think of the long reverberation times of a concrete-block bathroom, for example.)

Yes, JohninCR, I can hear cabinet diffraction - with pink noise (the ideal stimulus) it is strongest at a 135-degree angle with respect to the front surface of a conventional box speaker. It actually sounds like a little bitty tweeter emitting right at the cabinet edge.

I don't see any reason why smoothing out the termination of the front wave should degrade the rear wave. If the treatment is symmetric (front and rear), both should improve at the same time. The null region should improve most of all, no small gain, considering the importance of having the early room reflections have the same spectra as the direct-arrival wave.

That's my big beef with horns - the far off-axis region can get pretty ugly, with severe time distortions thanks to strong horn-edge diffraction effects. That's where all those narrow "spikes" in the 5 to 10 kHz polar pattern come from - the frequency and time response within those spikes is going to be pretty bad.

Is this audible? Oh yes. The ear uses fine-grained information in the 5 to 10 kHz region for localization and depth perception - that's what stimulates the outer ear (the pinna), which is used for precision localization. And guess what, depth information is an area where horns are not that good compared to the best direct-radiators.

Hard reflecting surfaces and edges have destructive effects on image quality and timbre - especially when the dimensions involved are similar to head, shoulder, and pinna dimensions, which are essential for localization of sound. Vocal timbres in particular are susceptible to artificial-sounding colorations, partly because human beings have such acute discrimination between voices, and also because vocal-tract dimensions can be similar to dimensions of loudspeaker diffraction effects (a few inches).

I can't emphasize enough the first millisecond (14 inches long) is the most crucial thing the loudspeaker does - although the first 3 milliseconds are pretty important too. Note that I'm discussing reflections that are in the 0 to 42 inches long, kind of awkward considering our speakers are the same size! Thus the importance of (any) smoothing techniques to soften the acoustic edges so they re-radiate less energy - and have smoother polar patterns.

What makes this more difficult is the dynamic range of the ear. A 20 dB reduction of edge reflection energy might seem a huge triumph, until we consider we really need 60 to 100 dB of reduction to say it's actually gone for good. This where we hope that auditory masking will save the day, so the quieter reflections fall below the auditory threshold.

[BudP]

Hi Paul,

I am sorry to report ignorance here. I do not know what CSD stands for. There are some time domain plots in my white paper posted at Positive Feed Back on line. Look in the last few pages. I did not perform these tests, Larry Arnst in Portland ran them from his Franklin (Mac copy) back in 1986. I do still have those original plots.

http://www.positive-feedback.com/Iss...ndingwaves.htm

I do not have the financial wherewithal to afford a MLSSA system, but the testing I did with a DOS based Lincoln Audio suit showed very little difference in before and after harmonic distortion plots, out to 9th order This is in line with listening tests, as the "character" does not change but a lot of extra information disappears.

I realize that sounds worrisome but the information that disappears is not related to the sources. It is just noise and since it is no longer in support of an information packet, the correlator finds it very interesting until it proves non threatening and then dismisses it and listens through it. It's removal is somewhat startling, and then you discover how deeply you can actually hear into the reproduced event. Not that you cannot already do that, but it can be tiring and post treatment it is effortless.

Bud

[planet10]

Originally posted by BudP
I do not know what CSD stands for. There are some time domain plots in my white paper posted at Positive Feed Back on line. Look in the last few pages. I did not perform these tests, Larry Arnst in Portland ran them from his Franklin (Mac copy) back in 1986.

CSD = cumulative spectral display = waterfall plot. Even these may not reveal all. The kinds of things i envisage enABLE doing (& similarily many of the tricks i use) are affecting the drivers performance 30-40 dB (or more) down from the main level of the signal. Rooting them out with currently en-vogue measuring kit can be tricky.

dave

PS a Franklin was an Apple II clone and had nothing to do with the Macintosh.

[BudP]

Hi Dave,

Thanks for the information.

The plots in the white paper come from that sort of analysis. I did not keep the original waterfall but the three time sequences are from that test. The final plot has always been the important one to me as it clearly shows the ringing going on in the untreated tweeter and just how much of that ringing has disappeared in the treated tweeter. Some of the vacant areas were already 60 dB down for the treated tweeter.

I would not have known the difference between an Apple 2, a Mac or an IBM / Microsoft computer at the time. Only barely do now.

Bud

[BudP]

Sun Ra

Were I building your dream I would put Mamboni's treatment on the back of all drivers bigger than 8" in diameter, the EnABL on the front, to clean up the 30% of ringing that will likely remain and do the same thing for the baffles.

On the Baffles the Mamboni will likely slow the back wave propagation just a tiny amount and allow the front wave to meet with it, at a greater than 180 degree reference angle for the null zone. I think that this might be beneficial in sound stage portrayal.

What do you guy's with actual experience with these things think?

Bud

[Lynn Olson]

Originally posted by johninCR
Bud,

I'm having trouble understanding exactly what the Mamboni and EnABL treatments are. I have some extra drivers I'm willing to risk. Can you direct me to specifics?

Start the OHM Acoustics "Walsh F" thread here and follow the comments by Mamboni, BudP, and c2cthomas. Around Page 13 or so I have some suggestions of my own about the Mamboni felt triangles.

Sound quality is the result of physics, acoustics, ingenuity, and the skill of implementation. This is why a Mamboni/EnABL-modified driver could quite easily outperform a $2000 boutique driver - it's not the name, it's how it's built that matters.