The first curve opened in the graphs is the yellow curve with the comment at the very bottom, I usually make that one the final or reference depending on what I'm showing. In that one, it's the response from the final pattern added that is shown in the picture at my earlier post here.
The blue curve is the original response from the driver before doing anything to the driver. That final yellow curve is with a typical thickness of damping on the surround (see it in the picture) and the pattern of damping "dots". It eliminates the surround mode resonance, smooths the breakup and smooths the overall response. A crossover would be far simpler for one. Some of the problems could not be eliminated with an analog crossover as has been done mechanically, such as that surround resonance.
BTW, I never smooth any of the measurements that I post at any time. You'll see that at my site as well. The only smoothing is the effective smoothing that is inherent in a window used on the impulse response for the FFT to get the frequency response. I'm usually trying to find and expose all issues with a driver.
One more thing. I did all of the testing of various patterns over a couple of days IIRC and never moved the mic nor the driver (in its original box). Upper frequency patterns can be affected by mic placement, especially something such as a full-range driver that can become more erratic in polar response at the top end. Attention to detail when testing is important. IMO, move the mic and the comparisons may be suspect in certain cases, but this is seldom discussed.
Dave
Last edited:
So, to review.
In each graph the graph of interest is the yellow one, that shows the effect of the treatment. The blue (in the case of two traces) is the as measured "stock" response?
But I don't see the blue curve being identical each time?
Fwiw, all your measurements have some smoothing applied... you'd have to go somewhere into the menu to really get all the smoothing off, IF they permit it. That never yields anything that most people would want to see - in fact it is very disturbing to see the extreme +/- variations so close together. At least that is what I saw when I got all smoothing truly disabled... that was some time ago. But for our purposes we can disregard my last comment.
_-_-bear
In each graph the graph of interest is the yellow one, that shows the effect of the treatment. The blue (in the case of two traces) is the as measured "stock" response?
But I don't see the blue curve being identical each time?
Fwiw, all your measurements have some smoothing applied... you'd have to go somewhere into the menu to really get all the smoothing off, IF they permit it. That never yields anything that most people would want to see - in fact it is very disturbing to see the extreme +/- variations so close together. At least that is what I saw when I got all smoothing truly disabled... that was some time ago. But for our purposes we can disregard my last comment.
_-_-bear
No (not all cases, yes for this one) and yes for the blue, you have to read the descriptions of the measurements. Go to the full web page, I think the descriptions are fairly complete. Most curves show treatment of one form or another. It all depends on what I'm trying to present for each case. In some cases I compare against untreated, other times only treated results.
No. See above.
Don't make assumptions. You are apparently not familiar with the measurement software I use. Praxis is used for display only, I use LAUD for measurement, but Praxis has full control over smoothing as well. I almost never use smoothing in the frequency domain.
As my father is so fond of saying, never say never. What you see in all of my measurements is without any frequency domain smoothing, period. I prefer to see all of the warts. If you think that you see smoothing, then the results are better than you have assumed. To repeat, the only smoothing is effectively that from the window applied for limited time domain data in the impulse response.
Dave
Last edited:
Dave,
There is 'smoothing' if you see a curve at all... otherwise you would see a great many up and down spikes, the center point of which is the "curve"... fyi. But let's not argue that. I could be wrong, but I think I am correct on this. The software is doing this for you, no matter what you dial up, apparently. But it's fine...
I did look at your website.
It is totally confusing to me as to what each color curve represents, since the yellow one is not "tagged".
Maybe it will make more sense if I look at it again... dunno.
_-_-bear
There is 'smoothing' if you see a curve at all... otherwise you would see a great many up and down spikes, the center point of which is the "curve"... fyi. But let's not argue that. I could be wrong, but I think I am correct on this. The software is doing this for you, no matter what you dial up, apparently. But it's fine...
I did look at your website.
It is totally confusing to me as to what each color curve represents, since the yellow one is not "tagged".
Maybe it will make more sense if I look at it again... dunno.
_-_-bear
no BL or BS
Randy...
feel free to go generate this data you desire... no one's suggesting otherwise...
I have more than a "seat of the pants" insight into boundary layer analysis. At Bell Aerospace, we studied bl's of wind flow over airfoils and fairings, metal convection and diffusion through electrolytes onto hardware surfaces, reactant concentration profiles around large arrays of axisymmetric laser nozzles operating at 1000 F in a deuterium/fluorine gas laser, ultrasonic and sonic scanning of metal surfaces and interiors for metallurgical defects to avoid catastrophic failure, etc.
Producing ink jet nozzles and print heads at Mead Digital and Burlington Industries (some up to 5' long), we were working on ultrasonic stimulation to produce uniform droplets for continuous spray applications, as well as synchronous electronic charging of said droplets for printing or not, under computer control. I also developed advanced metals and photolithographic methods for building corrosion resistant orifice plates with up to 5000 .003" holes each accurate to +/- .000015" themselves, all of which required intimate knowledge of boundary layers and whether or not they had any effect on success or failure of the hardware.
So my coments are directed at the acronym itself, eNabl, including boundary layer as a name and implied mechanism, since in no way can a boundary layer (fictitious or otherwise) contribute to the alledged effects yet to be shown as anything other than a mental construct (I'm being very polite in this characterization, btw) The wavelengths are simply too large and the lack of rigor in the explanations to0 blatantly obvious. These insights and observations, over several decades, leave me no choice but to interpret this hypothesis as null until someone proves otherwise (hasn't happened). I've heard several enabled systems, and they were anything but remarkable due to the enabling process.
John L.
Randy...
feel free to go generate this data you desire... no one's suggesting otherwise...
I have more than a "seat of the pants" insight into boundary layer analysis. At Bell Aerospace, we studied bl's of wind flow over airfoils and fairings, metal convection and diffusion through electrolytes onto hardware surfaces, reactant concentration profiles around large arrays of axisymmetric laser nozzles operating at 1000 F in a deuterium/fluorine gas laser, ultrasonic and sonic scanning of metal surfaces and interiors for metallurgical defects to avoid catastrophic failure, etc.
Producing ink jet nozzles and print heads at Mead Digital and Burlington Industries (some up to 5' long), we were working on ultrasonic stimulation to produce uniform droplets for continuous spray applications, as well as synchronous electronic charging of said droplets for printing or not, under computer control. I also developed advanced metals and photolithographic methods for building corrosion resistant orifice plates with up to 5000 .003" holes each accurate to +/- .000015" themselves, all of which required intimate knowledge of boundary layers and whether or not they had any effect on success or failure of the hardware.
So my coments are directed at the acronym itself, eNabl, including boundary layer as a name and implied mechanism, since in no way can a boundary layer (fictitious or otherwise) contribute to the alledged effects yet to be shown as anything other than a mental construct (I'm being very polite in this characterization, btw) The wavelengths are simply too large and the lack of rigor in the explanations to0 blatantly obvious. These insights and observations, over several decades, leave me no choice but to interpret this hypothesis as null until someone proves otherwise (hasn't happened). I've heard several enabled systems, and they were anything but remarkable due to the enabling process.
John L.
Last edited:
The yellow curves are tagged. As I said, the description associated with the curve is at the bottom of the frame, below the additional curves. This is just the way Praxis does this.
As for smoothing, any FFT-based system is going to have discrete points generated from the impulse response and the number of FFT points is user-selectable in general. I use the upper limit in LAUD of a 16,384 point FFT. With most drivers you're not likely to see wild gyrations in the graphing created from this unless you select a window stop time marker that does not adequately eliminate reflections.
You're going to see serious issues and it's not going to look smoothed until you are in the lower frequencies where the limitations due to the window occur. As an example, here's a tweeter before and after I modified it to eliminate an internal reflection, no smoothing. You can download the measurement files from my site if you care to examine the frequency interval in that 16384 point FFT. All FFT-based systems create a graph from the discrete data, so if you want to call that smoothing, fine, there's a limit to the resolution, especially given the small windows generally required. However, I think that you'll find that all of the sample points will fall onto the graph. The point spacing is about 3Hz. It's simply curve fitting to the data. The lower range will look smoothed due to the limited data and window tapering of the impulse response for the FFT.
Both of these curves were generated with essentially the same settings and mic positioning. It's likely that I did it without even moving the microphone, as I prefer to do them that way. I measure, make a mod, then re-measure immediately while everything is set up, no changes usually occur in between. As an example, the big dip you see starting around 15,600 to the return around 17,260 is 1660 Hz with approximately 550 data points. It looks like mostly straight lines, but it's still curve fitting. Now if you actually apply even 1/12 octave smoothing, you are going to see curves rather than the somewhat sharp changes. Then there's the log nature of the frequency axis added in.
Whoever may supply measurements, if done with an FFT-based system is going to have the same issues. You'll find quite often that others will take that same FFT data and apply 1/10, 1/6 or even 1/3 octave smoothing. Those are smoothed in the formal sense. I apply none. The smoothed data is still usually curve-fitted. It sounds to me as though you're referring to curve fitting the data. That's ubiquitous, not unique to anything I do. The software does allow me to specify no smoothing beyond what's inherent in an FFT-based system as a minimum.
I would suggest that when you see a significant amount of "up and down" points at the low end and midrange, that may be indicative of limitations in that measurement system and graphing as well, there's straight line drawing occurring rather than curve fitting. There must be data lacking in between those "up and down" points as well, so even that method of presentation could be considered to be misrepresenting the data in the opposite direction, indicating sharper gyrations than may actually be present. It assumes that the line between points should be a straight line.
Both ways have limitations. It's rather academic here, I think. We're veering pretty far OT.
Dave
sure
if you mean induction of expectation bias, maybe... depends on who's listening and what they believe in
if you mean induction of expectation bias, maybe... depends on who's listening and what they believe in
So John, would you please observe this short utube video and discuss the lack of a boundary layer shown? Not trying to start a fight, but, rather trying to understand the limits of applied terminology. Do please turn on the sound, I am sure you will find the discourse amusing.
YouTube - resonance
Bud
YouTube - resonance
Bud
It appears... How can any speaker cone avoid distortion?
Do some of the newer woofers that that have been made of a tactile woven composite have an effect on this phenomena?
Proof by irrelevance????
(I realize this was not addressed to me.)
It revealed how the woofer cone distorts and moves diffrerently when subject to certain frequencies. No? Like a variable mirror found in the Fun House.
knows no bounds
Hi Bud
yeah, I've watched that every time you post it, but not sure what boundary layer dynamics has to do with it?? All I see (with little to no information as to waveforms, etc.) is a visualization of standing waves, interference patterns due to variable resonances at differing frequencies, cone anomalies,etc. and some sort of description about the wonders of destructive and constructive interference.
What is the actual physical mechanism you propose, and have you observed it, for your "boundary layer" explanation (and I've read the part about how the patterns modify how the waves "lift" off the surface... that's a non starter)? Typically, boundary layer interpretations involve mass trasfer and rates of change, none of which occur with sound waves, other than periodic rarefaction and compression orthogonal to propagation. Boundary layer effects only influence conditions in the boundary layer; I fail to see how they could (if thay somehow exist) propagate to, for instance, a listening position located meters away, since they are defined by the boundary conditions set for the equation of state.
I'm not trying to fight either, just try to comprehend that much of what you propose as proof comes across as either disrespect for scientific and engineering protocols or misunderstanding of the profound effects of experimental bias in observations. BTW: nothing wrong with expectation (placebo effect) as a valid explanation for successful implementation and marketing of ideas. happens all the time. let's just call it what it is, until irrefutably proven otherwise
John L.
Hi Bud
yeah, I've watched that every time you post it, but not sure what boundary layer dynamics has to do with it?? All I see (with little to no information as to waveforms, etc.) is a visualization of standing waves, interference patterns due to variable resonances at differing frequencies, cone anomalies,etc. and some sort of description about the wonders of destructive and constructive interference.
What is the actual physical mechanism you propose, and have you observed it, for your "boundary layer" explanation (and I've read the part about how the patterns modify how the waves "lift" off the surface... that's a non starter)? Typically, boundary layer interpretations involve mass trasfer and rates of change, none of which occur with sound waves, other than periodic rarefaction and compression orthogonal to propagation. Boundary layer effects only influence conditions in the boundary layer; I fail to see how they could (if thay somehow exist) propagate to, for instance, a listening position located meters away, since they are defined by the boundary conditions set for the equation of state.
I'm not trying to fight either, just try to comprehend that much of what you propose as proof comes across as either disrespect for scientific and engineering protocols or misunderstanding of the profound effects of experimental bias in observations. BTW: nothing wrong with expectation (placebo effect) as a valid explanation for successful implementation and marketing of ideas. happens all the time. let's just call it what it is, until irrefutably proven otherwise
John L.
Last edited:
John, I have no dog in the fight regarding the terminology... that in and of it self, unless it has some significant aspect to the practical application, is meaningless to me. Your abilities and credentials are not in question from where I sit. If you want to argue the nomenclature or technical lack of boundary effect with Bud, that's fine with me, not my issue or concern.
Dave, I am only trying to understand what you are showing in your graphs... when I get a break I will look again and try to see what is what.
In general I think that what would help the understanding of the placement of "stuff" on the surface of the cone would be laser interferometry. That way one could actually start to see break up modes and the effect (if any) of the "stuff" on the surface...
My two wooden nickels...
_-_-bear
Dave, I am only trying to understand what you are showing in your graphs... when I get a break I will look again and try to see what is what.
In general I think that what would help the understanding of the placement of "stuff" on the surface of the cone would be laser interferometry. That way one could actually start to see break up modes and the effect (if any) of the "stuff" on the surface...
My two wooden nickels...
_-_-bear
John L and John K
With respect to the description of the movement of particles radially, out from the center, riding up at some form of interference and returning to the center of the activity shown. Does this movement across, rather than directly out ward from the involved surface show a boundary layer effect? Do we assume that the perimeter where the particles arise is due to an interaction between transverse waves and direct movement imparted activity? What is the mechanism driving the small droplets out beyond the main activity, which show the same form of movement of particles? Can we draw any analogies between these two areas of activity and the minimum phase portion and "break up" portion of a fixed geometry cone? If the large volume's size is an interference between transverse and compression wave activity would interruption of the transverse wave at that boundary eliminate that particular structure and cause the interference activity to arise further out, in a larger radial direction or would the loop back mechanism shown just disperse into local zones, as shown out beyond the interference point?
These are the question that the utube movie brings to my mind.
Another set of questions is the logical parallels between this 30 CM flat vibrating surface and a conical shaped, surface of the same radial distance from center to edge. Will the conical shape have fixed interference loop back structures that arise? Will they be frequency dependent or will a much wider range of frequencies give rise to an interference in the same radial place. What would be the effect of adding curvature to the cone for 1.3 of the radial distance?
A further set of questions pertain to what the analogue to this loop back mechanism would be for air molecules? Would the loops be larger, extending out into the room or would they be confined more closely to the vibrating surface? Would the conical shape limit this activity?
Just to add a comment, not a proof of anything. When I correct a driver I actually search for the loop back points on the cones. For flat cones there are basically an infinite number of locations, a perfect place to utilize SY's fractal pattern application of mass and local stiffness. Curved cones only show a need for loop back correction from the outer periphery of the curved portion to the surround, a relatively flat radial traverse. I assume that the curved portion is acting in a minimum phase fashion and that the "break up" portion is outside of the curve. Application to the curved portion changes nothing. Application of the patterns out beyond this zone is where the audible changes show up.
Not expecting answers here, just asking my questions.
And by the way Dave, very interesting work on that flat cone.
Bud
Various attempts over the years have been made to create a very very stiff and lightweight cone, one that will not be able to have breakup modes (at least in-band). These attempts have had varying degrees of success. Sometimes they seem to have exacerbated other factors in the the way a driver of the cone type operates.
Imo, every material that is used to form a cone will behave differently. There are going to be nodes and modes, the question is (imo) what is the "Q" of these going to be, and the amplitude, and phase relationship...
Thus we see a wide, wide, range of cone materials and methods of manufacture. Compare those mfrs whose drivers use the same motor and different cones - note the substantial difference in measured response?
Another thing that seems to stick in my mind is the sound propagation across or through the cone itself. In the case of most paper type cones (or woven materials) the cone is rather lossy, but in the case of say an aluminum cone it is rather not lossy at HF... so the termination at the perimeter starts to look even more important. Then there are the "diamond" and beryllium domes for tweeters, attempts at extreme stiffness and lightness.
Dlr shows a positive result from merely making the surround more lossy via a paint on "damping" compound.
So, from where I sit the "trick" to getting a good sounding dynamic driver is a fine balancing act between losses, absorptions vs. transmissions and radations... I think once you put a cone on the motor, the linearity of the system immediately is lost and it becomes a case of minimizing problems and maximizing the desired parts.
Seems to me that Bud's approach is intended to effect the nodes and breakup modes on the surface of the cone - never mind the "boundary layer" claim - and may have some merit IF one could in essence "fractalize" or "randomize" these modes on the surface (spreading them?) just the right amount. Finding the right amount and the right places is where it becomes not quite simple, and imo impossible by merely guessing or using a uniform approach and putting a given treatment in the same place all the time...
I say this with the understanding that what a cone driver is actually doing is in reality a fine mess to start with. Smoothing out the mess may not be a bad idea at all... (assuming you see it as I describe).
I think how to do that is the real issue and where the best solution(s) may be found??
_-_-bear
I may have missed something in this lengthy discussion and Bud may have already covered this factor.
Might be that each individual speaker needs first to be measured and categorized, say like a string on a guitar. Drum heads are effected in tone by differences in thickness and materials used, and even its surround. To apply the same damping technique in effecting the lowest string on a bass cello, and the lowest string on a guitar would render different results.
It may just be that what EnABL is doing is in a pioneer stage of development (which it is), but a valid one just the same. Once a formula is perfected for determining different drivers of the same diameter there can be rendered consistent results. The diameter of the speaker is one thing. Thickness and overall mass of the driver are variables that right now can not be determined with exactness. EnABL may need an unique formula for every different driver.
IMHO.. Gene
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.