Yes, it's important to note the distinction between lossy termination and lossy directivity control- though, of note, lossy termination is expected to lead to some extension downward of pattern control in midtweet horns as we're discussing. Naturally, bass and midbass horns are hardly, if at all, going to notice foam/felt of the sizes we're discussing here.
From a practical results perspective, is the difference one largely of lost efficiency ?
If we take a wide dispersion driver diaphragm and place it at the base of an inverted pyramid with extremely lossy walls, essentially what the design in the original post shows, we are constraining the directivity simply by absorbing anything beyond the angle that the lossy walls are occluding based on line of sight - energy is lost at the same time as limiting the radiation angle, so on axis response stays the same and net power response drops.
On the other hand an actual wave-guide is taking a plannar wavefront from the driver and via the taper in the throat "expanding" it into a spherical wavefront of a particular directivity, and doing so without absorbing any energy. In this case the net power response is about the same (assuming a wave-guide that provides little impedance transformation benefit) but the on-axis response is now increased due to focusing the energy over a smaller solid angle.
A good example being a CD waveguide where the on axis response increases a lot, particularly at the low end of its range, but the power response is not changed much. (Relative to normalised on axis response power response falls of course)
Controlling directivity by resistively absorbing unwanted radiation seems a bit wasteful to me, and likely to cause the driver to have to work a lot harder when the same directivity control could be achieved more effectively with an appropriate waveguide design with much greater efficiency and maximum SPL for the same diaphragm size ?
Last edited:
Simon
I agree with you that there is not a lot of attractiveness to doing things this way. Waveguides work very well for this task. There is no need for another solution. The biggest problem with waveguides - they need to be big to work optimally - is not alleviated with an entirely foam device, in fact its made even worse. And then the efficiency is lower. Not a lot there to like IMO.
I agree with you that there is not a lot of attractiveness to doing things this way. Waveguides work very well for this task. There is no need for another solution. The biggest problem with waveguides - they need to be big to work optimally - is not alleviated with an entirely foam device, in fact its made even worse. And then the efficiency is lower. Not a lot there to like IMO.
What I don't understand relates to response and efficiency when very significant damping is applied to a baffle.
Forget the notion of a foam waveguide, as that was not the intent, at least of the Stereophile reviewed unit. Assume the desire is to deal with energy moving down the surface that could create diffraction and reflection problems. If we have a flat baffle and place a fair amount of absorption on the surface, I would think that energy flowing along the surface would be absorbed and the wavefront would tend to curve into the absorptive surface. This is a bit like diffraction, energy bending around the corner, except the energy is lost rather than filling in the shadow region behind the baffle.
How does this impact the axial response? The frontal response is changed if we move a unit from 2pi or from a large baffle to a small baffle. Wouldn't a similar shift happen if we heavily damp a baffle?
Has anybody measured this before?
David
Forget the notion of a foam waveguide, as that was not the intent, at least of the Stereophile reviewed unit. Assume the desire is to deal with energy moving down the surface that could create diffraction and reflection problems. If we have a flat baffle and place a fair amount of absorption on the surface, I would think that energy flowing along the surface would be absorbed and the wavefront would tend to curve into the absorptive surface. This is a bit like diffraction, energy bending around the corner, except the energy is lost rather than filling in the shadow region behind the baffle.
How does this impact the axial response? The frontal response is changed if we move a unit from 2pi or from a large baffle to a small baffle. Wouldn't a similar shift happen if we heavily damp a baffle?
Has anybody measured this before?
David
2 things:
Ease of construction
Design flexibility
there are some options available to you with lossy barriers that don't exist with horns/waveguides in many instances. One could take a closely spaced cone/dome design with it's typical crossover dispersion flare and retrofit a shaping absorber to the top and sides and reduce the dispersion flare at/above XO.
Interesting questions, the following is just some postulation:
At high frequencies where the diffraction off sharp baffle edges is significant the absorption would probably be quite effective, so that little if any of the wave would reach the edge, therefore on axis diffraction ripples would be largely eliminated. I'm sure plenty of people have measured the effect of foam damping on a baffle at high frequencies.
However I can't see any way that the damping would change the fundamental full space to half space transition occurring at the baffle step frequency for two reasons - one is that absorption requires thickness, and any realistic thickness of foam damping on the surface of the panel would become ineffective below a certain frequency.
The other is that even if the foam was 100% effective absorbing low frequencies, only the part of the wave that impinged on the foam would be absorbed - the part of the low frequency wave immediately in front of the foam would still travel around behind the baffle at a distance corresponding to the baffle step frequency.
The lower frequencies would still wrap around the padded baffle at the edge of the baffle+foam because the wavelength is long compared to the size of the corner.
Not really thick foam, but I have tried a layer of 10mm thick open cell foam on the front of a wide flat baffle, and to be honest it didn't do that much - some reduction in baffle diffraction ripple effects above about 1.5Khz, but no real change below that. Mind you it was an 8" driver so driver directivity will limit how much of a wave is travelling along the baffle in the first place...in this example at low frequencies the foam is probably ineffective, while at high frequencies the driver is too directional to illuminate the baffle, so I only see a small change in the transition region between the two extremes.
It seems likely that realistic thicknesses of foam are only useful for tweeters or very small midrange drivers.
Last edited:
I don’t see it as a replacement for a waveguide, rather a better way of implementing the classic “2-way midwoofer + tweeter mounted on a baffle” design.
With respect to efficiency, I see it as absorbing excess energy that we may not want anyway.
Its far from a revolution, rather an inexpensive and flexible way to fine tune room response.
With respect to efficiency, I see it as absorbing excess energy that we may not want anyway.
Its far from a revolution, rather an inexpensive and flexible way to fine tune room response.
The fact that only energy traveling along the baffle is absorbed wouldn't prevent a change to axial response. Indeed this is the case with diffraction around a finite baffle: energy traveling along the baffle surface suddenly loses its boundary contraint and starts to expand backwards. The farfield axial response "sees" that the energy is lost to the front hemisphere and baffle step response is the result. Lost energy at the sides of the wavefront must be supplanted by frontal energy diminishing axial response.
If energy is lost into the absorption on the baffle then the same thing must be occuring. The question is to what degree. How effective an absorber the material is at the lower frequencies (as you said) is the only issue separating the damped infinite baffle case from the finite baffle with diffraction case.
Regards,
David
One would think. Wide angle radiation (of less than 90 degrees) would be then drawn toward the baffle as it seeks to integrate with the lower pressure caused by this 'lost energy' in the damping material. What I'm more concerned about is whether this new diffraction creates a backwave or whether it is somehow immune or is dissipated due to the damping material.
Just thought that I would make a comment regarding Norman's suggestion. I have a pair of Klipsch Heresy 3 running to which I had already added some high density fiberglass to the inside edge (top & bottom) of the K701 mid horn and after reading Norman's comments added a 5/8" square vertical column of fiberglass at each end of the K701 mid horn. As Norman found this resulted in a bigger difference to the sound than did the two strips of horizontally applied fiberglass. I am sure that a better horn would be nice but if you are working with what you have this simple mod works wonders. I plan to experiment with some 30 PPI open cell foam in the throats of the K701 as they are long and I would expect to hear some difference there.
An externally hosted image should be here but it was not working when we last tested it.
On the big horn, covering the sides made double the difference to covering the top bottom. Ideally it would be a smooth (no breaks) in the horn path, and no corners. That's why I've been looking at the jbl newer horn, the 2384 used in the 3622 and the 4622 (double 15" plus horn, like my setup). And I've also been looking at the stereolabs 400hz horns, but I don't think I could live with the lack of dispersion past 3-4khz (tractrix horn). I like my 90 degree dispersion setup.
It is well worth the time and ugliness to add the foam. It makes everything "sharper" as my wife says, less reverby, revealing detail that was obscurred before. A friend suggested trying small weatherstripping in all the corners, I'd imagine it would help, but only add a little.
Norman
Thank you Norman for posting your sub five minute fix. Say what kind of rubber bands did you use on those Klipsch anyway? Have a great day. Best regards Moray James.
This does come across as a worthy goal. Although it would be good to view a study, for the simpler approach I guess there are some obvious clues to the desired absorption thickness if you determine not bother with frequencies greater than the baffle, and considering the high angles of incidence involved it may not take much. I suspect bass trap density material is the value point here and I use same.
Allen, could you explain this further with respect to the different frequencies and radiation angles? I am having some trouble visualising how this all plays out.
With regard to horns I can't say much. I am inclined to think that using felt (or foam) on walls would be a negative. I've played around with small tweeters in small waveguides with the results being that the most important area is the throat, matching the dome dimensions to the diameter of the throat and the depth. Small amounts of cotton damping placed in the throat just in front of the dome was the most beneficial. I would not try anything on the walls.
Regarding the Lipinski product, it's nothing new. This is almost exactly the layout and felt placement used by John Dunlavy a long time ago. It allowed him to create first order systems with aligned acoustic centers that gave excellent measured on-axis response (at 10' IIRC). I've done similarly with ordinary, small stepped baffles. The Linpinski info is loaded with hype. It works, sure, but it's not special by any means. I also don't think of it in terms of horn loading, but it does alter the polar response and if thick enough it does tend to provide a bit of change directivity. In fact, they use a ring radiator, making reference to extra wide dispersion. This is contrary to all measurements I've made of the XT25 and XT19. A ring radiator, due to lack of occlusion of the far side of the ring, has increasing interference patters in the extreme off-axis. A typical dome of the same diameter has wider dispersion due to the occlusion by the dome off-axis.
What I do know is that placing felt on the baffle side, flush to the front baffle edge, provides practically zero benefit, still somewhat surprising to me. This is one reason I question the benefit of placing thin strips at the mouth of a horn. It's too thin to have a significant impact on the frequencies involved at that point or so I would think. Conjecture here, though.
There can even be a small change in the 4-pi response, but as the frequency drops, the less the effectiveness. It's possible to reduce the on-axis diffraction peak by as much as 2db with a smoother polar response in the transition area. Below a certain point, though, the response converges with the output when no treatment is used.
Dave
One thing I have noted consistently. In the off-axis, beyond maybe 45 degrees, there will be a range where the response will be a bit more ragged than the untreated case. Tweeters often are smoother off-axis in many cases. There's a transition region where there actually seems to be some diffraction from the felt itself. Given the axes involved, the greatly improved response at lower angles are worth it. I even measured this with a DXT.
Regarding the Lipinski product, it's nothing new. This is almost exactly the layout and felt placement used by John Dunlavy a long time ago. It allowed him to create first order systems with aligned acoustic centers that gave excellent measured on-axis response (at 10' IIRC). I've done similarly with ordinary, small stepped baffles. The Linpinski info is loaded with hype. It works, sure, but it's not special by any means. I also don't think of it in terms of horn loading, but it does alter the polar response and if thick enough it does tend to provide a bit of change directivity. In fact, they use a ring radiator, making reference to extra wide dispersion. This is contrary to all measurements I've made of the XT25 and XT19. A ring radiator, due to lack of occlusion of the far side of the ring, has increasing interference patters in the extreme off-axis. A typical dome of the same diameter has wider dispersion due to the occlusion by the dome off-axis.
I've never actually seen any description of precisely how felt improves the response. We used to debate that years ago back on the old Madisound board. My conjecture is that it's a combination when used on a flat baffle. It absorbs sound passing through, but it also may tend to "disrupt" the flow close to the surface making it a bit turbulent so that it is not a smooth flow when it reaches the edge.
What I do know is that placing felt on the baffle side, flush to the front baffle edge, provides practically zero benefit, still somewhat surprising to me. This is one reason I question the benefit of placing thin strips at the mouth of a horn. It's too thin to have a significant impact on the frequencies involved at that point or so I would think. Conjecture here, though.
I have made countless measurements over the years. I've measured the response of adding felt around a tweeter mounted on my 2m x 2m quasi-anechoic baffle. Since there is no diffraction occurring, one would assume that the response would show little change. However, there was always some deterioration in the response. This is, undoubtedly, due to the fact that felt is not perfectly absorptive. It is also reflective to some degree. That's where selection of density, shape, depth and thickness all have an impact. It's partially a balancing act.
There can even be a small change in the 4-pi response, but as the frequency drops, the less the effectiveness. It's possible to reduce the on-axis diffraction peak by as much as 2db with a smoother polar response in the transition area. Below a certain point, though, the response converges with the output when no treatment is used.
It can have a surprising affect on small midrange drivers as well.
True to a point. It depends on the directivity of the driver, the thickness and the baffle dimensions. Small drivers (including midranges) on narrow baffles will show more change than larger drivers or wider baffles. I suspect that the wider the baffle, the thicker the felt must be because you are trying to affect lower frequencies, a result of the wider baffle.
True, but what appears to occur is that the wave the passes through directly combines with the wave passing over the felt in such a way as to reduce the destructive interference somehow. There always seems to be an optimal thickness. Too little and it's useless. Too much and you get reflections and/or horn loading. DDF (a member at Madisound and now Parts Express) has frequently speculated that one could make a waveguide of sorts from felt. I agree, but I doubt it's ultimate usefulness. It would take quite a lot, certainly much more than shown in the photo at the link in the OP.
The benefit may be that there is a transition in the effectiveness related to wavelength, so there will be a transition range of improvement. Measurements tend to support this.
Short, but accurate I'd say.
Dave
One thing I have noted consistently. In the off-axis, beyond maybe 45 degrees, there will be a range where the response will be a bit more ragged than the untreated case. Tweeters often are smoother off-axis in many cases. There's a transition region where there actually seems to be some diffraction from the felt itself. Given the axes involved, the greatly improved response at lower angles are worth it. I even measured this with a DXT.
Last edited:
Interesting link, Dabbler. If I read his description right he is actually getting worse results with the circular foam ring. In otherwords it is fairly reflective and circular so it contributes the worst response variation. The felt blocks were better.
DLR, my experience is pretty close to yours. I've put lots of felt or foam on grille frames to kill reflections and usually got good results. I haven't played much with large expanses on the baffle, primarily because it was bad for cosmetics and I was actually trying to sell the speakers. (sad admission)
What I'm wondering about is whether there is an effect similar to diffraction whereby a broadly treated baffle causes axial response to fall, assuming that energy moving parallel to the baffle takes a right turn into the absorption. I realize that the change of edge reflections will create a complicating factor.
I need to set up my own experiment.
David
DLR, my experience is pretty close to yours. I've put lots of felt or foam on grille frames to kill reflections and usually got good results. I haven't played much with large expanses on the baffle, primarily because it was bad for cosmetics and I was actually trying to sell the speakers. (sad admission)
What I'm wondering about is whether there is an effect similar to diffraction whereby a broadly treated baffle causes axial response to fall, assuming that energy moving parallel to the baffle takes a right turn into the absorption. I realize that the change of edge reflections will create a complicating factor.
I need to set up my own experiment.
David
I've tested extensive treatments in a variety of ways. I've not seen anything I would describe as a fall in axial response other than what one would expect to see at extreme angles on which the wave passes through the felt. On the main axis, the only change is reduced diffraction effects. I've even been able to create an on-axis response above step of a raw driver on a small baffle that was almost exactly like that of the raw driver on a quasi-anechoic baffle.
I also don't think of the wave in terms of possibly "turning into" the felt. It enters into and passes through the felt as the sphere expands from a dome (for as much as it can be truly spherical vs. a true point source). Envision a pure expanding sphere. Then envision the portion of this sphere passing through the felt. It's magnitude reduces in level vs. the exterior portion and it may have a propagation speed change (as some postulate).
Then there is the portion of the sphere above the surface. For constant thickness felt, there will be an area that is "masked" if you will. A ray from the dome that is incident with the top edge of the felt and the surface of the felt describes an area "occluded" from direct wave movement. But a ray passing into the felt anywhere above the original baffle surface will eventually exit the top of the felt at increasing distances. So it provides a transitional area of direct wave vs. wave that has partially passed through the felt. As the rays are closer to parallel to the baffle surface, they pass though longer sections of the felt.
It's all very interesting, but also impossible for the DIYer to do much more than speculate as to the importance of any part of it. We can at least easily measure the final result. My article below has a bit of empirical testing of a few different applications, but I didn't go into an extensive off-axis demonstration. I've done that to satisfy myself that the off-axis was also largely improved when the on-axis was.
"Diffraction Doesn't Have to be a Problem"
It is unfortunate that felt is not a particularly good aesthetic in application. That doesn't stop me from using it, though.
Dave
Hi Dave,
I've been studying your extensive testing and agree that it doesn't show any effects beyond clearly killing the reflection artifacts. Several of your fully treated versions have more low end from the tweeter, so no hint that the felt is absorbing broad band to the point of pulling down low end.
Good stuff.
David S.
I've been studying your extensive testing and agree that it doesn't show any effects beyond clearly killing the reflection artifacts. Several of your fully treated versions have more low end from the tweeter, so no hint that the felt is absorbing broad band to the point of pulling down low end.
Good stuff.
David S.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.