What would happen if you covered a speaker with 2" of fiberglass to reduce diffraction?

[kemmler3D]

There is an interesting idea that has burbled around in the background for a while.
What if a waveguide was made from something that exhibits some absorption? It isn’t really clear what the answer might be, but the answer might not always be bad. Modelling it would be challenging, the usual BEM assumptions probably break badly. But there is the nagging feeling that maybe the idea could prove useful.
It opens up the whole box of possible meta material ideas as contributors to a design. But at a significant leap in complexity and general grief.
Perfect for hobbyists.

So, thinking about this and looking at the pic @Duke posted, I think a waveguide that absorbs energy above the crossover frequency of the driver in question could be really interesting, in that it would tend to absorb distortion harmonics but leave the intended output alone.

The difficulty is, distortion harmonics are mostly a problem for woofers and mids, which tend not to need or work well with deep waveguides. But say we put one on the midrange in a 3-way anyway. There's also the question of what kinds of materials absorb 2-4khz+ super-well, but leave 2khz and below alone.

There are probably also a lot of issues I'm not even thinking of.

 

[hex168]

There might be a benefit to clarity by reducing early reflections, so covering most of the front and sides of the cabinet may help in that regard. Of course, the on- and off-axis responses would have to be managed, but I think a region of controlled directivity followed by a quick drop-off further off-axis would be OK. I vaguely remember reading something about that being particularly desirable for multichannel, but cannot remember the source.

 

[Tim Link]

I built a pair of midrange horns years ago that are, to put it mildly, not the best sounding horns out there. I still use them but their frequency bandwidth is very limited now. As I came to understand the issues with these honkers, I at one point filled the mouth of the horns with a fiberglass panel. This acted like a high pass filter and drastically reduced the efficiency of the horns. It also made them sound dramatically more detailed and clear after appropriate EQ. I gave up on that approach and now just use them in the narrow band where they work without disturbing me.
I've been thinking about the benefits and costs of using absorption again recently. The most obvious cost is lost efficiency. It seems counterproductive to have a driver make sound only to then absorb a portion of it before it can be heard. It could be argued though that some of the sound is unwanted, and it may be reasonable in some circumstances to use immediate absorption near the driver.
I just did a little experiment with a Dayton 2" aluminum midrange dome. It gets beamy by 10 kHz. I've been pondering how that works and concluded it's simply that the center of the dome is where most of the sound is generated because that's the surface that's most perpendicular to the motion of the dome. The center has relatively less curvature than the entire dome, and the relatively flat center area is big enough that for higher frequencies they get beamed. So what if sound from the center of the dome were absorbed expertly so that the sound from all surfaces of the dome radiated evenly? That shouldn't beam so much, but it won't be as loud in the highs. Or what if I force the sound through a smaller opening in the middle? I pulled out an old wooden conical horn I had made with a 1" throat and put it in front of the driver, resting the throat opening on the grill, which creates a significant gap from the driver's dome to the throat. To control sound escaping to the sides I cut a 2" hole in some 2" thick fiberglass and put the driver in the hole, with the horn on top of all that. So any sound that doesn't make it into the throat of the horn ends up going into the fiberglass.

The result was a loss of output on-axis at the highest frequencies, but not off axis. So the off axis (30 degrees) and on axis were now the same. I didn't expect this because I though 1" was still going to be too big, but I guess the edges of the throat must have added some diffractive benefit. On the bottom end at 400 Hz there was a 6 dB gain both on axis and off axis. So this kind of worked! I could EQ it and have flat response on axis and 30 degrees off axis from 400 Hz to up over 13,000 Hz, with a 6 dB boost on the bottom. That's arguably better than the dome's flat baffle mounted performance. It'd be more ideal to collect all the sound coming off the dome's surface at the high frequencies and properly direct it with some kind of phase alignment device for maximum efficiency. But that's complicated. I saw in earlier tests that if I just stick the dome in a waveguide with a 2" throat the sound coming off the center of the dome will bounce off the waveguide and create a messy dispersion pattern. So something complicated like kef's tangerine slice thing would need to be constructed and carefully optimized to gain back that efficiency without making a mess of the dispersion. So for a DIY project it seems like a person could get a really nice dispersion pattern without needing to do extensive design and testing of a wave guide. Just make a straight sided cone, put the driver a short distance behind the throat and absorb any high frequencies that didn't go straight into the throat before they can bounce around and come out later.
On top of that, you could even potentially load some small woofers back there behind that insulation to further extend the low frequency response to whatever is possible to squeeze out of a 1" throat. This might be an impressive preformer that's not too complicated to build. I guess it's up to me to find out. I've got the drivers. Just got to buy and cut some wood and fiberglass, and maybe some small woofers.

 

[norman bates]

Diffraction Doesn't Have to be a Problem

This is a great article.
Is it the end all, be all ?
Nope, but a starting point.

I find (subjectively) it can increase sharpness or help the micro details, maybe it doesn't........

 

[Tim Link]

I find (subjectively) it can increase sharpness or help the micro details, maybe it doesn't........

I can't imagine it would hurt those things to smooth out the response both on and off axis. An experiment could be done on headphones where a music signal is altered to mimic the effects of the diffraction. The best way to do it I think would be to measure the anechoic response of a diffraction laden speaker at listening distance, and then convolve that impulse against the music and listen on headphones. I just need to figure out how to get an anechoic impulse response from a diffraction laden speaker.