DDF said:
DDF,
I believe the reason small boxes image well is that the time delay is so short that our ears/brain can't differentiate between the source and the diffraction. Note that with large baffles the spl will be lower at the edge (it decreases by 6db for each doubling of distance), so the pressures will be lower and diffraction effects as well, not to mention that they occur lower in frequency.
Don't get stuck with the idea that OB's have to be large. These remain my best imaging OB's. Note they're for a 4" driver, and the driver below provides the bass.
An externally hosted image should be here but it was not working when we last tested it.
johninCR said:
I've heard some big (and very big) horn systems that imaged like champs. I wonder if they behave differently than baffles?
If this edge diffraction/reflection thing is a problem that is causing phantom images, what can be done about it? I know what Bud will say! =)
Could the edges be treated with slots, mesh, towels, fur or something to lessen the abrupt pressure change? It seems to work well on Horn Edges.
I guess everybody has seen this?
"The Baffle Diffraction Simulator is a design program and education tool for predicting Baffle Diffraction Response and Baffle Step Gain in Closed Box Systems. It includes Multiple Driver Radiation Models for Piston and Ribbon Devices as Individual, MMT, MTM and small line arrays groupings, on three, four, and five sided Baffles with Sharp, Beveled & Rounded Edges. Program outputs Response Charts, Diagrammatic Layouts and Frequency Response Data files and maintains an extensive example design library with a session librarian capability."
http://www.pvconsultants.com/audio/diffraction/downloadbds.htm
Oops! Just noticed that Dave got there first in post 1305.
Cheers,
Mike
"The Baffle Diffraction Simulator is a design program and education tool for predicting Baffle Diffraction Response and Baffle Step Gain in Closed Box Systems. It includes Multiple Driver Radiation Models for Piston and Ribbon Devices as Individual, MMT, MTM and small line arrays groupings, on three, four, and five sided Baffles with Sharp, Beveled & Rounded Edges. Program outputs Response Charts, Diagrammatic Layouts and Frequency Response Data files and maintains an extensive example design library with a session librarian capability."
http://www.pvconsultants.com/audio/diffraction/downloadbds.htm
Oops! Just noticed that Dave got there first in post 1305.
Cheers,
Mike
BudP said:
It's a funny image for sure - but I've done it! I have a little pair of 2-ways that are covered in faux tiger fur. Looks cool and sounds terrrrrific!
I wanted to cover my big OBs with that long pile faux fur - in shocking pink. But the wife was having none of it. Oh well....
Think of the fuzzy windsocks that are so popular for microphones these days. Would that stuff reduce edge diffraction and scatter? Doubt it wll ever be a popular speaker treatment.
Hi Michael,
The best experimental baffle I had was a roll of several turns dense pile carpet either side of a minimum size wooden LS mounting board. It was like a figure of eight with the gap between the rolls closer together behind the LS such that the rear exit area equalled cone are, but it was not 'domestically' acceptable.
Cheers .......... Graham.
The best experimental baffle I had was a roll of several turns dense pile carpet either side of a minimum size wooden LS mounting board. It was like a figure of eight with the gap between the rolls closer together behind the LS such that the rear exit area equalled cone are, but it was not 'domestically' acceptable.
Cheers .......... Graham.
Waveguides behave quite differently than pistons in a baffle, especially as regards edge diffraction.
I think that there is a lot of evidence to say that smaller speakers image well because they are so small and the time delays are small. But what about output level? You simply cannot get decently high SPL's from these "good imaging" small speakers.
What can be done to get good imaging (per the above quote), low coloration AND decent SPL levels all at the same time? - read about the Summas as that is exactly what they were designed to do. These three things are not easy to achieve in a design, usually you will get two of the three, but seldom will you find a speaker that gets all three right.
panomaniac said:I have a little pair of 2-ways that are covered in faux tiger fur. Looks cool and sounds terrrrrific!
I wanted to cover my big OBs with that long pile faux fur - in shocking pink. But the wife was having none of it. Oh well....
Think of the fuzzy windsocks that are so popular for microphones these days. Would that stuff reduce edge diffraction and scatter? Doubt it wll ever be a popular speaker treatment.
These kinds of materials can be very good at sound absorption if the sound propagates along them, but they tend to not work so well for incident sound, as on a wall - too thin.
My daughter has requested a pair of my speakers in "hot pink" - maybe I'll also cover them in pink Faux fur!! I'm not sure that they would be a big seller in the marketplace though.
In post #1375 above, JohninCR wrote;-
>> Too bad the scientists can't explain exactly what happens with edge diffraction <<
Hi Bratislav,
All the diffraction software, and your link are excellent, but maybe there are differences relating to the visualisation of this audible interference.
Source radiation might be nulled in line with a baffle, but air flow perpendicular to a baffle is maximum, and thus any edge disturbing that airflow and modifying the pressure gradient must cause an equal-but-opposite-reaction sonic disturbance.
Hi JohnK,
Does your program examine from a wavefront viewpoint only ?
What about the quadrature related air flow, and the new radiation which alternating air flow past the edge of a stationary object cannot fail to generate ?
In post #1375 above, JohninCR wrote;-
>> With dipoles the pressure change at the edge is double, << (compared to monopole)
So does JohninCR's second statement not suggest that a bipole would produce not only less baffle step radiation change with frequency , but also less edge generated secondary interference wrt to original radiation, and thus by keeping the bipole narrow the source point could remain small through a much greater range of AF ? A bipole will also increase SPL.
( I say this knowing that one of the cleanest sounds I ever heard was from a 12" driver on a drawing board whilst a student, but that this was at a distance of at least 30m in a very large non-domestic room. )
Cheers .......... Graham.
>> Too bad the scientists can't explain exactly what happens with edge diffraction <<
Hi Bratislav,
All the diffraction software, and your link are excellent, but maybe there are differences relating to the visualisation of this audible interference.
Source radiation might be nulled in line with a baffle, but air flow perpendicular to a baffle is maximum, and thus any edge disturbing that airflow and modifying the pressure gradient must cause an equal-but-opposite-reaction sonic disturbance.
Hi JohnK,
Does your program examine from a wavefront viewpoint only ?
What about the quadrature related air flow, and the new radiation which alternating air flow past the edge of a stationary object cannot fail to generate ?
In post #1375 above, JohninCR wrote;-
>> With dipoles the pressure change at the edge is double, << (compared to monopole)
So does JohninCR's second statement not suggest that a bipole would produce not only less baffle step radiation change with frequency , but also less edge generated secondary interference wrt to original radiation, and thus by keeping the bipole narrow the source point could remain small through a much greater range of AF ? A bipole will also increase SPL.
( I say this knowing that one of the cleanest sounds I ever heard was from a 12" driver on a drawing board whilst a student, but that this was at a distance of at least 30m in a very large non-domestic room. )
Cheers .......... Graham.
Another aspect comes to mind relating to simulation, in that it is swept frequency and sinusoidal ONLY.
Real music waveforms are composite and asymmetrical.
What happens at a computed diffraction point when there is considerable low frequency air flow ?
Are there any cross/intermodulation effects due to the virtual source being shifted over the edge boundary ?
Cheers .......... Graham.
Real music waveforms are composite and asymmetrical.
What happens at a computed diffraction point when there is considerable low frequency air flow ?
Are there any cross/intermodulation effects due to the virtual source being shifted over the edge boundary ?
Cheers .......... Graham.
And here is the cone treatment to suit.
http://www.google.com/patents?id=ULEeAAAAEBAJ&printsec=abstract&zoom=4#PPA3,M1
Bud
I've nothing to add to the minimum phase question, but I'm surprised this diffraction issue is so tough to clear up. I always thought of it in the same terms as lynn, wave impedance and reflected power. See if this makes sense, A wave front is emitted by the speaker and travels outwards along the front baffle. When it reaches the edge the sudden change in impedance causes some of the sound to reflect, and I believe this reflection has it's phase flipped. So in the case of a sealed box you should have the original sound source in the center and weaker secondary sources along the outline of the front baffle that are out of phase with the original source (at least for wavelengths that are long relative to the baffle dimensions).
With an open baffle of the same shape you also have a primary sound source in the center and weaker out of phase sound coming around the baffle edges from the rear radiation of the speaker so it's a different mechanism and we don't call it diffraction, it's just the nature of a dipole, but it is a similar result. Presumably you have the same set of primary sound source and out of phase baffle edge sound sources as the boxed monopole but the out of phase secondary sources are IN phase with the rear radiation and coming from the same location, so they should sum with the rear output for an infinitely thin baffle and you're back to simple dipole action.
In my reading of Linkwitz's diffraction section this seemed to be what he was saying. You could say a dipole has far more diffraction than a monopole because more out of phase sound is coming from the edges of the baffle, but it's sort of stating the obvious and maybe misusing the term "diffraction". It's a dipole, it has in phase and out of phase sound sources, you'd expect this.
I'm no expert and I'm not emotionally invested in any of this so please feel free to set me right if I'm missing something here.
With an open baffle of the same shape you also have a primary sound source in the center and weaker out of phase sound coming around the baffle edges from the rear radiation of the speaker so it's a different mechanism and we don't call it diffraction, it's just the nature of a dipole, but it is a similar result. Presumably you have the same set of primary sound source and out of phase baffle edge sound sources as the boxed monopole but the out of phase secondary sources are IN phase with the rear radiation and coming from the same location, so they should sum with the rear output for an infinitely thin baffle and you're back to simple dipole action.
In my reading of Linkwitz's diffraction section this seemed to be what he was saying. You could say a dipole has far more diffraction than a monopole because more out of phase sound is coming from the edges of the baffle, but it's sort of stating the obvious and maybe misusing the term "diffraction". It's a dipole, it has in phase and out of phase sound sources, you'd expect this.
I'm no expert and I'm not emotionally invested in any of this so please feel free to set me right if I'm missing something here.
Good Golly Bud! Fuzzy speaker cones!
You're the cone treatment expert, Bud - what do you think?
Hey, I was only 1/2 kidding. There does seem to be something to the furry-fuzzy surface. Fur is grown as an insulator and skin protector, but it seems to have some interesting sonic properties because of the way it traps air.
I have heard good results with felt and even sandpaper inside horn mouths. Earl G is right that this type of furry stuff is too thin to do much good on reflected sound, but when the sound is moving perpendicular to the hair, it's another matter. That was what the sandpaper was said to do - create a slight turbulent layer of air that lowers reflections of the horn walls.
It's seems silly, but it's easy enough to play with. I seem to be getting good results with towel material or faux fur over foam at the horn mouth. Mr. Earl fills the wave guide with foam.
But that a subject for 2 threads that already exist.
Geddes on Waveguides and Horn Honk and Towels.
However, strange baffle edge treatments should be fair game here.
You're the cone treatment expert, Bud - what do you think?
Hey, I was only 1/2 kidding. There does seem to be something to the furry-fuzzy surface. Fur is grown as an insulator and skin protector, but it seems to have some interesting sonic properties because of the way it traps air.
I have heard good results with felt and even sandpaper inside horn mouths. Earl G is right that this type of furry stuff is too thin to do much good on reflected sound, but when the sound is moving perpendicular to the hair, it's another matter. That was what the sandpaper was said to do - create a slight turbulent layer of air that lowers reflections of the horn walls.
It's seems silly, but it's easy enough to play with. I seem to be getting good results with towel material or faux fur over foam at the horn mouth. Mr. Earl fills the wave guide with foam.
But that a subject for 2 threads that already exist.
Geddes on Waveguides and Horn Honk and Towels.
However, strange baffle edge treatments should be fair game here.
Hi
John K, already illuminated at least ONE way to go. There is comb filtering due to the baffle as an obstacle creating different path lengths from the front an back side of the speaker AND there is the second source created right at the baffle edge due to diffraction
This second source radiates SPL from the same place in space from where the back wave arrives ( the baffle edge ) - seen from the listening position ( no wall reflections taken into account ). Therefor this second source can be seen as a REAL HELPER to improve the overall performance.
As John K outlined with his few simulations the thickness is one variable that can be addressed for optimisation as the two monopoles at the front and the rear edge of the baffle form a dipole of different strength when set at different distance ( T ). An other variable is the distance of the edge to the driver and I bet there are combinations of T and D that work better or worse.
For a rectangular or any non circular baffle shape this would require to vary the thickness of the baffle at its edge related to the distance to the driver.
Sure, this does not fix anything in the time domain but with a dipole speaker the time domain issue is inherent to the principle of operation as I see it.
SL outlined at his page that the upper frequency range becomes even less critical in this respect as the edge is illuminated less due to directivity ( wave guides may help further but also may introduce new hurdles to overcome, Earl and some others seem to be specialised in this topic )
-----------
Lynn, what makes me a little bit reserved about the mesh treatment is that it's basically the same approach that lead to the stuffing of boxes. To me ANY absorbent material adds to what you want get rid of – coloration of sound.
If this wouldn't be so, absorbent material would be used in excess.
I never was able to find the optimum of damping a certain resonance and not making the sound dull at the same time.
Beside the dampening effect these materials also seem to alter sound speed which makes things no better ( gradual multi path effects here as well ).
Greetings
Michael
John K, already illuminated at least ONE way to go. There is comb filtering due to the baffle as an obstacle creating different path lengths from the front an back side of the speaker AND there is the second source created right at the baffle edge due to diffraction
This second source radiates SPL from the same place in space from where the back wave arrives ( the baffle edge ) - seen from the listening position ( no wall reflections taken into account ). Therefor this second source can be seen as a REAL HELPER to improve the overall performance.
As John K outlined with his few simulations the thickness is one variable that can be addressed for optimisation as the two monopoles at the front and the rear edge of the baffle form a dipole of different strength when set at different distance ( T ). An other variable is the distance of the edge to the driver and I bet there are combinations of T and D that work better or worse.
For a rectangular or any non circular baffle shape this would require to vary the thickness of the baffle at its edge related to the distance to the driver.
Sure, this does not fix anything in the time domain but with a dipole speaker the time domain issue is inherent to the principle of operation as I see it.
SL outlined at his page that the upper frequency range becomes even less critical in this respect as the edge is illuminated less due to directivity ( wave guides may help further but also may introduce new hurdles to overcome, Earl and some others seem to be specialised in this topic )
-----------
Lynn, what makes me a little bit reserved about the mesh treatment is that it's basically the same approach that lead to the stuffing of boxes. To me ANY absorbent material adds to what you want get rid of – coloration of sound.
If this wouldn't be so, absorbent material would be used in excess.
I never was able to find the optimum of damping a certain resonance and not making the sound dull at the same time.
Beside the dampening effect these materials also seem to alter sound speed which makes things no better ( gradual multi path effects here as well ).
Greetings
Michael
I can see one error I made already, where I said "for wavelengths that are long relative to the baffle dimensions". In that case for the monopole there should be little or no diffraction because it doesn't really "see" the baffle. I was thinking about how at higher frequencies the baffle edge sound source could reach the listener IN phase with the direct sound, but wanted to focus attention on the phase reversal of the edge source. Hope I'm not hopelessly confused
Graham Maynard said:
Graham,
visualization and audibility are subjective side of this equasion. I'm not even going to attempt to go there.
But Huygens principle is a powerful tool, and with abundance of computing power it becomes very easy to implement. Basically it treats every molecule of air that gets hit by the wavefront as a new source of sound that radiates 4 pi. Together with all of the other particles that get excited at the same time it integrates into a smooth spherical wavefront. When a solid obstacle is hit (baffle edge), the molecules behind it can't be excited by original wavefront (they are shielded). Hence, molecule right at the edge suddenly gets to "radiate" into a free, unexcited area of air. We get another semispherical wavefront starting right there (semi as molecules above the obstacle still carry the original wavefront).
This is easily simulated. It works and it is proven again and again in all fields of physics. Saying that simulation doesn't work is plain silly. What could perhaps happen is that maybe our model isn't right (GIGO principle), but simulations work.
Period.
Bratislav
Bratislav said:
Bratislav,
No one is arguing that Huygens principle isn't correct. In fact open alignments wouldn't work like they do if the front and rear waves didn't both act like the animations in your link. No one said simulations don't work either. Neither, however, explain what is happening at the edge to my satisfaction, nor have I seen any good explanation. For all I know the problem could be all attributable to some kind of surface thing that BudP talks about. I do know there's something going on out there at the edge that requires further explanation, and its effects are greater with open alignments.
Now that I think about it in terms of molecule spreading like the animation in your link, the problem may very well be occuring right at the surface where these new little sources begin as molecules bump into each keep being constrained by the baffle. Truthfully, I'd be flabbergasted if his little blocks work on baffles, but if it is something going on right at the surface that builds up until hits a boundary, the the little blocks may to something to diffuse that energy much like they do in the wavetank. I never really put any thought into what goes on with the molecules at the surface as waves travel parallel to a solid surface. I'd still like to have something more to go on than just paint, but I'll give about anything a go if I can't find a hole in the concept. Hmmm.
