Baffle Step - Discussion

[brianpowers27]

It is my understanding that a typical baffle has 2 dimensions that will affect overall Baffle step loss, not just one dimension. It is my theory that both dimensions count. I have heard this voiced in such a way that describes only the minimum dimension as applicable. THis is usally width. Typically one dimension of a baffle is long enough that the Fc becomes very low and nearly insignifigant.

Furthermore I believe that is one dimension is butted against another object or a wall it will experience a 3db loss at the Fc versus a 6db loss, for the one baffle step dimesion. This assumes that the object is of inifinite size...

Could someone more experienced chip in on this discussion?

[gumert]

I'm just regurgitating what I've read:


At frequencies where the wavelength is shorter than the distance from the center of the driver to the shortest edge of the baffle, the baffle tends to block the sound from radiating around the sides and rear of the speaker, and redirect it forward, like the reflector on a flashlight.

http://www.geocities.com/cc00541/bafflestep.html

No comment about putting things up against walls because I don't know enough to say anything meaningful.

[Ludo]

Brian,

I wish I could find the paper I read from a while ago which talks about baffle step as a reflection phenomena; where every part of the baffle contributes to reflected, delayed versions of the sound wave which then sum at the listening position. I will search around some more and post it when I find it, but I believe you are right that to get an accurate view of the impact of baffle edge diffraction on the frequency response, the x/y dimensions of the baffle as well as the chape and contour must be accounted for.

http://www.trueaudio.com/st_diff1.htm

Louis

[brianpowers27]

It is my understanding that the baffle diffraction is a reflection while the baffle step loss is an acoustic impedance/radiation change.

[Æ]

There is a small program called Best Place by Allison Acoustics that will calculate a bass augmentation curve based on where you place your loudspeaker in relation to the floor, back wall, side wall, etc. The program was originally designed to run on Windows 3.1 and would also run on Win95/98. Not sure about anything else. You may be able to locate a copy on the net. I have a copy on my hard disk, not sure if it is usable, because the copy has been around back and forth a couple of times. But I'm sure someone here could locate a good copy for you.



I'm just regurgitating what I've read:



http://www.geocities.com/cc00541/bafflestep.html

No comment about putting things up against walls because I don't know enough to say anything meaningful.

[brianpowers27]

There is a small program called Best Place by Allison Acoustics that will calculate a bass augmentation curve based on where you place your loudspeaker in relation to the floor, back wall, side wall, etc. The program was originally designed to run on Windows 3.1 and would also run on Win95/98. Not sure about anything else. You may be able to locate a copy on the net. I have a copy on my hard disk, not sure if it is usable, because the copy has been around back and forth a couple of times. But I'm sure someone here could locate a good copy for you.

XBaffle does something very similar but is designed for open baffles. I misplaced the link for this. If someone knows where to find it , please let me know.

[curt_c]

Have you played with the Edge program? Simple, but illuminating.

http://www.tolvan.com/

C

[brianp]

It is my understanding that a typical baffle has 2 dimensions that will affect overall Baffle step loss, not just one dimension. It is my theory that both dimensions count. I have heard this voiced in such a way that describes only the minimum dimension as applicable. THis is usally width. Typically one dimension of a baffle is long enough that the Fc becomes very low and nearly insignifigant.


http://www.musicanddesign.com/Equivalent_Baffles.html

The cited article relates specifically to open (dipole) baffles, but shows that the baffle induced response derives from the complex sum of ALL the distances from the center of the driver to EVERY point on the edge of the baffle -- not just the shortest distance.

[moosespeaker]

Brian - I believe this is the link to that program.:D

http://www.classicspeakerpages.net/links.html.

Go to the Site Map and under "Other" is the program (version 1.0) I am at work and cannot play with it to see if it is the one refereenced.

George

[Dirk]

Have you played with the Edge program? Simple, but illuminating.

http://www.tolvan.com/

C

I was just trying to get the Baffle Diffraction Spreadsheet to work under Office 2007. It does, sorta.

I just simulated a baffle that is 10"x24" and a baffle that was 10"x44". I assumed that the drivers didn't move, so it was a sort of "montor vs floor stander" simulation. The baffle step corner frequency for both were identical (as was the final -6dB loss). The only real difference was that the magnitude of the ripple was slightly higher for the small box. When I say "slightly", I mean about .25dB

In the other thread, Brian said:


Why is it based off of the smallest dimension? Shouldn't the change in impedance affect the total acoustic output from all 4 sides? Doesn't the change in acoustic impedance from the floor or other boundary account for anything?

I think there are two effects here. There's baffle step and boundary gain. Remember how Jeff talked about how putting a speaker up against a wall will bring back bass, but still leave a hole in the response? That's because boundary gain doesn't kick in at the same frequencies.

[brianpowers27]

I went to the response modeler 2.1 and tried various combinations. I don't seem to be able to acheive any Baffle curve with a transfer loss much greater than 6 db. I would expect to lose 6db on the x dimension and 6 db on the y dimension. I don't see this. Now my question is one of why...

I think there are two effects here. There's baffle step and boundary gain. Remember how Jeff talked about how putting a speaker up against a wall will bring back bass, but still leave a hole in the response? That's because boundary gain doesn't kick in at the same frequencies.


Agreed... There is boundary gain and baffle loss. I find an inwall speaker to be the only siutation where there is no boundary loss.

Now let's imagine a setup where there is a tower of near infinite hieght where the woofer is sufficiently far from any boundary as not to exhibit any boundary gain. Let's say 40" tall x 10" wide with the woofer near the top of the 40".

I would expect the woofer to have a full step loss on the x dimension (10") and a half step loss on the y dimension (40"). The Y-dimension would only be a half step loss due to one end of the dimension never allowing for the transition to 4pi space...

I know I am not the one to answer the question. I am just clarifying the question.

[Dirk]

I would expect the woofer to have a full step loss on the x dimension (10") and a half step loss on the y dimension (40"). The Y-dimension would only be a half step loss due to one end of the dimension never allowing for the transition to 4pi space...

I just tried extending the height of the cabinet to 200" (the highest I could convince BDS to take without errors in Office07). Not infinitely tall, but the best I could do. The diffraction ripple was the same basic shape and magnitude. The graph just took on a bunch of teeny-tiny squiggles, like the previous run had smoothing applied to the result. Or maybe that the 2nd sim was just higher in resolution. I can try and post screen captures if that makes no sense.

My take is that secondary dimension makes so little difference as to not be worth mentioning. That makes sense to me -- taking take your "acoustic impedance" notion a step further. If you have a 200ohm resistor in parallel with a 4ohm resistor, where will most of the energy go? Baffle step losses affect ALL frequencies below the corner. Why should some of those frequencies "wait around" to be affected by the longer dimension? They're long gone, gettin' all four-pie with their bad selves.

[brianpowers27]

I suppose the post above from John K's site has applicability even in the non-ob world. His article suggests that the corner Frequency is the integral of all distances from source to edge.

[billfitzmaurice]

It is my understanding that a typical baffle has 2 dimensions that will affect overall Baffle step loss, not just one dimension. It has an effect over 360 degrees, so accurately charting it would require a series of polar plots. What one is mainly concerned with is dispersion and response on the horizontal plane, so what one mainly considers is the baffle width.

[dlr]

Brian,

Baffle step is simply one of the effects of diffraction. All baffle edges come into play, including any oddly shaped baffles. The common misunderstanding that only the width matters is primarily due to tall, thin baffles. Even for those, if a tweeter is located near the top, then the three near edges have a strong influence whereas the other has little due to the distance, thus the frequency.

The other reason for a strong influence of one or the other is the included arc that is defined by the edge. The diffraction signature for any point in space is an integration of all edge diffraction. Several things influence the impact at any given point on the baffle edge. These are distance (from all points on the driver), driver directionality, edge shape, angle from any point to the sample point in space and secondary diffraction of the original diffracted wave (usually not included in most calcs I think).

Read the page I host at my site by Andy Unruh, "Understanding Cabinet Edge Diffraction". I have a couple of other pages dealing with treating it.

Step appears to be unique simply due to the magnitude and eventual loss in the transition from 2-pi to 4-pi. It's probably the most influential due to the loss at the low end.

Stand-mounted 2-ways are the most sensitive to it primarily due to the short vertical dimension. But mount the tweeter right at the top of a tall baffle and the vertical edge takes on significance.

Be sure to use the highest settings in the BDS to see the most realistic responses.

dlr

[Ludo]

It is my understanding that the baffle diffraction is a reflection while the baffle step loss is an acoustic impedance/radiation change.

Brian,

My understanding is that the two are tightly inter-related. Whenever the acoustic wave encounters a change in the acoustic impedance there is a corresponding reflection. This is consistent with baffle step occuring even with a spherical cabinet.

As the acoustic wave propagates out from the driver it encounters the edge of the baffle where there will be a sudden drop in acoustic impedance. This will cause part of the wave to start propagating towards the rear of the cabinet, and a phase-reversed, reflected acoustic wave to propagate in the forward direction.

At frequencies where the phase-reversed, reflected wave is delayed by less than 1/4 wave length relative to the primary forward wave, the two will add destructively and there will be a net reduction in the acoustic output in the forward direction. This is the case below the baffle step frequency.

At frequencies where the phase-reversed, reflected wave is delayed by more than a 1/4 wave length relative to the primary wave, the two will add constructively and there will be a net increase in the acoustic output in the forward direction. This is the case above the baffle step frequency.

Louis

[rone]

It is my understanding that the baffle diffraction is a reflection while the baffle step loss is an acoustic impedance/radiation change.

Nope, they are the exact same phenomena - a reradiation caused by an impedance change. I helped out a bit with the BDS, so you know I am not blowing smoke. As far as explaining it, I won't waste my breath on what others have done better. There are some nice Diffraction articles in JAES by Bews, Hawksford, Backman, Lipshitz, Vanderkooy.

Andy Unruh's article is based on the latter two and is not really complete, IIRC, but it has been years....