Can someone say authoritatively if BSD "reflections" are the same as room reflections in their mathematical interaction with the acoustic wave? If the mental model of the pressure wave hitting a discontinuity (edge of baffle) and "bouncing back" is true, than it would seem intuitive that in a non-round baffle, BSD reflection effects would occur with non-constant delay; e.g. non minimum phase.
Simply being able to say "yes" or "no" to this question would seem to resolve the matter.
That does seem to imply something interesting about round baffles, doesn't it?
Simply being able to say "yes" or "no" to this question would seem to resolve the matter.
That does seem to imply something interesting about round baffles, doesn't it?"If what you say was true, then an equalizer could be used to correct all response anomalies caused by the room."
The room is so complex, but looking at one room mode in isolation a regular EQ can counteract it IF the room mode is measured carefully and the EQ circuit is tailored to exactly mimik the Q of the mode.
Also a loudspeaker is a resonant device in itself, but still considered a minimum phase system. A in band resonance in a driver does not change the driver into a non minimum phase device..? It just change the Fr. and phase according to the Q of the resonance. With a simple LC or RCL circuit the response can be flat and corrected in phase as well, resulting in a transient perfect system as far as a driver can be that.
"The strong implication would be that true BSC (phase as well as FR) would require FIR filters."
I do not think so.
I think BSC brings back frequency response and phase to normal.
Lets take a step repsonse, the leading edge leaves the driver and travels towards listener and baffle edges. After 0.5ms or so the wave reach the edge and suddenly experience a pressure drop due to some energy diffract around the cabinet.
What we need is a circuit that is ready to push some extra current into the driver at this point in order to counteract the baffle step, since we can not have a transient perfect system that is not flat in Fr. The delay in the inductor is mathced in time to the time it takes the signal to reach the baffle edge.
If this simple baffle step circuit would change the system into a non minimum phase system, we would see a phase turn similar to a high order (non 1st order) crossover.
For now I´m convinced I was right from the beginning but I am ready to change my mind instantly if someone comes with a better idea.
/Peter
The room is so complex, but looking at one room mode in isolation a regular EQ can counteract it IF the room mode is measured carefully and the EQ circuit is tailored to exactly mimik the Q of the mode.
Also a loudspeaker is a resonant device in itself, but still considered a minimum phase system. A in band resonance in a driver does not change the driver into a non minimum phase device..? It just change the Fr. and phase according to the Q of the resonance. With a simple LC or RCL circuit the response can be flat and corrected in phase as well, resulting in a transient perfect system as far as a driver can be that.
"The strong implication would be that true BSC (phase as well as FR) would require FIR filters."
I do not think so.
I think BSC brings back frequency response and phase to normal.
Lets take a step repsonse, the leading edge leaves the driver and travels towards listener and baffle edges. After 0.5ms or so the wave reach the edge and suddenly experience a pressure drop due to some energy diffract around the cabinet.
What we need is a circuit that is ready to push some extra current into the driver at this point in order to counteract the baffle step, since we can not have a transient perfect system that is not flat in Fr. The delay in the inductor is mathced in time to the time it takes the signal to reach the baffle edge.
If this simple baffle step circuit would change the system into a non minimum phase system, we would see a phase turn similar to a high order (non 1st order) crossover.
For now I´m convinced I was right from the beginning but I am ready to change my mind instantly if someone comes with a better idea.
/Peter
Well Pan, I respect what you are saying, but I think you are oversimplifying in assuming the steady state. In a listening room 20' long with the listening position in the middle, the back wall reflection is arriving 20ms after the start of the tone. There is no way you can correct this reflection!
Now, you may well be right that BSD is a different beast than this. But I haven't seen an argument that persuades me yet.
It's an interesting discussion, I just have a feeling that someone like MarkMck already knows the answer.
Now, you may well be right that BSD is a different beast than this. But I haven't seen an argument that persuades me yet.
It's an interesting discussion, I just have a feeling that someone like MarkMck already knows the answer.
What the speaker do and what the room do are two separate things. Of course they affect eachother (i´m not stupid ) but you must realize that you do not only have the back wall reflection, you have lots of reflections that come and go, adds and subtracts.
You can´t blame a speaker or a BSC circuit for the mess that the room contributes with. Even though you´d maybe like to balance the amount of BSC against the contribution from the surfaces of the room.
"but I think you are oversimplifying in assuming the steady state"
Actually I did talk about transient behaviour, not steady state. Allthough as far as the speaker goes you will get the steady state from the impulse or step response (but not the other way around).
/Peter
) but you must realize that you do not only have the back wall reflection, you have lots of reflections that come and go, adds and subtracts.You can´t blame a speaker or a BSC circuit for the mess that the room contributes with. Even though you´d maybe like to balance the amount of BSC against the contribution from the surfaces of the room.
"but I think you are oversimplifying in assuming the steady state"
Actually I did talk about transient behaviour, not steady state. Allthough as far as the speaker goes you will get the steady state from the impulse or step response (but not the other way around).
/Peter
Baffle diffraction -- of which Baffle Step is just a component -- is a continuum. At high frequencies, when the wave hits the discontinuity of the cabinet edge we get a reflection/reradiation of the wave. At low frequencies the wave does not even see the cabinet. At HF we have radiation into 2 pi steriradians, at LF 4 pi. In between is a continuum transitioning from one to the other.
In the last years my preference is to deal with the LF phenomenom by eliminating any baffle step from my boxes at the source -- a bi-pole (or alternatively a 0.5 speaker on the back if i'm not building with the FRs i prefer). As long as the box is symetrical back-to-front this perfectly compensates for Baffle Step (at the cost of an extra driver)
dave
In the last years my preference is to deal with the LF phenomenom by eliminating any baffle step from my boxes at the source -- a bi-pole (or alternatively a 0.5 speaker on the back if i'm not building with the FRs i prefer). As long as the box is symetrical back-to-front this perfectly compensates for Baffle Step (at the cost of an extra driver)
dave
Have you accurately simulated that or measured outdoors/anechoic chamber?
Mirage seems to have problems with getting a 100% result if my memory serves me.
My mental model of this is that the enclosure needs to be thin as a credit card in order to make this work perfect... but on the other hand, what is perfection.
I understand the basics and the function of this. It´s just that I have not spent much time on calculating exact what happens at what frequency so to speak.
/Peter
Mirage seems to have problems with getting a 100% result if my memory serves me.
My mental model of this is that the enclosure needs to be thin as a credit card in order to make this work perfect... but on the other hand, what is perfection.
I understand the basics and the function of this. It´s just that I have not spent much time on calculating exact what happens at what frequency so to speak.
/Peter
Pan said:
Nope... just a really good thot experiment and listening experience
They are multiway bipoles with a single bass driver if memory serves me right.
My mental model of this is that the enclosure needs to be thin as a credit card in order to make this work perfect... but on the other hand, what is perfection.
As long as the 2 bass drivers are within a half-wavelength below the baffle step they should be fine.
dave
I saw this thread a bit late, but I'll share my view on the subject:
A driver and the diffraction produced at the edges is not a minimum-phase system, but as we shall see, it can be seen as one in some cases anyway.
In the simplest case with the circular baffle, the diffraction from all locations along the baffle edge will coincide and form a secondary impulse in the impulse response, and such a system is not minimum phase. The circular baffle results in a baffle step with large oscillations (in the frequency domain) due to the sharp secondary impulse.
This means that a circular baffle with the driver in the centre is a bad idea, in fact it is about the worst I can imagine. It is better to smear the secondary impulse from the edge by assuring that the driver-to-edge distance varies in different directions. Parts of the edge should be really close to the driver, other parts should be far away. In that way, a clever driver positioning can smear the secondary pulse in such a way that the amplitude and phase response becomes very close to that of a minimum phase system at low and mid frequencies. At high frequencies, the driver directivity reduces the radiation toward the edge, which makes the edge reflections and thus the minimum phase issue less important. All in all, this can make the resulting response close to a simple first order (one pole, one zero) system, and this can be compensated by means of a simple coil and resistor.
The error that remains is mostly small compared to other errors in the design, and can even be utilised to compensate other errors in some cases.
You can try this with my little "Edge" hack, and see that with a clever positioning of the driver, and with a simple compensating network, both phase and amplitude can become reasonably flat.
One final thing: The result of the diffraction at the edge is principally similar to a reflection at a wall in the sense that it produces an"echo" in the impulse response. However, the time that elapses between the first pulse and the reflection is typically larger in the wall reflection case due to a longer distance. The two cases are quite different in psychoacoustical terms. If the delay is short (as in the baffle diffraction case) the reflection merges psychoacoustically with the original sound and results in a colouration of the sound. In the wall case we will hear the reflection as reverberation or an echo if the delay is really long. The colouration can be compensated by a simple passive network, on the other hand the wall reflections probably need digital means to be compensated to inaudibility.
A driver and the diffraction produced at the edges is not a minimum-phase system, but as we shall see, it can be seen as one in some cases anyway.
In the simplest case with the circular baffle, the diffraction from all locations along the baffle edge will coincide and form a secondary impulse in the impulse response, and such a system is not minimum phase. The circular baffle results in a baffle step with large oscillations (in the frequency domain) due to the sharp secondary impulse.
This means that a circular baffle with the driver in the centre is a bad idea, in fact it is about the worst I can imagine. It is better to smear the secondary impulse from the edge by assuring that the driver-to-edge distance varies in different directions. Parts of the edge should be really close to the driver, other parts should be far away. In that way, a clever driver positioning can smear the secondary pulse in such a way that the amplitude and phase response becomes very close to that of a minimum phase system at low and mid frequencies. At high frequencies, the driver directivity reduces the radiation toward the edge, which makes the edge reflections and thus the minimum phase issue less important. All in all, this can make the resulting response close to a simple first order (one pole, one zero) system, and this can be compensated by means of a simple coil and resistor.
The error that remains is mostly small compared to other errors in the design, and can even be utilised to compensate other errors in some cases.
You can try this with my little "Edge" hack, and see that with a clever positioning of the driver, and with a simple compensating network, both phase and amplitude can become reasonably flat.
One final thing: The result of the diffraction at the edge is principally similar to a reflection at a wall in the sense that it produces an"echo" in the impulse response. However, the time that elapses between the first pulse and the reflection is typically larger in the wall reflection case due to a longer distance. The two cases are quite different in psychoacoustical terms. If the delay is short (as in the baffle diffraction case) the reflection merges psychoacoustically with the original sound and results in a colouration of the sound. In the wall case we will hear the reflection as reverberation or an echo if the delay is really long. The colouration can be compensated by a simple passive network, on the other hand the wall reflections probably need digital means to be compensated to inaudibility.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.