This is what people have been trying to explain for decades.
Impulse response = freq resp
Same thing, different way of looking at it.
The goal would be accurate recreation of the signal. As the driver executes the steady signal, we can predict some of the errors it will make looking at the IR. So as you say, there is what happens in reality, or in complex signal instead of an impulse, to consider. That difference I'm interested in is Transient vs Steady state. I also think that steady state is just a succession of transients in some way, as in, every oscillation is a transient in its own right. In comparison to its resting position, easily. In music/media the transients we refer to are the true peak transients, that live generally some 15db over max average. High Excursion always exaggerates the character flaws of the IR. A true peak transient doing 112db/1m will be highly excitatory in comparison to the average spl oscillations, and the decay of the True peak could cause the average spl to be less accurate, in its wake. This is called timbre I believe.
Mass isn't exactly the big deal because if the Q specs says things are highly dampened, they arrived there with Mass obviously already in the formula. Still there will a difference between light and heavy cones. Its inertia. Inertia is higher when mass is higher. This is why I think we will see lower distortions from quality drivers with light cones, at lower spl, but they cannot stay rigid enough to compete with higher mass cones, so high quality drivers on this side of the scale will have better distortion specs at high spl.
Mass isn't exactly the big deal because if the Q specs says things are highly dampened, they arrived there with Mass obviously already in the formula. Still there will a difference between light and heavy cones. Its inertia. Inertia is higher when mass is higher. This is why I think we will see lower distortions from quality drivers with light cones, at lower spl, but they cannot stay rigid enough to compete with higher mass cones, so high quality drivers on this side of the scale will have better distortion specs at high spl.
Is this suggesting that there is no other care as long as FR is acceptable in regards to what you choose to include in the passband? In other words, not even Le matters, as long as sensitivity is reasonable its all good>? I think the consequences of high Le will be seen in IMD/THD well before it attenuates the actual frequencies you want to play.
Yes, that is far far more correct than not. .......... IF frequency response is both frequency magnitude and phase.
Tis the most fundamental, yet still least understood, principle in audio I think.
Well, a perfect reproducer of the input signal - in any terms. So while the FR-Pulse relation is mathematically correct, I doubt that we can do a slow sweep during 30 seconds and then say, look - her is the pulse response of this speaker. If I could choose to do one measurement only of the two, I would choose the a proper pulse as a stimuli. Why, because the sweep lacks the jolt that a pulse will induce into the DUT. It so much harder to cope with that than a slow sweep. And music is much more about onset than constant tones so I think also from that perspective, it would be a better test. Now, apparently it is hard from a S/N perspective to do this kind of test. Hence why we never see them. But it could be explored more.
//
Well, unfortunately this is a bit of can of worms if you know how certain manufacturers determine these unfortunately
I was just talking about what the impulse response (or step response) is?
Obviously you also have to look at distortion and IMD.
It surprises or rather confuses me from some people with a certain background to be honest.
The way step/impulse responses work and how to mathematically transform these to other things, is something most engineers learn at control theory courses?
I totally understand that most people forget the math behind it (I also don't remember), but just the higher level principles are mostly important.
Anyway if you're still unsure, just check the mechanical equivalent circuit of a loudspeaker.
It immediately shows how and why the Mms and Le interact.
Last edited:
Basically yes.
Le matters for the linear HF response of a driver as it attenuates the response. For a woofer, where the response is usually LF limited, Le will not be much of a factor (although its value needs to be taken into account in a passive Xover.)
This is wrong.
The IR completely defines the linear response to any input signal, steady state or otherwise. Nonlinear response is not so easy, but for all practical purposes drivers are linear. It takes a serious nonlinearity to affect the linear response to any appreciable degree. For example it would take a level of THD of >50% to affect the linear response by a mear 3 dB. Typical values of <10% have far less than a 1 dB effect on the linear response.
Linearity does not depend on context.
Good loudspeakers are basically linear devices - certainly linear enough so as nonlinearity is not an audible factor. So what if nothing is perfect, if it is good enough.
Strange response, since in your own response you say that linearity depends on context.
"basically", "enough" "not an audible factor" and "So what if nothing is perfect, if it is good enough." are all additions and words to say that within a certain context, something can be assumed to be linear.
But that doesn't mean the system is linear by itself.
The nuance I wanted to point out with my previous response, is that it's important to understand what context we are talking about.
But the real elephant in the room here, is that it's unclear for many people what "good enough is".
@camplo
Unfortunately I couldn't find the right graphs for woofers atm, but here are some visual examples what things like Mms, BL and Le do with the frequency response for compression drivers.
Although they are not exactly the same as woofers, the same effects can be expected.
If you really want, I can find the ones for woofers as well, but it has been a while so I have to dig deep.
Unfortunately I couldn't find the right graphs for woofers atm, but here are some visual examples what things like Mms, BL and Le do with the frequency response for compression drivers.
Although they are not exactly the same as woofers, the same effects can be expected.
If you really want, I can find the ones for woofers as well, but it has been a while so I have to dig deep.
D.B.Keele considered an effect of BL factor on woofer efficiency
Comparison of Direct-Radiator Loudspeaker System Nominal Power Efficiency vs. True Efficiency with High-Bl Drivers
Comparison of Direct-Radiator Loudspeaker System Nominal Power Efficiency vs. True Efficiency with High-Bl Drivers
Link doesn't seem to work, maybe this one works?
https://www.xlrtechs.com/dbkeele.com/PDF/Keele (2003-10 AES Preprint) - Nom vs True Eff High BL.pdf
otherwise it's attached below
https://www.xlrtechs.com/dbkeele.com/PDF/Keele (2003-10 AES Preprint) - Nom vs True Eff High BL.pdf
otherwise it's attached below
System efficiency or rather usability is an entirely different thing.
An higher BL also shifts up the frequency response, as can be seen in the paper as well.
Meaning that for the same acoustic response, we have to boost more in the low-end.
So eventually we will end up in a circle of compromises between acoustic volume, acoustic power, electrical power and the amount we have to EQ. (VAS also has a tiny bit of importance in here)
For a very specific custom project for a customer (can't disclose anything unfortunately), we have been investigating a custom subwoofer.
We ended up lowering the BL actually, because otherwise a boost of more than 6dB was needed to get a certain target response.
Which can be troublesome for either the amplifier or pre-amplifier stages/active filter.
Green = higher BL
Yellow = lower BL
Except for BL, all other parameters are the same.
EQ curve:
Freq resp:
This is the power graph (VA curve), as you can see, the amount of extra power is only a little but more;
Even for people with big mamma woofers, this is important.
Unless you just buy the biggest amplifier you can buy with a pre-stage that has plenty of headroom.
But otherwise you really run the risk of clipping and/or wasting power for nothing.
An higher BL also shifts up the frequency response, as can be seen in the paper as well.
Meaning that for the same acoustic response, we have to boost more in the low-end.
So eventually we will end up in a circle of compromises between acoustic volume, acoustic power, electrical power and the amount we have to EQ. (VAS also has a tiny bit of importance in here)
For a very specific custom project for a customer (can't disclose anything unfortunately), we have been investigating a custom subwoofer.
We ended up lowering the BL actually, because otherwise a boost of more than 6dB was needed to get a certain target response.
Which can be troublesome for either the amplifier or pre-amplifier stages/active filter.
Green = higher BL
Yellow = lower BL
Except for BL, all other parameters are the same.
EQ curve:
Freq resp:
This is the power graph (VA curve), as you can see, the amount of extra power is only a little but more;
Even for people with big mamma woofers, this is important.
Unless you just buy the biggest amplifier you can buy with a pre-stage that has plenty of headroom.
But otherwise you really run the risk of clipping and/or wasting power for nothing.
Last edited:
Exactly, and therefore, one can't trust the FR-> Impulse calculation that REW et. al. does.
//