• Welcome to ASR. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Speaker "Speed"

Back to your “speed” definition if you have a small vs large driver and both have the same bandwidth, in the same enclosure type, do you think they have the same rise time from 0 to max?

I am not sure what you are talking about. The shape of the initial part of the step response is governed by the minimum-phase characteristics of the driver. Low frequencies = slow rise time. High frequencies = fast rise time. Place a high-pass filter and you can increase the rise time.

I would be interested to read that Linkwitz article that you mentioned?
 
"Speed" is a vague term. Do you mean acceleration or velocity?

^100%^
Usually people referring to slow or flabby etc. are usually talking about transcient response.
We don’t know if the OP means that, or not.
We should have them say what they mean.

"Speed" is a vague term. Do you mean acceleration or velocity?

Anyway, let's work it out from first principles. Let's define some terms:

- Displacement: how far the cone moves (meters)
- Velocity: how fast the cone moves (meters/sec)
- Acceleration: how fast the velocity changes (meters/sec^2)

For a given speaker driver producing a constant sine wave tone at a constant SPL:

- Frequency = ω/2π (ω = angular frequency in radians). Or: ω = 2πf. It is simpler to think of ω as frequency.
- Displacement follows this formula: x(t) = A * sin (ωt), where x(t) is the displacement of the cone at that moment of time, A is the maximum amplitude of displacement. Or put in plain English: the displacement of the cone equals the sine of the frequency (expressed in angular frequency) at that moment in time.
- Velocity is 90deg ahead of displacement (e.g. when the cone reaches the limit of its excursion, velocity is zero). So we take the cos function and multiply it by the angular frequency: v(t) = A * cos (ωt) * ω
- Acceleration is 180deg ahead of displacement (e.g. when displacement is at maximum, acceleration is at maximum but in the opposite direction). So we are back to a sine function. But to obtain acceleration, we have to multiply displacement by the square of the angular frequency. So a(t) = A * sin (ωt) * ω^2

If you want to ignore all that mathematics stuff, here is a simpler explanation:

- Frequency = 1 over time (1/s)
- Displacement is how far the cone moves (m)
- Velocity (m/s) is displacement multiplied by frequency
- Acceleration (m/s^2) is displacement multiplied by the square of frequency

Remember that SPL = the swept volume of the driver. Meaning, cone area * maximum displacement. So for a large driver to produce the same SPL as a smaller driver, it doesn't have to move as much (displacement is less). Looking at those equations again, if A is less, then so is maximum velocity and acceleration.

So yes, larger drivers are "slower" than small drivers if they are asked to produce the same frequency at the same SPL.

People talk about XMAX like it is the main thing to know, and I suppose that it is a straying to push the
Push the coil out of motor is not really a great idea.

In reality the SPL acronym has he “P” for “Pressure”.
There is no direct relationship for SPL to Position, Distance, or Velocity… it is all the 2nd derivative.
So we can sort of ignore position.

In the extreme, if we went to DC, we would see the motor providing a constant acceleration to “the cone”, which clearly have some limits.
(But there are rotary drivers)

In the operating range of a subwoofer, the pressure is
The Sound Pressure Level (SPL) can be related to acceleration and area through the formula: SPL = 20 log10(p/p0), where p is the sound pressure in pascals, which can be derived from the acceleration (a) and area (A) using the relationship p = ρaA, with ρ being the density of the medium. However, the direct SPL equation typically does not use acceleration and area explicitly

SPL = 20 log10(p/p0),
Or
SPL = 20 log10(ρaA/p0),

If we double the area (A) then we get to reduce the acceleration (a) by half.
Every motor is non-linear to some extent, but as the cone get larger and larger, then the total throw required gets to be less and less, and it is easier to stay in a more linear region as the throw approached zero displacement.


So yes, larger drivers are "slower" than small drivers if they are asked to produce the same frequency at the same SPL.

Yeah no.
More massive drivers are only slower at the same throw or velocity, using the same power/force.


Using a steady-state sine wave hides the transient behavior, that’s why the unit step signal or tone burst is used instead.

With a steady state sine-wave you can’t observe damping or any oscillatory behavior, even at resonance its still a sine-wave SPL, albeit larger.

If we use the example of a 0-60mph vehicle test, at T=0 the drive slams his/her foot on the accelerator, commanding the ECU for maximum torque (or in the old days wide open throttle) - in effect this is a step signal, and the response is the time it takes for the vehicle speed to reach a predefined amount… 60) or a stead state maximum (aka top speed). But now imagine the driver at T=0 only partially opens the throttle, then a second later, a bit wider and so on, a bit like your sine-wave example.

A shaped tone burst is also a good way of seeing the finite acceleration in the first half of burst cycle and indeed the overshoot when the burst is shut off. Siegfried Linkwitz provides a good explanation of how shaped tone burst are a great tool for examining transient response, and distortion too.

The Kicker data sheet has both step and tone burst performance.

Back to your “speed” definition if you have a small vs large driver and both have the same bandwidth, in the same enclosure type, do you think they have the same rise time from 0 to max?

In ^this example^, it may be better to think of the “speaker driver” as what happens at the instant when the throttle is mashed.
We don’t need the car or the cone to get up to 60 MPH, the pressure is instantaneous.
So it is more like “at the instant” when the throttle is matched the front if the car lifts up immediately. It doesn;t start moving and then the front end lift and the hind end squats down. It lift and squats immediately.

And on the cone, the pressure is instantaneous with the acceleration.[/QUOTE][/QUOTE]
 
^100%^
Usually people referring to slow or flabby etc. are usually talking about transcient response.
We don’t know if the OP means that, or not.
We should have them say what they mean.



People talk about XMAX like it is the main thing to know, and I suppose that it is a straying to push the
Push the coil out of motor is not really a great idea.

In reality the SPL acronym has he “P” for “Pressure”.
There is no direct relationship for SPL to Position, Distance, or Velocity… it is all the 2nd derivative.
So we can sort of ignore position.

In the extreme, if we went to DC, we would see the motor providing a constant acceleration to “the cone”, which clearly have some limits.
(But there are rotary drivers)

In the operating range of a subwoofer, the pressure is


SPL = 20 log10(p/p0),
Or
SPL = 20 log10(ρaA/p0),

If we double the area (A) then we get to reduce the acceleration (a) by half.
Every motor is non-linear to some extent, but as the cone get larger and larger, then the total throw required gets to be less and less, and it is easier to stay in a more linear region as the throw approached zero displacement.



Yeah no.
More massive drivers are only slower at the same throw or velocity, using the same power/force.




In ^this example^, it may be better to think of the “speaker driver” as what happens at the instant when the throttle is mashed.
We don’t need the car or the cone to get up to 60 MPH, the pressure is instantaneous.
So it is more like “at the instant” when the throttle is matched the front if the car lifts up immediately. It doesn;t start moving and then the front end lift and the hind end squats down. It lift and squats immediately.

And on the cone, the pressure is instantaneous with the acceleration.
[/QUOTE]
[/QUOTE]
Well there is no practical air load on the cone, so what I was referring to is the position of the cone vs time, clearly not instantaneous.
 
I don’t think that can ignore the physics though.

Could we draw the SPL as a sine wave, in one color and plot the position of the cone in another colour?
Maybe we are not thinking of the same thing. Here is a useful graph from matlab concerning the response of a system to a step input. Consider this to be the cone, because after the inital motion the SPL doesn’t following this because when the cone reaches its steady state position, there is no further motion and thus no spl, but its the first part of the response where the dynamics of the cone, suspension, and diaphragm loading by the air in the enclosure control the rise time are important. So this is what I am talking about, Tr = rise time. BW~0.35/Tr. Standard systems theory, nothing new.

I don’t have much more I can add on this topic. I feel some of the posts are mixing poorly understood theory with practical engineering design and we might get caught in an endless debate about “speed” whatever that means.

So I’ll duck out on this one. Thanks for the exchanges.

1778885372901.png
 
Maybe we are not thinking of the same thing. Here is a useful graph from matlab concerning the response of a system to a step input. Consider this to be the cone, because after the inital motion the SPL doesn’t following this because when the cone reaches its steady state position,

No - the SPL goes to zero when the acceleration goes to zero.
The main point I was making is that SPL has absolutely nothing to do with position.



I am not sure I understand the plot:
- Is the MATLAB model have voltage going out endlessly?
- Why it is stopping at Y-Final? (Is the motor force being opposed by the suspension forces?)

1) It starts at Y=-1 - versus having started at Y=0 ??
2) When I twist the iPad around it looks like the “The rise” is pretty much a linear line until about T=2.
- I would have expected more of a curve staying closer to -1 for a while as the inertia was holding it back, and then exponentially start going up.
 
Last edited:
No - the SPL goes to zero when the acceleration goes to zero.
The main point I was making is that SPL has absolutely nothing to do with position.




I am not sure I understand the plot:
- Is the MATLAB model have voltage going out endlessly?
- Why it is stopping at Y-Final? (Is the motor force being opposed by the suspension forces?)

1) It starts at Y=-1 - versus having started at Y=0 ??
2) When I twist the iPad around it looks like the “The rise” is pretty much a linear line until about T=2.
- I would have expected more of a curve staying closer to -1 for a while as the inertia was holding it back, and then exponentially start going up.
The cone steady state position is fixed (its a DC current) so no acceleration, so of course the SPL goes to zero!

We are in agreement that SPL has nothing to do with position, I did not sy that it did.

Maybe for an alternative on rise time, just google step response and rise time and read Wikipedia
 

We are in agreement that SPL has nothing to do with position, I did not say that it did.

I didn’t say you did, but it is surprising how many want to draw the SPL curve on top of the position of the driver plot.
And it is sort of easy to arrive at that with the Steady-state approach.


The cone steady state position is fixed (its a DC current) so no acceleration, so of course the SPL goes to zero!

We are in agreement that SPL has nothing to do with position, I did not say that it did.

Maybe for an alternative on rise time, just google step response and rise time and read Wikipedia

I think we are largely in agreement.
Especially with the transcient response.

The OP could look at a variety of subwoofers of various sizes and the step function responses.
And they could do ^all that^ for sealed and ported arraignments.

The bottom line is that at some defined SPL the larger cone doesn’t have to move as much, or as fast, or as quickly… as the smaller cone.
So it either comes out as a wash, or the larger is cone is actually “faster” because the cone mass scales with area, whereas the coil is more of a constant added to it.

And all of the subwoofer drivers are an order of magnitude faster than the crossover, or some DSP bandwidth, requires them to be anyhow.
So the “speed” is mostly all about the box.

Do you have any links for the step function response for actual drivers in boxes? Etc.?
 
Using a steady-state sine wave hides the transient behavior
Not true if the system is close enough to linear and time invariant. You can derive impulse, step, CSD, tone burst, etc. response from steady state magnitude and phase. If the system is known to be minimum phase, you only need magnitude.

For a given speaker driver producing a constant sine wave tone at a constant SPL:
This is incorrect. A direct radiator produces a constant SPL vs frequency when the acceleration is constant[1], not the displacement—
a(t)=A×sin(ωt), v(t)=-A×cos(ωt)/ω, and x(t)=-A×sin(ωt)/ω².

[1] Assuming the diaphragm is sufficiently small compared to the wavelength.
 
This is incorrect. A direct radiator produces a constant SPL vs frequency when the acceleration is constant[1], not the displacement—
a(t)=A×sin(ωt), v(t)=-A×cos(ωt)/ω, and x(t)=-A×sin(ωt)/ω².

[1] Assuming the diaphragm is sufficiently small compared to the wavelength.

Thank you for the correction, but i'm not sure if that is true. I am under the impression that the swept volume determines the SPL. Swept volume = cone area * displacement. So if you have a larger cone area, you need smaller displacement. But since both large and small driver needs to move at the same frequency, the larger driver has to cover a smaller displacement in the same amount of time. Therefore it needs to accelerate less and reach a lower maximum velocity?
 
Not true if the system is close enough to linear and time invariant. You can derive impulse, step, CSD, tone burst, etc. response from steady state magnitude and phase. If the system is known to be minimum phase, you only need magnitude.


This is incorrect. A direct radiator produces a constant SPL vs frequency when the acceleration is constant[1], not the displacement—
a(t)=A×sin(ωt), v(t)=-A×cos(ωt)/ω, and x(t)=-A×sin(ωt)/ω².

[1] Assuming the diaphragm is sufficiently small compared to the wavelength.
Yes^ you can derive those things but that wasn’t my point, you can’t observe transient behavior… by definition its “steady state”
 
I didn’t say you did, but it is surprising how many want to draw the SPL curve on top of the position of the driver plot.
And it is sort of easy to arrive at that with the Steady-state approach.




I think we are largely in agreement.
Especially with the transcient response.

The OP could look at a variety of subwoofers of various sizes and the step function responses.
And they could do ^all that^ for sealed and ported arraignments.

The bottom line is that at some defined SPL the larger cone doesn’t have to move as much, or as fast, or as quickly… as the smaller cone.
So it either comes out as a wash, or the larger is cone is actually “faster” because the cone mass scales with area, whereas the coil is more of a constant added to it.

And all of the subwoofer drivers are an order of magnitude faster than the crossover, or some DSP bandwidth, requires them to be anyhow.
So the “speed” is mostly all about the box.

Do you have any links for the step function response for actual drivers in boxes? Etc.?
I don’t have any to hand, but frankly any good loudspeaker designer will make those measurements, so there must be lots of examples around on ASR too.

I’d recommend reading para 7 on Siegfried’s web page here, he was a very well respected RF Engineer from HP and a person that made contributed in many ways to our understanding of loudspeaker design.

 
i'm not sure if that is true
See this post (case 3 and summary point 4). If displacement is constant, the SPL increases +12dB/octave with frequency.
Your basic conclusion is correct: a larger cone needs less displacement and therefore lower velocity and acceleration to generate a given SPL.
 
See this post (case 3 and summary point 4). If displacement is constant, the SPL increases +12dB/octave with frequency.
Your basic conclusion is correct: a larger cone needs less displacement and therefore lower velocity and acceleration to generate a given SPL.

^“Therefore“ is in the wrong place.^

Velocity is a byproduct of acceleration.
And displacement is a byproduct of the velocity.

It should read as “ a larger cone needs a lower acceleration and therefore a lower velocity, and less displacement, in order to generate a given SPL.”
 
@Waveform Fidelity I showed you how it's posible to stay on the on the side that group delay is minimal even to a sub bass, stil tone burst neads about 30 ms to settle down which ain't a thing to full standing waves under 45 Hz and still very good. What Linkwitz didn't mention and Keith got wrong (high pass with be taken for actual crossover) you cut it with high enough order self low PEQ filter and explained the range where to do it as easy as possible. In order to show everyone how it's posible to get there. You always take compromises in design, ideally is jest ideal.
Did you heard the story about guy who mesured 10 sub's in order to make it's own?
Pay attention to step (impulse) response and how relatively also impacts it especially how long it takes to settle down on tone burst if step is bad (yes it's more relevant to beaming and where it starts to happen higher in FP but when you count in harmonics it's relevant to all). B&W 10" does it best, followed by Neumann (which uses active DSP to achieve all it achieves) and it's own 10" which is in between those two (still quite good and DSP-ed to achieve it). After he whose done he run away and didn't disclose build to DIY. It's about using kicker in your head (brain). You can't expect it in of shelf pasive design and you won't get it in most cases even if it uses active DSP but you can make it relatively simple if you understood what I said.
Hire is another example that is very easy to understand how changing tuning and reliefing burden of what it struggles with impact groop delay and THD of course.
Take a look at music preset and compare it to others (it's reference with a bit of Butterwort extension to it). If you ask me finally one of the manufacturers started using DSP properly.
 
It's not the hole story of course, there is space and refractions ratios to direct sound, their timing difference which adds to it all and also getting detail idea of how human hearing works regarding loudness scale and how delay progressively goes up in sub bass and that there is no woofer in known history that can take SPL increase in order to compensate (offset) for it. Included in ELC ISO 226 2022 and later revisions (after ISO 226 2003 really). That's why we perceive sub bass as a tail lagging in time with rapid intensity fall, most other mamels don't even bother including super predators like family of cats which hear to 40 Hz and use pow cousins to detect seismic changes instead. It's really seismic more than sound as Anna (Lapwood) use to say.
It's posible to use space boundary fundamental for the purpose in controled DSP-ed manner, you just need long enough room (8, 9m) so F is on place to do the job.
 

Did you heard the story about guy who mesured 10 sub's in order to make it's own?
Pay attention to step (impulse) response and how relatively also impacts it especially how long it takes to settle down on tone burst if step is bad …

There is some pretty nice work in there.
And those step function responses are pretty varied.
 
There is some pretty nice work in there.
And those step function responses are pretty varied.

That's because the step response is a function of the minimum-phase nature of the subwoofer! @ZolaIII and @Waveform Fidelity

1778931162338.png


I'll prove it. Here are two simulated subwoofers. Both have corner frequencies of 20Hz and 100Hz. The red one has a minimum-phase Butterworth 4th order bandpass. The green one is Butterworth 2nd order.

1778931342785.png


Now look at the step response. See the steeper ("faster") rise time of the green slope?

Take away: if we examine ONLY the pure minimum-phase response of subwoofers (with no extra funny stuff going on like all-pass filters or additional DSP), the amplitude response determines the step response.
 
PLUS the driver phase which is what you are not showing. Do the same simulation and use LINFIR you can see the effect of the driver phase and crossovers phase and play with these parameters and watch the resulting step response. You can also se how using an FIR filter to flatten the phase response affects the impulse response.
 
Back
Top Bottom