"Fast" and "slow" subwoofers: can we put them to bed?

[twelti]

My subwoofers are all "slow". I know this because I usually can catch up to mine in a dead heat, with favorable winds.

But seriously, I do see this subject crop up all the time. I'm sure it has been discussed before in different threads, but I'm feeling ambitious today and thinking maybe I could finally put the idea to bed. Or, maybe someone can convince me the idea has merit, in which case I will have to eat crow .

The crux of my argument is that the impulse or step response of a subwoofer, or any linear, time invariant system, is completely described by the magnitude and phase response. If anyone disagrees with that, I can only suggest some reading of any competent text on the subject. The following points should then be considered:

1. Any two subwoofers with the same magnitude and phase response will have the same impulse or step response, and the same "fastness". It doesn't matter if one is larger than the other, or one has a lead diaphragm and one has a carbon nanotube diaphragm.

2. Subwoofers are generally minimum phase devices, which means that the phase response is directly linked to the magnitude response. If you know one you know the other. So, really, all we need to look at is the magnitude response, leading to point 3:

3. We have much more experience interpreting the magnitude response vis a vis subjective qualities, than phase or impulse responses. I know there are some who feel more comfortable looking at impulse responses, but to me, frequency response plots are more intuitive. Aside from that, there is a body of scientific research relating magnitude, and to a lesser extent phase, to subjective impressions. There is much less such research using time domain data. Floyd Toole and Sean Olive, who are responsible for some of the existing research on audibility of resonances, were able to determine thresholds for their audibility using magnitude plots, based on gain, Q, and frequency of each resonance. However they could only speculate about how to predict audibility of the resonances using only the time domain view of the data. The data is there, but not easy to see.

4. The crux of the matter seems to be the cutoff frequency and alignment of the subwoofer at the low end. We have paramaters to describe the variations very handily in the frequency domain (Fo, Q), but in the time domain, what do we have? Rise time (or time constant) is the only thing I can think of, and I have a hard time relating that to what I hear. I know that a long time constant will smear transients, but that is about all. I can suppose that two impulse or step responses could have the same time constant, but resulting form two different resonances, which would be easy to see in the frequency domain, but harder to see in the time domain.

5. If you want to throw non-linear behaviour into the mix, all bets are off, but I don't think that is the issue here.

I'm sure there is something here that someone will take issue with, so, fire away...

 

[Steve williams]

Nicely written Todd. I must admit that I have been one of those guys to use the term "fast" for subs. Thanks for making it clear.

 

[amirm]

Nicely done Todd.

Question: wouldn't the decay time in a room subjectively be thought to impact the bass response as "slow or fast?" If the decay time is long enough, would it not have the same impact as reducing the impact of the next beat? You mention this in #4 but I don't understand how that then supports the fact that time domain is not as revealing. Or are you saying that for the sub itself in an anechoic chamber that is not a good way to study it?

 

[microstrip]

Focal Grande Utopia EM has a very interesting large area woofer speaker - in order to get a high magnetic field in the gap they adopted an Electro-Magnet for their 16” (40cm) woofer. The speakers have a separate DC power supply to polarize it and the user can adjust the current to tailor its response. I have read in a review that adjusting this current changes the efficiency of the loudspeaker, changing the frequency response, but also changes its the damping and the type of bass - more "slow" or more "fast" and that this action is quite different from what you can get using an equalizer.

 

[DonH50]

If we are only talking steady-state, I tend to agree it's a moot point. My experience is related to what Amir said, and in fact deals with what is generally considered non-linearity: the "ringing of the woofer". I have no way to prove this with documents now, but in listening and measuring tests in the 80's, I was trying to determine how the new "active" sub designs compared to the standard (then) "passive" designs. It was around then I first heard people talk about the "speed" of a subwoofer, and folk discussed various issues with the attacks. With the permission and help of the store owner, we set up a bunch of tests. In waveshaping, there is the initial attack, decay, sustain, and release. What we found was that the decay was equally important, if not more so, in determining how a (sub)woofer sounded. Here are the salient points as I recall them:

1. The initial attack was heavily influenced by how the subwoofer integrated into the system. Proper phase matching with the mains to ensure a coherent wavefront (or, cleanest/sharpest pulse edge/impulse response) was the dominant factor here, after getting enough of an amp to do the job.

2. Ringing after the initial attack would "smear" or "muddy" the sound. In some systems, the ringing was actually quite large, and would continue for a number of cycles. The effect heard was a loss of transient attack, although what was really happening was the decay was messed up by all the ringing. The active designs were generally much better at controlling this decay ringing.

All IMO, FWIWFM, my 0.000001 cents - Don

 

[FrantzM]

Hi

Count me as one of those who doesn't espouse this view of "fast" and "slow" subs. I do however think that subs sound different. There is an audible difference between the type of alignments.. Ported sounds different from sealed and bandpass are even more different at least that is I think I've perceived.

 

[twelti]

Nicely done Todd.

Question: wouldn't the decay time in a room subjectively be thought to impact the bass response as "slow or fast?" If the decay time is long enough, would it not have the same impact as reducing the impact of the next beat? You mention this in #4 but I don't understand how that then supports the fact that time domain is not as revealing. Or are you saying that for the sub itself in an anechoic chamber that is not a good way to study it?

Thanks for the question. You brought up a point that perhaps I should have included in my original post. The decay time of the room is much longer and perhaps a better candidate for"fast" or "slow" response than the subwoofer. In any case, I'm not saying that a subwoofer (or room for that matter) cannot have a fast or slow response. My main point is that "fast" or "slow" response is not a new and unique characteristic of subwoofers. It is included in, and more easily characterized by the frequency response, IMHO.

 

[twelti]

Hi

Count me as one of those who doesn't espouse this view of "fast" and "slow" subs. I do however think that subs sound different. There is an audible difference between the type of alignments.. Ported sounds different from sealed and bandpass are even more different at least that is I think I've perceived.

No argument there. They do sound different. I wonder though, how many times they are judged to sound different when auditioned is some room, with the sub located at some particular location in the room, with the listener in some particular other location in the room... giving perhaps a different room transfer function...? just a thought.

 

[twelti]

Focal Grande Utopia EM has a very interesting large area woofer speaker - in order to get a high magnetic field in the gap they adopted an Electro-Magnet for their 16” (40cm) woofer. The speakers have a separate DC power supply to polarize it and the user can adjust the current to tailor its response. I have read in a review that adjusting this current changes the efficiency of the loudspeaker, changing the frequency response, but also changes its the damping and the type of bass - more "slow" or more "fast" and that this action is quite different from what you can get using an equalizer.

Hmmm. Interesting. It seems I have heard of this before. So, if you increase the current, you increase the Bl, and the efficiency goes up. Sounds fishy on some level, but I'm not a transducer designer.

My only thought is, that if the system remains linear, the effect of varying the current should be reproducable by some conbination of liner filter and gain. Of course, at higher levels, subwoofer motors do become non-linear, so this discussion could get complicated in that case.

 

[twelti]

If we are only talking steady-state, I tend to agree it's a moot point. My experience is related to what Amir said, and in fact deals with what is generally considered non-linearity: the "ringing of the woofer". I have no way to prove this with documents now, but in listening and measuring tests in the 80's, I was trying to determine how the new "active" sub designs compared to the standard (then) "passive" designs. It was around then I first heard people talk about the "speed" of a subwoofer, and folk discussed various issues with the attacks. With the permission and help of the store owner, we set up a bunch of tests. In waveshaping, there is the initial attack, decay, sustain, and release. What we found was that the decay was equally important, if not more so, in determining how a (sub)woofer sounded. Here are the salient points as I recall them:

1. The initial attack was heavily influenced by how the subwoofer integrated into the system. Proper phase matching with the mains to ensure a coherent wavefront (or, cleanest/sharpest pulse edge/impulse response) was the dominant factor here, after getting enough of an amp to do the job.

2. Ringing after the initial attack would "smear" or "muddy" the sound. In some systems, the ringing was actually quite large, and would continue for a number of cycles. The effect heard was a loss of transient attack, although what was really happening was the decay was messed up by all the ringing. The active designs were generally much better at controlling this decay ringing.

All IMO, FWIWFM, my 0.000001 cents - Don

Very interesting, and thanks for commenting. One thing I would differ on is that the ringing is "non-linear". Ringing is a property of linear systems. To avoid any confusion, let me be clear that when I say "non-linear" I dont' mean non-linear phase, I mean non-linear in the sense of not distorting. These are not the same.

What do you mean by "active designs". Do you mean powered subs, or servo subwoofers.

I certianly would not discount your observations. It would be interesting and illuminating to look at all the data from a magnitude/phase point of view. For example, if the sub isn't integrating well, I would expect to see that the level and/or phase is not summing up consistantly over the crossover region. That is pretty easy to spot.

 

[amirm]

Todd, what if lack of deep bass is associated with the perception of "fast bass?" Wouldn't higher frequency bass have the sensation of tighter bass? If I had different size drums and pounded on each, wouldn't the smaller one sound tighter and hence "faster?"

In that sense, the frequency response would still be revealing of this fact but perhaps not in the way we interpret it.

 

[twelti]

Todd, what if lack of deep bass is associated with the perception of "fast bass?" Wouldn't higher frequency bass have the sensation of tighter bass? If I had different size drums and pounded on each, wouldn't the smaller one sound tighter and hence "faster?"

In that sense, the frequency response would still be revealing of this fact but perhaps not in the way we interpret it.

Hi Amir, well, your very good questions make me want to answer with another question. What IS "fast" or "slow" bass? Do we even agree on that? I'm not sure, becuase I haven't really thoguht about bass in those terms.

Your intuition is that higher frequency bass would be associated with "fast" bass, because higher frequency bass is itself "faster" (shorter period)... right? So, if we move the resonant frequency of the sub higher, (all other things being equal) the time constant of the ringing is shorter, thus faster bass. So, if we have a sub that only goes down to 100 Hz, it will be very very "fast" ?!? Is this a good thing? Also, don't forget that the time constant of ringing is equally dependent on slope of the high pass response. So if we hold the cutoff frequency the same and reduce the slope (say we switch to a sealed box) we get more low frequency energy, but the system rings less ("faster"). This seems counter to your intuition.

 

[amirm]

I am with you Todd on the definition problem . I personally never look at whether a sub is fast or slow. I am just thinking out loud what perceptual factor may lead to someone saying that. Wonder if we could have poll a few people believing in this characterization and then correlate that with the frequency response curve. If my guess is right, we may find that the less low frequency extension, the less the descriptor "fast" would be used. Alas, such a test would have to be in a chamber as room effects would make this a hard test to run.

 

[microstrip]

And how should we consider the influence of the amplifier in the "fast" - "slow" bass? Considering equal frequency responses, only the damping of the amplifier should affect its sound. But from my experience building a large 15" subwoofer with an high quality speaker, there was a large difference in bass performance between amplifiers with similar damping factors.

 

[twelti]

And how should we consider the influence of the amplifier in the "fast" - "slow" bass? Considering equal frequency responses, only the damping of the amplifier should affect its sound. But from my experience building a large 15" subwoofer with an high quality speaker, there was a large difference in bass performance between amplifiers with similar damping factors.

Well, if the amplifier damping factor is small (as it normall is) it shouldn't really be a factor. If it were not small, the effect should be reflected in the frequency response.

 

[DonH50]

Very interesting, and thanks for commenting. One thing I would differ on is that the ringing is "non-linear". Ringing is a property of linear systems. To avoid any confusion, let me be clear that when I say "non-linear" I dont' mean non-linear phase, I mean non-linear in the sense of not distorting. These are not the same.

What do you mean by "active designs". Do you mean powered subs, or servo subwoofers.

I certianly would not discount your observations. It would be interesting and illuminating to look at all the data from a magnitude/phase point of view. For example, if the sub isn't integrating well, I would expect to see that the level and/or phase is not summing up consistantly over the crossover region. That is pretty easy to spot.

I was not clear enough, sorry... When I said "considered non-linearity" I meant "considered incorrectly as a non-linearity". That is, I have run into many people who think "ringing" is non-linear. It can be, of course, depending on the circuit (too deep for this thread -- ringing in a large-signal circuit that contains a voltage-dependent capacitance (e.g. transistor junction or depletion C) would be an example, and I suppose you can argue the ringing is linear but modulated by the non-linear circuit elements), but in this case it is a distortion that is certainly not non-linear.

I think of distortion as nonlinear operation, but in this case I would assert that ringing, by adding undesirable signal content to the original signal, can be considered distortion. Another way of saying it, more in line with my background, is that if I measure a circuit with ringing on a spectrum analyzer, the ringing adds an undesirable spur. Linear, yes, but still a bad thing.

Active design = servo subs to me, i.e. with a control (PID, FB and/or FF, whatever) loop. I think of a powered sub as simply moving the amp block along the same line in the signal flow graph...

Great, I managed to mess up twice in one post...

I agree you can clearly see if the phase is not continuous through the crossover region, and for that matter should be able to see the ringing frequency as a peak on the Bode plot.

I also just noticed this is your Expert Forum, my apologies. I usually click on the links in "What's New" and did not notice I was trespassing. I try to not butt in, but it's an endemic problem I have spent years honing, sorry. - Don

p.s. There is much more to consider than DF in amplifiers... Charge storage, for one thing, and large-signal response to back-emf. For that matter, DF should be considered a complex quantity, and not measured just by an RMS meter with a resistive load. I have measured many amps with the same DF spec that did not have the same DF on a VNA. But, I've yapped enough.

 

[twelti]

Hi Don, is there such a thing as "tresspassing"? Anyway your comments are welcome. Yes, there is confusion about linearity. For me, being an EE, it is perhaps simpler, as there is a mathematical definition which is pretty unambiguous. non-mathematically, you could say that the input/output plot of a linear system is a straight line that goes through zero. So, normal resonances meet this definition, and even systems with excess phase in their response meet this definition.

Say you have a resonance and you drive it a frequency near the resonant frequency. The steady state response will still only have the same frequency component as the driving frequency, so linear.

Some of the other things you mentioned I'm not too sure about! In any case, I'm just trying to limit the discussion to the linear stuff, at first anyway, until we get bored with that...

 

[microstrip]

(...) p.s. There is much more to consider than DF in amplifiers... Charge storage, for one thing, and large-signal response to back-emf. For that matter, DF should be considered a complex quantity, and not measured just by an RMS meter with a resistive load. I have measured many amps with the same DF spec that did not have the same DF on a VNA. But, I've yapped enough.

Don,
This is too interesting to be forgotten. Could you educate us on what you mean by charge storage?
Or perhaps you could start one of your excellent threads just about the damping factor!

 

[twelti]

Don,
This is too interesting to be forgotten. Could you educate us on what you mean by charge storage?
Or perhaps you could start one of your excellent threads just about the damping factor!

Separate thead would probably be better...

 

[Mark Seaton]

Hi Don, is there such a thing as "tresspassing"? Anyway your comments are welcome. Yes, there is confusion about linearity. For me, being an EE, it is perhaps simpler, as there is a mathematical definition which is pretty unambiguous. non-mathematically, you could say that the input/output plot of a linear system is a straight line that goes through zero. So, normal resonances meet this definition, and even systems with excess phase in their response meet this definition.

Say you have a resonance and you drive it a frequency near the resonant frequency. The steady state response will still only have the same frequency component as the driving frequency, so linear.

Some of the other things you mentioned I'm not too sure about! In any case, I'm just trying to limit the discussion to the linear stuff, at first anyway, until we get bored with that...

Hi Todd,

I commend the effort.

The biggest problem I see lies in the connotative definitions of the descriptors of "fast bass" and the resulting extrapolations so many make which are non-causal, but ARE coincidental.

I regularly push to change the vocabulary to more specific descriptions, as with further probing, I've seen many use the "fast/slow" descriptor for a wide range of subjective observations. The most common causes of such observations can be tracked directly to the room's transfer function at the listening position, along with integration issues with the main speakers, and even differences in the ground plane behavior of the subwoofer. More often than not so many audiophile hypotheses stem from likely coincidences with common room transfer functions and measurement blind crossover implementations.

The ultimate questions come back to answering the perceptions of why so many perceive an 18" woofer as "slow" and a 10" as "fast", or a ported 12" vs. sealed 12" may have similar perceptions.

The reality is that it is possible to create examples that clearly break these perceptions, but they are not what is commonly available for many obvious reasons.

Maybe this will throw a little spark to the thread again.