The sonic benefits of an active crossover. A discussion.

DonH50

Member Sponsor & WBF Technical Expert
Jun 23, 2010
3,539
4
38
Monument, CO
Thanks, Don. Here is the seminal question for a non-techie like me. If I am working with a former speaker designer who has worked extensively on Wilson speakers over the last 15 years...and he is 'highly confident' that he can use the Wilson Active Crossover to bypass the passive bass network (ie, handle the technical issues of the crossover frequency and roll-off but also the phase issue you refer to above)
that it would not surprise you if connecting a great amp (perhaps a second Gryphon) directly to the bass drivers should result in a superior sound than using the current passive crossover network. Fair interpretation of your posts? If not, please correct me...only trying to understand, not to put words in anyone's mouth. Thank you for helping!
There is no way to be 100% sure, but my answer would be yes. The biggest unknown is how much difference there would be; my experience has been dramatic to almost none. IMO!
 

DonH50

Member Sponsor & WBF Technical Expert
Jun 23, 2010
3,539
4
38
Monument, CO
@opus111: My answer in a nutshell: The product of two linear systems can produce a nonlinear result.

I have measured distortion due to back-EMF. It was not a listening test and I am not sure I have ever actually performed a listening test that isolates distortion due to back-EMF. Nor do I recall the actual measurement data (I have notes but it was several decades ago). As you have said it is highly dependent upon the driving impedance. v = L * di/dt (and thus i is the integral of voltage over time divided by L) so there is a relationship between voltage and current in the coil.

As I said, in ADCs and DACs (themselves nonlinear products) the charge kick at the input/output (essentially the same impact as back-EMF from voice coils though much different basis) is a huge problem.
 

LL21

Active Member
Dec 26, 2010
10,572
3
38
There is no way to be 100% sure, but my answer would be yes. The biggest unknown is how much difference there would be; my experience has been dramatic to almost none. IMO!
Thanks, Don! as the saying goes, YMMV!
 

Groucho

New Member
Aug 18, 2012
685
0
0
UK
@Don and opus

So it would only be back-EMF that caused nonlinearity when driving a cone via a series impedance? There's nothing about the mechanics of the system that means that 'loss of control' of the cone gives rise to distortion? I guess it's simply a case of a signal driving a resonant system, so there are no fundamental reasons why it should distort; it will merely give rise to phase shift/attenuation over certain frequency ranges.
 
Last edited:
Feb 11, 2012
1,286
0
0
Hangzhou, China
No - in my understanding back EMF is perfectly linear. You only get distortion because of the inductance modulation of the voice coil as it physically traverses the gap. A shorting ring which reduces voice coil inductance will also reduce the non-linearity. So no, no fundamental reasons at all why adding a series impedance would cause any distortion other than by this effect.

@DonH - how can you be sure it was the back EMF which caused your distortion? I reckon it was inductance modulation. To verify you could replace the drive unit with its electrical equivalent (L,C and R model) - if the back EMF was responsible then the distortion wouldn't change because the electrical equivalent gives precisely the same back EMF. My bet is it'd go away provided the inductor's core material wasn't being driven into non-linearity.
 

DonH50

Member Sponsor & WBF Technical Expert
Jun 23, 2010
3,539
4
38
Monument, CO
@Groucho: I would not characterize the speaker system as a resonant system though I suppose you could do so (it does have impedance peaks and valleys so does exhibit resonances). There are many factors that can cause distortion, including cone modulation (flexing of the cone itself, sometimes called "breakup" -- B&W had a nice demo of that many years ago, and it was discussed in my grad acoustics class) and modulation of the driving voltage due to finite driving impedance. In speakers, "loss of control" often refers to the inability of the amplifier to keep the speaker (driver) from exactly following the input signal. The amplifier's linearity, its output impedance, the impedance of any wire and the crossover, driver characteristics (electrical and mechanical, e.g. voice coil and magnet electrical characteristics, cone flex and momentum, etc.) all have their say in how well the speaker reproduces the input signal. All to say that if the amplifier was a perfect voltage source connected directly to the speaker, you would still see distortion from the speaker itself, but lower than if there was a lot of impedance between amp and speaker. In the tech forum I posted a thread on the impact of amplifier impedance on frequency response. I did not attempt to perform time-domain analysis and the model did not include sufficient detail to show all these factors but it may be interesting to see how FR is impacted by amplifier impedance given somewhat realistic speaker models.

There is also "distortion" due to different signals arriving at different times due to differences in driver placement, size, crossover alignment, or whatever. The signals may be linear leaving the cones, but if they are not time-aligned then pulse integrity (the "attack" and "decay") is lost even if a spectrum analyzer does not show additional spurs. How much this really matters is, surprise, a matter of some debate.

@opus111: I would not distinguish between "inductance modulation" and back-emf caused by the inductance of the voice coil in response to cone movement. You drive the voice coil, it reacts, and mechanical systems being what they are there is lag and hysteresis that modulates the signal given a finite (non-zero) driving impedance. This leads to nonlinearity (distortion). Transformers and RF mixers have similar effects and they all produce distortion. The signal produced in response to movement is linear, true (under your assumptions), but the effect of a linear signal modulating another linear signal still produces a nonlinear result in this case. It is similar in the derivation to an RF mixer or for example clock modulating the input signal of an ADC. Linear signals go in, but the output is nonlinear. A*B is not a linear operation.

My analogy is the charge kick-back from an ADC input, a nasty problem for high-speed circuit design such as I have done in the past. The charge kick from the input devices switching modulates the input signal and leads to distortion that appears at the output. I model the charge kick from a speaker (voice coil) the same way, except it is lower in amplitude and slew and audio amplifiers have lower output imepdance so it has less impact. Plus we are less sensitive to distortion; 1% distortion in a speaker will go unnoticed, but -40 dBc spurs in many a radar system would be a disaster.

I suppose we could go on debating but I'm tired; you can have the last word.
 

Groucho

New Member
Aug 18, 2012
685
0
0
UK
Very very interesting. So it may be how I thought it might be: that the cone has mechanical motion/hysteresis/resistance/memory of its own that is 'mixed' with the voice coil's force, so that if the driving signal is less than vice-like, some genuine nonlinear distortion may occur. It would be really interesting to have a ballpark estimate for what that distortion level might be...
 

DonH50

Member Sponsor & WBF Technical Expert
Jun 23, 2010
3,539
4
38
Monument, CO
The voice coil and magnet structure together are part of the whole "back-emf" thing. I made measurements decades ago and flat-out do not remember what the results were. It is highly dependent upon the driving source; I used a big amp (I think it was a Krell or ML class A monoblock) and stepped in resistors to increase its output impedance. I think if it was big I would have remembered... One of the interesting things I do remember is that super-high damping is less critical than expected. At that time (late 70's) high damping factors, like several hundred to 1000+, were all the rage. My study was actually focused on that, not strictly the impact of speaker cones causing back-EMF via the voice coil (although both relate to amplifier output impedance). IIRC for most speakers once you get over about 20 or so it no longer matters, at least for hearing tests, and by maybe 50 even measurements were in the mud. That would imply it is a non-issue for any modern amp, and probably inaudible even for many tube amps. I am guessing it was well below 1%, and with speakers typically generating 1% to 10% or more the whole damping factor and back-EMF issues were in the noise.
 
Feb 11, 2012
1,286
0
0
Hangzhou, China
Very very interesting. So it may be how I thought it might be: that the cone has mechanical motion/hysteresis/resistance/memory of its own that is 'mixed' with the voice coil's force, so that if the driving signal is less than vice-like, some genuine nonlinear distortion may occur. It would be really interesting to have a ballpark estimate for what that distortion level might be...
Attached a paper by Wolfgang Klippel which covers this in a fair amount of detail - not for the faint-hearted though :)

View attachment Loudspeaker_Nonlinearities–Causes_Parameters_Symptoms_01.pdf
 

Groucho

New Member
Aug 18, 2012
685
0
0
UK
Attached a paper by Wolfgang Klippel which covers this in a fair amount of detail - not for the faint-hearted though :)
Thanks for that. Very comprehensive!
 

Duke LeJeune

[Industry Expert]/Member Sponsor
Jul 22, 2013
149
5
18
Princeton, Texas
The purpose [of comparing active and passive versions of the same speaker] is so blatantly obvious. Compare them side by side to see what the actual audible differences are. Plain and simple. Geddes has done this with his speakers. His conclusion was the difference in sound quality between the active and passive versions wasn't worth the added cost of an active version.
Earl was commissioned to build an active version of his Summas, and did so. He compared that version side-by-side with his passive version and what he told me was that there was no audible difference, and so he would never build an active version again.

As an aside, while Earl is definitely an objectivist, he may be the most intense listener I have ever met. He has a friend who is a solo piano recording artist. Earl listened to and critiqued the sonics of one of his friend's albums, and his friend was so impressed with Earl's ability to hear little problems that everyone else had missed that he said he would never put out another album without Earl listening to and critiquing it first.

Regarding passive vs active, as a speaker designer, there are things I can do actively (and especially with DSP) that I can't do passively, and that is very attractive.

Over in the studio speaker world, actives totally dominate the mixing monitor end of the spectrum. Those are the smaller speakers you see on top of consoles (or, more recently, firing over a bank of computer monitors). That sucks for me because I have an idea for a somewhat unorthodox mixing monitor whose radiation pattern would work really well for that application, but I'm not an amplifier technician and don't want to manufacture something that I can't fix.

At the other end of the spectrum, big mastering monitors, passives outnumber actives, though not by a huge margin. I spent some time on a prosound site reading threads on the active vs passive debate, as far as mastering monitors go. Here's what I came away with: There is relatively little sound quality difference between top-flight implementations of the two, but better amps are available if you go passive, unless of course you go active with separate amps, which is the most expensive route. Also, if an amp goes down in a passive system, you can throw in a substitute and keep on working; if an amp or crossover module goes down inside an active speaker, a critical part of your studio is shut down until it gets repaired, unless you have a spare active speaker on hand (which many studios do, just because the money they stand to lose if their mastering capability is down for a week or more could be crippling).

There is a third application, less well known to the outside world, that of "tracking monitor". Those are the speakers that the engineer listens to while the musicians are laying down a track on the other side of the big glass window. And then when the musicians come back into the control room and say, "dude, how'd we sound??", the engineer plays the raw take back for them over the tracking monitors. Superb accuracy isn't required, but superb dynamics are, because they want to hear the same mojo they had going on a few minutes ago. And the engineer really wants to impress his clients - the musicians - so they need to sound really good even if they aren't objectively super-accurate. Anyway, passives dominate the dedicated tracking monitor market.

Imo tracking monitors and main monitors are the applications that most closely approach high-end home audio; mixing monitors have a highly specialized task that emphasizes articulation over other things that also matter a lot in home audio. Not that this precludes actives for home audio by any means!

I was a dealer for ATC for many years, and was sorely tempted to become a Klein & Hummel dealer at one point, but the price of entry was too steep for me at the time.
 
Last edited:
Jul 1, 2010
8,713
0
0
Earl was commissioned to build an active version of his Summas, and did so. He compared that version side-by-side with his passive version and what he told me was that there was no audible difference, and so he would never build an active version again.

As an aside, while Earl is definitely an objectivist, he may be the most intense listener I have ever met. He has a friend who is a solo piano recording artist. Earl listened to and critiqued the sonics of one of his friend's albums, and his friend was so impressed with Earl's ability to hear little problems that everyone else had missed that he said he would never put out another album without Earl listening to and critiquing it first.

Regarding passive vs active, as a speaker designer, there are things I can do actively (and especially with DSP) that I can't do passively, and that is very attractive.

Over in the studio speaker world, actives totally dominate the mixing monitor end of the spectrum. Those are the smaller speakers you see on top of consoles (or, more recently, firing over a bank of computer monitors). That sucks for me because I have an idea for a somewhat unorthodox mixing monitor whose radiation pattern would work really well for that application, but I'm not an amplifier technician and don't want to manufacture something that I can't fix.

At the other end of the spectrum, big mastering monitors, passives outnumber actives, though not by a huge margin. I spent some time on a prosound site reading threads on the active vs passive debate, as far as mastering monitors go. Here's what I came away with: There is relatively little sound quality difference between top-flight implementations of the two, but better amps are available if you go passive, unless of course you go active with separate amps, which is the most expensive route. Also, if an amp goes down in a passive system, you can throw in a substitute and keep on working; if an amp or crossover module goes down inside an active speaker, a critical part of your studio is shut down until it gets repaired, unless you have a spare active speaker on hand (which many studios do, just because the money they stand to lose if their mastering capability is down for a week or more could be crippling).

There is a third application, less well known to the outside world, that of "tracking monitor". Those are the speakers that the engineer listens to while the musicians are laying down a track on the other side of the big glass window. And then when the musicians come back into the control room and say, "dude, how'd we sound??", the engineer plays the raw take back for them over the tracking monitors. Superb accuracy isn't required, but superb dynamics are, because they want to hear the same mojo they had going on a few minutes ago. And the engineer really wants to impress his clients - the musicians - so they need to sound really good even if they aren't objectively super-accurate. Anyway, passives dominate the dedicated tracking monitor market.

Imo tracking monitors and main monitors are the applications that most closely approach high-end home audio; mixing monitors have a highly specialized task that emphasizes articulation over other things that also matter a lot in home audio. Not that this precludes actives for home audio by any means!

I was a dealer for ATC for many years, and was sorely tempted to become a Klein & Hummel dealer at one point, but the price of entry was too steep for me at the time.
I would put it differently - mixing monitors emphasize precision, clarity,revelation. It is the part of the process in which you must be able to hear everything. The other parts? Sounding more like typical home audio is more important. What are the other things that also matter in home audio that you refer to above?
 

Groucho

New Member
Aug 18, 2012
685
0
0
UK
Robh3606 said:
The purpose [of comparing active and passive versions of the same speaker] is so blatantly obvious. Compare them side by side to see what the actual audible differences are. Plain and simple. Geddes has done this with his speakers. His conclusion was the difference in sound quality between the active and passive versions wasn't worth the added cost of an active version.
Earl was commissioned to build an active version of his Summas, and did so. He compared that version side-by-side with his passive version and what he told me was that there was no audible difference, and so he would never build an active version again.
Have you looked at the Geddes web site recently?:
The Abbey 12c is the latest model with a new crossover, and an active crossover option.
Baffle + drivers + crossover (active or passive)
Abbey 900 1025 1250 1400 1800 2700 3000
Nathan 400 525 725 800 (850 Active) 1000 1400 1600
Harper 350 450 600 675 (800 Active) 800 1100 1200
"Crossover" contains all electronic parts or a preprogrammed DSP box for active filter usage
When comparing active and passive versions of the same speaker, is the aim to duplicate the passive crossover actively, or to take advantage of the possibilities for steeper slopes etc. that the active version offers? Analogue active, or DSP active? Time alignment of drivers? I would still expect the line level duplicate of the passive network to sound different from the passive version especially when driven loud.

I recently read this review of a DSP active speaker:
http://www.hificritic.com/downloads/digital/HIFICRITIC_MeridianDSP7200.pdf

and found the reviewer's comments on the sound uncannily like some comments I had made concerning DSP active speakers in general. He really does hear a specific, identifiable difference between passive and active speakers:
Right away certain specific characteristics were evident. It’s unquestionably ‘active’, with the grip, near effortless dynamic range, convincing integrity and authority that is typical of the breed. The stereo image was simply excellent, in depth width and focus. There was no aural confusion here, as it sounded almost effortlessly clear, with crisp stable imaging of believable width, coupled with stable off-axis placement where obviously phase displaced content so dictates.

The bass is unusually good, as powerful at very low frequencies as the two hard working 8.5in (216mm) bass units could supply, still more powerful at somewhat higher bass frequencies, and with a gracefully controlled overload character allowing it to be driven harder than one has any right to expect. The bass clearly sounded ‘different’, even compared with very large and extended low frequency alternatives. Something about the 7200 got closer to the truth, with tighter control, better tune playing, and an ability to differentiate confidently between percussive and sustained bass sounds. Tracks
combining both at once can tend to blur into one sound, but not so with the 7200.

Overall it sounds essentially neutral, if marginally rich and comfortable, giving a slightly distant effect that caresses rather than assaults the ears, even when playing very loud.
The above comments are exactly how I feel about the sound of DSP actives. "marginally rich and comfortable, giving a slightly distant effect that caresses rather than assaults the ears, even when playing very loud" strikes a chord with me - and that was not was I was expecting when I first experimented with this stuff. Overall I was expecting a cool, thin sound rather than "rich and comfortable". The above reviewer implies that most passive speakers "assault the ears" when playing loud, and this also rings true to me.

All of this stuff will be measurable, but we then run into the question of what it is that people prefer, rather than what appears to be objectively best. There are comments aplenty that suggest that people like the way that a single amp and passive network 'unifies' the drivers, so that amplifier distortion caused by driving the bass then permeates into the mid and treble, for example (although they wouldn't put it that way, I'm sure). And what music are they listening to? Has their musical taste already withered and ossified around jazz and girl-and-guitar to suit valve amps and full range horn speakers, so that they can't actually perceive the advantages of the active version anyway?
 
Jul 1, 2010
8,713
0
0
Right away certain specific characteristics were evident. It’s unquestionably ‘active’, with the grip, near effortless dynamic range, convincing integrity and authority that is typical of the breed. The stereo image was simply excellent, in depth width and focus. There was no aural confusion here, as it sounded almost effortlessly clear, with crisp stable imaging of believable width, coupled with stable off-axis placement where obviously phase displaced content so dictates.

The bass is unusually good, as powerful at very low frequencies as the two hard working 8.5in (216mm) bass units could supply, still more powerful at somewhat higher bass frequencies, and with a gracefully controlled overload character allowing it to be driven harder than one has any right to expect. The bass clearly sounded ‘different’, even compared with very large and extended low frequency alternatives. Something about the 7200 got closer to the truth, with tighter control, better tune playing, and an ability to differentiate confidently between percussive and sustained bass sounds. Tracks
combining both at once can tend to blur into one sound, but not so with the 7200.

Overall it sounds essentially neutral, if marginally rich and comfortable, giving a slightly distant effect that caresses rather than assaults the ears, even when playing very loud.
This. Exactly this. And when amps are designed for, or carefully matched to individual drivers and built into the speakers instead of being built to handle anything you might wire them to, into a box that is the amp's most expensive component, the active solution should be very efficient, delivering much more bang for the buck than any passive system approaching its performance level. If its not, back away slowly, with one hand on your wallet.

Tim
 
Jul 1, 2010
8,713
0
0
I thought you were asking me what mixing monitors don't do that matters for home audio. If you were asking about something else, obviously I missed it.
No, I got it, but if the monitors you've heard lack low-end extension and impact, large-scale dynamics, sense of immersion or envelopment, wide sweet spot, you need to hear bigger monitors, or smaller ones paired with subs, in a live room. If you heard unnatural timbre, they just weren't very good.

Tim
 

About us

  • Founded in 2010 What's Best Forum invites intelligent and courteous people of all interests and backgrounds to describe and discuss the best of everything. From beginners to life-long hobbyists to industry professionals we enjoy learning about new things and meeting new people and participating in spirited debates.

Quick Navigation

User Menu

Steve Williams
Site Founder | Site Owner | Administrator
Ron Resnick
Site Co-Owner | Administrator
Julian (The Fixer)
Website Build | Marketing Managersing