I simply made the point there's more than one way to create a fully active speaker system besides putting it all in one box. No dismissal intended but both approaches have their strong points.
-
WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. 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!
Passive speakers, separate boxes...help me understand the appeal
- Thread starter svart-hvitt
- Start date
- Joined
- Jun 19, 2018
- Messages
- 6,652
- Likes
- 9,408
I just had a read of the Smith paper and didn't find much relevant to the current discussion.
Firstly, @levimax, the definition of transient response you're working with is not the definition used in that paper. The paper is interested in the classical definition of transient response, which encompasses both ringing and group delay. Given that the paper investigates only relatively low-order filters (1st and 2nd order), ringing is negligible, and it is group delay that is responsible for the observed differences between the filters used. I don't think there's enough info in that paper really to draw any conclusions, as the individual responses of the filters/drivers used are only described in broad terms, and were not measured.
Secondly, the filter types used in the passive networks were then-standard Butterworth filters, which have now fallen out of favour due to their inability to correctly sum at the crossover point in the frequency domain. In most modern passive speakers, these Butterworth networks have been replaced by Linkwitz-Riley networks or variations thereof, which do sum correctly in the frequency domain. The active filters used in the study were Crowhurst filters, which are a filter type now long-forgotten but which (from what I can gather from the online sources available) were optimised for transient response and which were technically superior to the Butterworth networks to which they were being compared.
So it seems to me that all the paper demonstrates is that by using active filters, the authors were able to achieve better transient response than with the Buterworth passive networks originally used in the speakers they investigated. This did of course result in better transient response, but since the authors did not make any effort to design their own passive networks, all they are really comparing are the speakers' stock passive networks with the authors' customised active networks.
From this I think the only conclusions that can be drawn are that:
Things have obviously moved forward a long way since that paper, and of course modern DSP-based active filters are capable of far better transient performance than both the passive and active networks used in the paper.
Finally, the authors began from the assumption that the degree of group delay introduced by the passive networks they looked at was audible. Even half a century later, this assumption is yet to be experimentally demonstrated.
There is some more information on the Crowhurst networks the authors investigated here.
In any case, I think that paper is of little interest to the present discussion.
Firstly, @levimax, the definition of transient response you're working with is not the definition used in that paper. The paper is interested in the classical definition of transient response, which encompasses both ringing and group delay. Given that the paper investigates only relatively low-order filters (1st and 2nd order), ringing is negligible, and it is group delay that is responsible for the observed differences between the filters used. I don't think there's enough info in that paper really to draw any conclusions, as the individual responses of the filters/drivers used are only described in broad terms, and were not measured.
Secondly, the filter types used in the passive networks were then-standard Butterworth filters, which have now fallen out of favour due to their inability to correctly sum at the crossover point in the frequency domain. In most modern passive speakers, these Butterworth networks have been replaced by Linkwitz-Riley networks or variations thereof, which do sum correctly in the frequency domain. The active filters used in the study were Crowhurst filters, which are a filter type now long-forgotten but which (from what I can gather from the online sources available) were optimised for transient response and which were technically superior to the Butterworth networks to which they were being compared.
So it seems to me that all the paper demonstrates is that by using active filters, the authors were able to achieve better transient response than with the Buterworth passive networks originally used in the speakers they investigated. This did of course result in better transient response, but since the authors did not make any effort to design their own passive networks, all they are really comparing are the speakers' stock passive networks with the authors' customised active networks.
From this I think the only conclusions that can be drawn are that:
- the stock passive networks were sub-optimal
- it was easier to produce a speaker with superior transient response using the active filter networks chosen by the authors
Things have obviously moved forward a long way since that paper, and of course modern DSP-based active filters are capable of far better transient performance than both the passive and active networks used in the paper.
Finally, the authors began from the assumption that the degree of group delay introduced by the passive networks they looked at was audible. Even half a century later, this assumption is yet to be experimentally demonstrated.
There is some more information on the Crowhurst networks the authors investigated here.
In any case, I think that paper is of little interest to the present discussion.
IME there is a big difference between traditional active speakers, where the crossover is still analogue, and a DSP digital active speaker, where the corrections can be much more accurate and targeted in amplitude and phase an resonance cancellation.
I thought so too until I actually got into building both types. Check out the famous Linkwitzlab web page.... http://www.linkwitzlab.com/filters.htm .... if you are willing to do the work you can get a traditional analog filter to do pretty much anything but it is time consuming and expensive. An alternative to analog filters is something like "rephase" which creates FIR DSP filters in software. As was mentioned earlier an analog filter and a digital version of that same filter act the same... including the "problems" such as group delay and ringing. FIR adds a new wrinkle in that the "cost" of added filter complexity is not "phase shift"as in an analog filter but rather "latency" (time delay). As a software solution DSP is obviously much easier and cheaper and more flexible. While DSP is great it has limitations in what it can accomplish and since any filter added to the signal chain has some performance cost DSP is more of a incremental fine tuning and cost savings (especially for mass market audio) than a break through. For real high end speakers it always comes back to the mechanical issues of good quality drivers and enclosures which limit the amount of filtering required.
The demo I heard actually used the DSP engine to simulate a passive filter as one of the presentations. I don’t recall exactly which parameter pairs the engineer said could not be both corrected simultaneously in an analogue filter that could with DSP, it was a few years ago.
I have been interested by Linkwitz for decades.
Maybe not practical to correct two things simultaneously in a passive analog filter rather than impossible? I wish DSP could make my speakers sound like Salon 2 ..... more likely to make next Salon (3?) sound incrementally better.
Whilst latency is of little import in a purely audio playback system, it is a real problem in live sound or audio for video where the sound has to match the pictures, or in the case of live sound, what's happening on stage. Consequently, the use of FIR isn't universal and therefore the value of DSP-based crossovers over analogue derived filter has to come from other aspects.
In my view, the great benefits of DSP-derived crossovers are the following:-
1) Greater slopes over what's feasible with op-amps and passive components. 48dB/octave is easy with DSP, 24dB/octave is a more practical limit with analogue methods.
2) Very much easier ability to change crossover frequencies and slopes 'on the fly' and put these changes into memory, so one filter box can be used for any number of different loudspeakers.
3) DSP lends itself to parametric and graphic EQ at very little extra cost. This can also include LF room compensation.
4) DSP can be used for time delay compensation, phasing, drive unit protection and in PA, feedback reduction, all much more difficult to do with analogue circuits.
4) Development is much easier given how many libraries exist for DSPs like the Sharc.
For a manufacturer of loudspeakers, there may be a cost benefit from a simple analogue active crossover once all the design work is done using DSP, but if the crossover is at all complex, or other facilities are offered, then DSP becomes far simpler and cheaper to implement.
Just as I can't think of any benefit to passive crossovers over active, I can't think of any benefits to analogue active crossovers over DSP except possibly in the very cheapest and simplest cases.
S.
In my view, the great benefits of DSP-derived crossovers are the following:-
1) Greater slopes over what's feasible with op-amps and passive components. 48dB/octave is easy with DSP, 24dB/octave is a more practical limit with analogue methods.
2) Very much easier ability to change crossover frequencies and slopes 'on the fly' and put these changes into memory, so one filter box can be used for any number of different loudspeakers.
3) DSP lends itself to parametric and graphic EQ at very little extra cost. This can also include LF room compensation.
4) DSP can be used for time delay compensation, phasing, drive unit protection and in PA, feedback reduction, all much more difficult to do with analogue circuits.
4) Development is much easier given how many libraries exist for DSPs like the Sharc.
For a manufacturer of loudspeakers, there may be a cost benefit from a simple analogue active crossover once all the design work is done using DSP, but if the crossover is at all complex, or other facilities are offered, then DSP becomes far simpler and cheaper to implement.
Just as I can't think of any benefit to passive crossovers over active, I can't think of any benefits to analogue active crossovers over DSP except possibly in the very cheapest and simplest cases.
S.
Agree completely. The better the EAC (in Line with DAC ;-)) is engineered , the filter order and complexity can be reduced. Notecthat when i talk about filter order, i mean the acoustical order, not the electrical order.
That's a ridiculous generalization. A worldly-wise and successful speaker designer with a good reputation will tell you that the transition zone within which both drive units are energized is bad news. Even with a theoretically perfect crossover in which the two outputs have identical characteristics (one rising, the other falling), the region in which two (non-identical) drive units are contributing to the output is troublesome and it is the view of the best designers (as evidenced by their results) that it is preferred to get in to and out of this troublesome region as quickly as feasible.
While there might be some speaker designs that make a really good fist of first order crossovers I have yet to hear what I would describe as a ’general purpose’ box speaker in which the first order approach excels. There's a certain amount of common sense in this reasoning: a typical two-way design might cross over at, say, 3kHz. A first order filter rolls off at 6dB/8ve or 20dB/decade. This means that you have to use a tweeter that is at its best at 3kHz (highly critical as it's near the ear's most sensitive region) while not producing too much distortion two octaves below, at around 800Hz. To help appreciate what's required here, 20dB / decade means that your little tweeter which has to be at its best behaviour at 3kHz must still produce a reasonably clean output (albeit at 20dB lower level) at 300Hz!
This is a truly tall order, electromechanically speaking.
Not only non-identical but non-coincident.
Point taken but I said lower order not first order.... In a discussion of active vs passive where very high order filters may be involved lower order to me is 4th or less.... I will be more specific.
Just to be clear I am not arguing passive is better or preferable, rather all things considered especially the added complication and incompatibilities with existing equipment a well executed passive speaker can be a reasonable choice.
- Joined
- Jan 27, 2019
- Messages
- 7,336
- Likes
- 12,302
I'm left unclear on what you refer to as a "general purpose" box speaker, vs whatever other type of box speaker you might have in mind.
I would say that Thiel speakers, *especially* the last Jim Thiel iteration 3.7 using the final version coax drivers, excel in sonic terms, to say the least.
They got uniform rave reviews for a reason. Obviously Jim Thiel was keenly aware of the driver/crossover challenges in first order crossover design so he continued to refine the crossover/motors and finally designed drivers in-house especially to meet those challenges. The Thiel 3.7 used a specially designed coax driver midrange driver that was "flat" up to 20 kHz. I forget the specs for the tweeter but they were impressive too. Not to mention the midrange in that coax design was essentially "flat" shaped as well (though ribbed) so as not to impede tweeter dispersion.
I've owned many speakers of most design types, and I recently completed almost 2 years of speaker auditioning to see if I could find a replacement for the Thiel 3.7s (because, visually, they were a bit big for my room). I couldn't find anything that performed top to bottom as well as the 3.7s. The 3.7 disappeared so well, provided such a huge soundstage, with such image precision, and with no audible resonance or obvious box coloration from top to bottom. And they remain the most coherent box speaker I've ever encountered. I had to end up finding a pair of the sightly smaller Thiel 2.7s using the same coax/design principles to finally let go of the bigger Thiels.
- Joined
- Feb 23, 2016
- Messages
- 20,792
- Likes
- 37,693
I always thought Jim Thiel's speakers would be even better if they had used 24 db/octave crossovers.
- Joined
- Jan 27, 2019
- Messages
- 7,336
- Likes
- 12,302
As you are no doubt aware, that would defeat one of his main design goals of time/phase coherency, of course. So that was never going to happen.
I've never been in a position to verify via the necessary tests the actual audible-or-not contribution of time/phase coherency in the Thiel design.
But Thiel certainly was single-minded in his efforts to perfect that design. Anecdotally, having had at one point the old Thiel CS6, as well as more recently the 3.7 and 2.7, among the salient attributes of their sound was imaging and soundstaging precision - a purported benefit of time/phase coherency. The Thiels just seem to "line up" all the sonic information associated with a sound source in a particularly dense and convincing manner. When I listen to other speakers, even ones that image really well, they always sound a bit more "swimmy" and less organized than the sound coming from the Thiels. Something you notice when you get used to it.
Again, I'm aware of the controversies about Time/Phase coherence, so I don't have a commitment either way. But the Thiels have a certain precision that really seems lacking in most other speakers of my acquaintance, wherever that attribute comes from in the design I'm not sure.
Simple: a standard cuboid-ish speaker that most people would find palatable, most of the time
Sometimes, one gets the impression that a speaker has been designed with an overriding objective to meet a particular design goal at the expense of (nearly) everything else. I'm not going to name what I'm thinking of because I have no desire to upset treasured apple carts.
- Joined
- Jan 27, 2019
- Messages
- 7,336
- Likes
- 12,302
Simple: a standard cuboid-ish speaker that most people would find palatable, most of the time
Sometimes, one gets the impression that a speaker has been designed with an overriding objective to meet a particular design goal at the expense of (nearly) everything else. I'm not going to name what I'm thinking of because I have no desire to upset treasured apple carts.
Er....ok, do some sensitive ASR members own Green Mountain Audio speakers or something? ;-)
Sorry – don't get that. Cultural divide....
GMA, like Thiel and Vandersteen, adhere to the dogma of first-order passive crossovers with physical time-alignment through slanted/stepped baffles. Unlike more recent DSP techniques (especially FIR filtering), this is a laughably imprecise approach. Moreover, in tolerating severe compromises in directivity (already well established by Geddes, Toole, Olive and others to have a major effect on perceived sound) and distortion for audibly-marginal - if any - gains (see Lipshitz and Vanderkooy for their classic paper on audibility of phase shift) in time alignment and linear phase these brands miss the forest for the trees.
Such brands, though not totally pseudoscientific or science denialists, are at least selectively blind. They do three things: they single-mindedly focus on optimising parameters that do not have any evidence that point at major audibility (not for want of trying); in doing so, they severely compromise other parameters long-established to be pertinent and critical to our perception of loudspeakers in-room; moreover, they use antiquated, inefficient and highly-imprecise methods of optimising the parameters they prioritise, where there are deployable technologies that do the same thing with far less compromise.
- Joined
- Jun 10, 2018
- Messages
- 6,249
- Likes
- 9,389
Some people, though not totally pseudo scientific or science denialists, are at least selectively blind. They do three things: they single-mindedly focus on optimizing parameters that do not have any evidence that point anywhere (not for want of trying); in doing so, they severely compromise other parameters long-established to be pertinent and critical to our perception; moreover, they use antiquated, inefficient and highly-imprecise methods of optimizing the parameters they prioritize.
I wouldn't argue with anything written in the two posts above, but to what does the phrase...
It's clear that any speaker design which prioritizes one particular holy grail over everything else is unlikely to succeed overall. Above all, speaker designers must not fall victim to failure to see the wood for the trees.
refer?
It's clear that any speaker design which prioritizes one particular holy grail over everything else is unlikely to succeed overall. Above all, speaker designers must not fall victim to failure to see the wood for the trees.
Similar threads
