The idea came to me when reading an interview with Vandersteen founder.
We all heard of the often advise even from experience designers that the quality of the components "in the signal path" is important. For example the series cap on the tweeter xover ... But based on electrical circuit analysis, any components on the tweeter xover affects the signal, not just the series cap, then why the cap in series with the tweeter is so important and I don't doubt this based on listening either from my own or from experience designers who talk about it.
Let's look at an example. A first order vs. second order xover for the tweeter. For the 2nd order, there is an inductor to ground. Now even if the inductor is NOT in the signal path based on our definition, it shouldn't affect the sound that much right? But it obviously does. To me 1st order seems to have a more transparent treble vs. 2nd order. So the inductor obviously has some affect on the sound. But you could say this is more like 1st vs. 2nd order filter so it's not a valid comparison.
Comparing two 2nd order xover, the one with better cap sounds better but why would the cap has such an overriding affect to the sound and not the inductor?
Another case is when the on the same for example tweeter xover, we have a RLC notch filter. Now designers will tell you that the quality of the notch filter components are not that important, but when you perform the electrical analysis, it may even have a more severe and abrupt affect on the sound in term of amplitude attenuation and phase change vs. the series cap.
So how does one reconcile that the electrical circuit analysis says there are change in signal regardless of whether the component in the signal path or not but the hearing tells something that may be different?
The idea came to me after I read the article from Vandersteen. Vandersteen is known for his 1st order time aligned speaker and who strongly believe in it. He said a experiment was done in which the listeners were asked if he/she could tell the difference between a change in signal phase vs. amplitude. The listeners was not sensitive to amplitude change within a small db or so but very sensitive to even minor change to phase. But he said that well obviously we are all sensitive to amplitude changes when we change the volume knob just a bit, but he said the important distinction is that when the volume changes, it changes the amplitude of the entire audio range, not just a small range in frequency .
So to me when we talk about component "in the signal path", it is the components that affect the entire signal range, whereas notch filter component only affect a small range of frequency. So any component that affects the entire range will have the most affect on the signal. Any component that on affects part of the frequency range, I think will affect the sound too but probably only to a small amount.
Now for example, the 2nd order xover on the tweeter, does the inductor to ground affect the sound since I would think it modifies the freq. response? From personal experience, it seems to me when using 1st order, with just a cap and also a notch filter on the tweeter, the treble sounds a little more transparent at least to my hearing. But the inductor is supposedly not "in the signal path".
Anyhow, that's how I think about it.
We all heard of the often advise even from experience designers that the quality of the components "in the signal path" is important. For example the series cap on the tweeter xover ... But based on electrical circuit analysis, any components on the tweeter xover affects the signal, not just the series cap, then why the cap in series with the tweeter is so important and I don't doubt this based on listening either from my own or from experience designers who talk about it.
Let's look at an example. A first order vs. second order xover for the tweeter. For the 2nd order, there is an inductor to ground. Now even if the inductor is NOT in the signal path based on our definition, it shouldn't affect the sound that much right? But it obviously does. To me 1st order seems to have a more transparent treble vs. 2nd order. So the inductor obviously has some affect on the sound. But you could say this is more like 1st vs. 2nd order filter so it's not a valid comparison.
Comparing two 2nd order xover, the one with better cap sounds better but why would the cap has such an overriding affect to the sound and not the inductor?
Another case is when the on the same for example tweeter xover, we have a RLC notch filter. Now designers will tell you that the quality of the notch filter components are not that important, but when you perform the electrical analysis, it may even have a more severe and abrupt affect on the sound in term of amplitude attenuation and phase change vs. the series cap.
So how does one reconcile that the electrical circuit analysis says there are change in signal regardless of whether the component in the signal path or not but the hearing tells something that may be different?
The idea came to me after I read the article from Vandersteen. Vandersteen is known for his 1st order time aligned speaker and who strongly believe in it. He said a experiment was done in which the listeners were asked if he/she could tell the difference between a change in signal phase vs. amplitude. The listeners was not sensitive to amplitude change within a small db or so but very sensitive to even minor change to phase. But he said that well obviously we are all sensitive to amplitude changes when we change the volume knob just a bit, but he said the important distinction is that when the volume changes, it changes the amplitude of the entire audio range, not just a small range in frequency .
So to me when we talk about component "in the signal path", it is the components that affect the entire signal range, whereas notch filter component only affect a small range of frequency. So any component that affects the entire range will have the most affect on the signal. Any component that on affects part of the frequency range, I think will affect the sound too but probably only to a small amount.
Now for example, the 2nd order xover on the tweeter, does the inductor to ground affect the sound since I would think it modifies the freq. response? From personal experience, it seems to me when using 1st order, with just a cap and also a notch filter on the tweeter, the treble sounds a little more transparent at least to my hearing. But the inductor is supposedly not "in the signal path".
Anyhow, that's how I think about it.
there's a 75 yr old 'app' for that: Bode Sensitivity Analysis
and no reconcilliation' required if 'hearing tells' from memory, after part swapping by your own hand, not Double Blind, level matched
there are higher order XO that can be 'transient perfect'
1st order slopes don't qualify to me as effective XO - the drivers need decade overlaps, and are are still moving with out of band freqs at high amplitude, adding distortion
and no reconcilliation' required if 'hearing tells' from memory, after part swapping by your own hand, not Double Blind, level matched
there are higher order XO that can be 'transient perfect'
1st order slopes don't qualify to me as effective XO - the drivers need decade overlaps, and are are still moving with out of band freqs at high amplitude, adding distortion
I agree that higher xover sounds "cleaner" and probably less distortion but too me the higher the xover order, the dryer and to some degree sounds a bit more sterile.
Although I don't think there is anything magical about 1st order, it seems to me in general the lower the order, going from 4th ... 2nd ... 1st, the sound becomes more vivid, with more present and just that more musical. It may have to do with less abrupt phase change or something to that affect. I don't know.
1st order is not perfect but I like the trade off.
I had some early digital recordings which sounds cold but through the 1st order, it actually sounds nice.
I hate the whole 'signal path' thing as a concept.
Current flows in loops (always) and will take every available loop in the ratio of their conductances so anything forming any part of such a loop is "in the signal path", and that includes every part of a crossover, makes the concept somewhat pointless.
Some things you do to a circuit are WAY more audible then others, not news, but I don't think that reasoning about in or shunting the signal path is a reasonable way of determining what matters.
Regards, Dan.
Current flows in loops (always) and will take every available loop in the ratio of their conductances so anything forming any part of such a loop is "in the signal path", and that includes every part of a crossover, makes the concept somewhat pointless.
Some things you do to a circuit are WAY more audible then others, not news, but I don't think that reasoning about in or shunting the signal path is a reasonable way of determining what matters.
Regards, Dan.
Any component in the crossover that introduces non-linearity , harmonic or amplitude, will make it's presence felt, audible or subjective. Loudspeaker drivers are by no means linear devices whose parameters will change somewhat with excursion, and, are also subject to cone resonances and non linearites in the suspension system as the signal goes through transients and then there are the practical limitations of a driver's X max. to consider.
The most non-linear component used in passive crossovers will be the iron cored inductor and that will only become obvious when the signal current drives the inductor close to its saturation point. Iron cored inductors are best used in low frequency crossovers where harmonic distortion is less apparent to the ear. Air cored inductors, audio quality capacitors and resistors won't introduce any notable harmonic distortion at all regardless of their placement in a passive crossover.
C.M
The most non-linear component used in passive crossovers will be the iron cored inductor and that will only become obvious when the signal current drives the inductor close to its saturation point. Iron cored inductors are best used in low frequency crossovers where harmonic distortion is less apparent to the ear. Air cored inductors, audio quality capacitors and resistors won't introduce any notable harmonic distortion at all regardless of their placement in a passive crossover.
C.M
Vandersteen's report about phase condradicts my own experience. I designed a speaker (detailed elsewhere in these pages) and use a DSP device for my DAC to run the system. I've done tests with phase corrected for an essentially perfect impulse response, vs. a phase response typical of multiway speakers (higher group delay and phase roll near crossover points). The difference is.... maybe(?) I can hear it?
At a local audio club meeting, I took the speakers and DSP setup and demonstrated the difference with and without perfect FIR phase response with a variety of music, most minimally produced. About a third of those attending (maybe 25 people) said they could hear it, another third said they couldn't, the rest were in the "maybe" camp. Zero people reported a dramatic change or big preference for the phase detected version. One of the "yes, I hear it" people prefered the non-corrected version (I suspect that might be from an aversion he had to the idea of digital and not really understanding about the digital hardware not being any different with or without correction!).
So, not saying it's inaudible, but I think the whole first-order-crossover thing is overblown. I also sense that there is a big overlap between people who think simplist-crossover-is-always-better and people-who-haven't-a-real-clue-how-to-design-a-crossover!
At a local audio club meeting, I took the speakers and DSP setup and demonstrated the difference with and without perfect FIR phase response with a variety of music, most minimally produced. About a third of those attending (maybe 25 people) said they could hear it, another third said they couldn't, the rest were in the "maybe" camp. Zero people reported a dramatic change or big preference for the phase detected version. One of the "yes, I hear it" people prefered the non-corrected version (I suspect that might be from an aversion he had to the idea of digital and not really understanding about the digital hardware not being any different with or without correction!).
So, not saying it's inaudible, but I think the whole first-order-crossover thing is overblown. I also sense that there is a big overlap between people who think simplist-crossover-is-always-better and people-who-haven't-a-real-clue-how-to-design-a-crossover!
Again, as I said, I don't think there is anything magical or unique about 1st order xover. But what I said also true is that as you go from higher order such as 4th to 2nd to 1st, there is a progression to the sound from sounding "clean, somewhat dry, and maybe less distortion at least in term of amplitude" to "sounding more musical, more three dimensional" and I think this has been backed up by others.
OK so you say higher order sounds better well then I guess there's nothing I can do about.
Now as far as the speakers ability to "time align" and able to reproduce a perfect step response or square response, I don't know whether it will ultimately affect the sound or not, this part I am not sure.
But to my own listening experience, a first order whether time aligned or not, since you can make a 1st order but still not time aligned, has a musical quality that is not there in the higher order.
OK so you say higher order sounds better well then I guess there's nothing I can do about.
Now as far as the speakers ability to "time align" and able to reproduce a perfect step response or square response, I don't know whether it will ultimately affect the sound or not, this part I am not sure.
But to my own listening experience, a first order whether time aligned or not, since you can make a 1st order but still not time aligned, has a musical quality that is not there in the higher order.
Small aside: 🙂 Vandys have the Most complex /too many parts Crossovers that one can find without some serious looking.
Typically those Suck massive watts Gee how surprising.
Point being that Yes Vandys can be decent sounding (or at least better than others) but Vandersteen..Giving Xover advise seems oddly out of place IMO
Typically those Suck massive watts Gee how surprising.
Point being that Yes Vandys can be decent sounding (or at least better than others) but Vandersteen..Giving Xover advise seems oddly out of place IMO
I rarely use crossovers.
I do for tweeters and then it is usually just a 2u2 cap in series.
The speakers seem to sound fine without.
I do for tweeters and then it is usually just a 2u2 cap in series.
The speakers seem to sound fine without.
It means that you put in use some very limited in band loudspeakers 😛
The speakers, thou, are part of the circuit, as the cables are.
A consideration to make: in series Xovers the speakers are even more important
..or no
😕
The speakers, thou, are part of the circuit, as the cables are.
A consideration to make: in series Xovers the speakers are even more important
..or no
😕
I recall a Berklee College of Music lecturer saying that in testing phase rotation audibility of LR4 ~ 1 kHz XO the (student) test subjects trained up to good discrimination in 4-5 hrs
so yes phase in XO is audible, but commercial success of the many speakers with LR4 or higher suggests that few home listeners or even 'Golden Eared' reviewers practically care that much
so yes phase in XO is audible, but commercial success of the many speakers with LR4 or higher suggests that few home listeners or even 'Golden Eared' reviewers practically care that much
Last edited:
Thanks for sharing - I like it when real tests show that carefully applied Phlebotinum does nothing much.
This is me. If passive crossovers were all that existed, I'd stick to 1st and 2nd order. Higher than that is a dark art.
Q:What is the main goal of a loudspeaker industry?
A: To sell the products.
Q: How will the company achieve its goals?
A: By making up a concept that will likely make the target crowd believe that their story makes sense to them.
Q: Why do they need to make up concepts like that at all?
A:They don't need to do that but they believe and not without reason, that speaking the naked truth will steer the people away from spending money and that contradicts the main purpose.
Q: What can an average audiophile do in order not to make a too grand of a mistake?
A: Listen to advice that speakerdave at least once wrote down in this forum, to make your own facts by experimenting.
Speaker design is nowadays closer to general public by means of free simulators and relatively cheap measurement gear of all kinds that it's painfully obvious what's true and what isn't.
A: To sell the products.
Q: How will the company achieve its goals?
A: By making up a concept that will likely make the target crowd believe that their story makes sense to them.
Q: Why do they need to make up concepts like that at all?
A:They don't need to do that but they believe and not without reason, that speaking the naked truth will steer the people away from spending money and that contradicts the main purpose.
Q: What can an average audiophile do in order not to make a too grand of a mistake?
A: Listen to advice that speakerdave at least once wrote down in this forum, to make your own facts by experimenting.
Speaker design is nowadays closer to general public by means of free simulators and relatively cheap measurement gear of all kinds that it's painfully obvious what's true and what isn't.
My own testing has revealed that the lower the crossover point, the more audible phase distortion is. Move that 1Khz XO to 2kHz, and the ability to discriminate would probably go away. Also, switch from LR4 at 1K to LR2, and the ability to discriminate would likely go away.
Additionally, the ability to discriminate in ABX testing is pretty far removed from having a preference.
My senior design project in college was to fix the transient response of a speaker. Now I don't bother. Directivity is much more important.
To get back to the main theme of this thread, every component in the crossover impacts the phase and/or magnitude of the signal going through the speaker, so they are all important. Choose components that can handle the power you intend to put through them (some parts of a circuit see more power than others). Some attributes of a component's impedance matter more in certain situations. For instance, the ESR of a capacitor in a zobel circuit probably isn't as important as the ESR of a capacitor in a first order XO. Similarly, inductor resistance doesn't need to be super low if you're putting it in series with a resistor anyway, so go ahead and use a smaller gauge there than you would otherwise.
Last edited:
I've found that as I discovered and treated the room acoustics issues in successive trials, finer aspects of loudspeaker performance in room began to emerge and become much more audible. It's like a little continuous improvement loop: upgrade room acoustics first, then upgrade loudspeaker performance, then upgrade room acoustics again...etc.
I've found that notable types of room acoustics are often overlooked when making such comparisons as you have described, Bill. All it takes is a very few early reflections around the loudspeakers--ceiling, floor, front wall, side walls, racks, large loudspeaker cabinets, etc.--to mask the effects of outstanding loudspeaker phase coherence. Or conversely, the loss of LF performance due to placing the loudspeakers away from room boundaries unbalances the entire presentation so that the listener is only aware of the lack of bass issues...etc.
[This includes the listening rooms of notable loudspeaker manufacturers, audiophile magazine "golden ear" reviewers, and independent audio shop listening rooms. Early reflections are usually detectable from pictures of the rooms alone--like those pictures placed on the web to advertise their "in-house capabilities".]
One clue that you mentioned above is that you took your speakers to an audio club meeting (i.e., not your listening room). That indicates to me a significant probability that the room had early reflection issues. Most audiophiles don't pay attention to early reflections I've found, and are much more concerned about how things look rather than how they perform acoustically. If you treat their room acoustically for them, they usually are sitting there with sweat pouring out of them trying to contain their hostility toward you for bringing room treatments...
So unless you brought with you a stack of absorption tiles that you placed in acoustically effective positions willy-nilly around the front of the room (including the ceiling and floor), most audiophiles usually don't pay attention to that sort of in-room acoustic treatments, and generally don't consider that to be acceptable behavior by their "audio acquaintances". 😱
In my experience, every part in a speaker crossover has an effect on sound quality, but some have a greater effect than others. Most of my DIY projects use 12db slopes, and the series cap in the high-pass has the most impact. After that the low-pass inductor, and after that the two shunt elements are about the same. But all of the parts make an important contribution. The parts in a zobel network may be even less critical but they too are important.
- Status
- Not open for further replies.