Hi mates, I heard top designers promoting the idea of "less is more", and sometimes the idea of "no crossover is the best crossover".... and other top designers advocating very complex crossover designs. 🙂
Some say that crossovers (the electronic components) can affect the sound in a way which makes it stray from the original source signal (of course excluding the intended purpose of crossovers which are HPF, LPF, etc). This might be, but how, and by what degree ?
DSP might sound like a great alternative but again the source signal has to be converted to digital and the digital signal has to be converted back to analogue. AD (Analogue to digital) and DA(Digital to analogue) conversions can also distort the sound if the clocking is not right(precise enough) and other elements can distort the sound as well. 😕 Are you aware of other negatives which DSP cross overs might have ?
Based on your experience, tests and honesty, do you think that electronic crossovers do actually deteriorate the fidelity of a system ?
If so, how exactly, what kind of distortions does it provoke? and those distortions are caused by which electronic components/group of electronic components?
Thank you very much in advance
Some say that crossovers (the electronic components) can affect the sound in a way which makes it stray from the original source signal (of course excluding the intended purpose of crossovers which are HPF, LPF, etc). This might be, but how, and by what degree ?
DSP might sound like a great alternative but again the source signal has to be converted to digital and the digital signal has to be converted back to analogue. AD (Analogue to digital) and DA(Digital to analogue) conversions can also distort the sound if the clocking is not right(precise enough) and other elements can distort the sound as well. 😕 Are you aware of other negatives which DSP cross overs might have ?
Based on your experience, tests and honesty, do you think that electronic crossovers do actually deteriorate the fidelity of a system ?
If so, how exactly, what kind of distortions does it provoke? and those distortions are caused by which electronic components/group of electronic components?
Thank you very much in advance
Last edited:
Your questions as simple as they may seem to be are quite complex to explain. There are some distortions that can be deadly sonicly simply by bad choices made in the design process, crossover distortion being one.
There are times when a zobel filter (impedence smoothing) is required and no active crossover can address that. Not that you would want to use one with an active setup.
Active even with a lowly miniDSP 2x4 can sound exceptionally good, in the right hands.
Analogue can have as good or better sound if done properly. The cost difference between the two can easily push one to DSP based digital active for it's flexibility.
First order networks work well with drivers that can fit into that design goal, problem is finding them and the cost they typically run.
Popularity of F.A.S.T setup where one uses a full range down to <200 typically 100 and crosses to a supplimental sub for the bottom end takes care of many problems as far as crossover issues are concerned. The compromise with this is reduced dispersion up high, many times being laser like = head in a vise narrow sweet spot directly on axis.
There are compromises to every design, take your pick of what is and isnt an issue for your design goals, build from there.
My system (under construction) uses 2 miniDSP 2x4's with 8 100w channels. Each speaker is a quasi 4 way, MTMMMWW. 6-6.5"midbasses and an AMT. Dual MLTL subs built in.
Playback of Chesky, Spanish Harlem, Rebeca Pigeon is heavenly, clean and clear, nothing bad or or slightly glaring. Even when played back in mono from one speaker. Depth of sound stage is incredible with If I could sing your blues. Each clearly defined. Actually everything on Chesky Demonstration disk is superb. Can I hear the difference between an mp3 vbr/ ogg vorbis @320k and say and ape or flac, certainly. But even some of the ages old 128k mp3 I have sound good enough to listen to, just not critically 😉
There are times when a zobel filter (impedence smoothing) is required and no active crossover can address that. Not that you would want to use one with an active setup.
Active even with a lowly miniDSP 2x4 can sound exceptionally good, in the right hands.
Analogue can have as good or better sound if done properly. The cost difference between the two can easily push one to DSP based digital active for it's flexibility.
First order networks work well with drivers that can fit into that design goal, problem is finding them and the cost they typically run.
Popularity of F.A.S.T setup where one uses a full range down to <200 typically 100 and crosses to a supplimental sub for the bottom end takes care of many problems as far as crossover issues are concerned. The compromise with this is reduced dispersion up high, many times being laser like = head in a vise narrow sweet spot directly on axis.
There are compromises to every design, take your pick of what is and isnt an issue for your design goals, build from there.
My system (under construction) uses 2 miniDSP 2x4's with 8 100w channels. Each speaker is a quasi 4 way, MTMMMWW. 6-6.5"midbasses and an AMT. Dual MLTL subs built in.
Playback of Chesky, Spanish Harlem, Rebeca Pigeon is heavenly, clean and clear, nothing bad or or slightly glaring. Even when played back in mono from one speaker. Depth of sound stage is incredible with If I could sing your blues. Each clearly defined. Actually everything on Chesky Demonstration disk is superb. Can I hear the difference between an mp3 vbr/ ogg vorbis @320k and say and ape or flac, certainly. But even some of the ages old 128k mp3 I have sound good enough to listen to, just not critically 😉
Do crossovers(passive/active) deteriorate the fidelity of the source sound?
You know what we know results in a crappy speaker? Very nonlinear frequency response. So there goes the "minimal/no crossover" argument right there, because much of it is based on nonsense that some magic x matters more than FR.
That said, sometimes quite good response can come with a minimal crossover. Such cases almost always happen because all the drive units were bespoke to that specific product, eg KEF Q100. As a DIYer, don't bet on it.
You know what we know results in a crappy speaker? Very nonlinear frequency response. So there goes the "minimal/no crossover" argument right there, because much of it is based on nonsense that some magic x matters more than FR.
That said, sometimes quite good response can come with a minimal crossover. Such cases almost always happen because all the drive units were bespoke to that specific product, eg KEF Q100. As a DIYer, don't bet on it.
Hi,
Very short answer is no. They are nearly always necessary.
They can be done non-optimally, they work, but not well.
rgds, sreten.
Very short answer is no. They are nearly always necessary.
They can be done non-optimally, they work, but not well.
rgds, sreten.
The main issue with crossovers is not on the electrical side of things, it's simply the fact that most speakers with crossovers have center to centrer spacing greater than 1/4 wavelength in the crossover region, and thus you will get vertical off-axis lobing error.
Now with odd-order/phase-quadrature crossovers, this lobing error can be significant and uneven, and very room/distance/angle-dependant in terms of what the listener gets.at the seating position. With even-order/in-phase crossovers, this lobing error is symmetrical and never results in peaks, but results in a dip in sound-power response.
Now coaxials or synergy horns and many low crossover points don't have these issues, so it doesn't really matter there. Like I said, it's an acoustical issue, not an electrical one.
The only other issue is transducer selection. You have to pick the right parts to cross over before you do so. Trying to fit a square peg in a round hole... is what a lot of "designers" like to think they're very good at.
Back to the electrical part of things:
Adding an electrical filter to a system changes its phase response. But that's not inherently bad. Minimum phase is tied into frequency response - the flatter your driver's frequency response, the better the minimum phase looks too. Guess what, you add an electrical filter to that, and you're actually IMPROVING its phase response in terms of changing it. That''s assuming you properly corrected improved the frequency response abberation.
Now to address your concerns:
Yes, less parts is more $ in your pocket.
Going ::too:: far into either extreme could degrade measured acoustic response. The goal is always "whatever gets you the desired ACOUSTIC response. Not of the individual drivers, but of the system as a whole, measured both on and off-axis (The Spin-o-rama). Of course, these measurements need HIGH resolution to be useful - no more than 1/20th octave smoothing.
A degree much smaller than the dozens of times the source material has gone through other electronic components before it was ever pressed into physical media.
Those issues you mention are insignificant in my opinion. The REAL issues with DSP crossovers are that you need to
1) Avoid digital clipping of the sound due to filter-implementation.
2) Use high enough bit resolution that losses in bit resolution from digital attenuation are not significant (this, by the way, is the reason why mastering systems use 24-bit resolution even though 16-bit is adequate for playback).
3) Avoid analog clipping of the input/output.
I think that good crossovers ameliorate the fidelity of the system more than pretty much every other aspect of system design.
Now with odd-order/phase-quadrature crossovers, this lobing error can be significant and uneven, and very room/distance/angle-dependant in terms of what the listener gets.at the seating position. With even-order/in-phase crossovers, this lobing error is symmetrical and never results in peaks, but results in a dip in sound-power response.
Now coaxials or synergy horns and many low crossover points don't have these issues, so it doesn't really matter there. Like I said, it's an acoustical issue, not an electrical one.
The only other issue is transducer selection. You have to pick the right parts to cross over before you do so. Trying to fit a square peg in a round hole... is what a lot of "designers" like to think they're very good at.
Back to the electrical part of things:
Adding an electrical filter to a system changes its phase response. But that's not inherently bad. Minimum phase is tied into frequency response - the flatter your driver's frequency response, the better the minimum phase looks too. Guess what, you add an electrical filter to that, and you're actually IMPROVING its phase response in terms of changing it. That''s assuming you properly corrected improved the frequency response abberation.
Now to address your concerns:
Yes, less parts is more $ in your pocket.
Going ::too:: far into either extreme could degrade measured acoustic response. The goal is always "whatever gets you the desired ACOUSTIC response. Not of the individual drivers, but of the system as a whole, measured both on and off-axis (The Spin-o-rama). Of course, these measurements need HIGH resolution to be useful - no more than 1/20th octave smoothing.
A degree much smaller than the dozens of times the source material has gone through other electronic components before it was ever pressed into physical media.
Those issues you mention are insignificant in my opinion. The REAL issues with DSP crossovers are that you need to
1) Avoid digital clipping of the sound due to filter-implementation.
2) Use high enough bit resolution that losses in bit resolution from digital attenuation are not significant (this, by the way, is the reason why mastering systems use 24-bit resolution even though 16-bit is adequate for playback).
3) Avoid analog clipping of the input/output.
I think that good crossovers ameliorate the fidelity of the system more than pretty much every other aspect of system design.
Last edited:
Depends on the execution of the XO and the system is in.
I turfed a Marchand XO (upgraded) because it was introducing a veil into the system. The Ace Audio i have is better but still does. The Beringer DCS2496s i've had listened to are unusable if you use analog in (haven't tried digital in), but there versatility (and a mic) were usefuil is design exporation.
The most transparent i've used are PLLXO, but are limited and need to be matched with carefully considered speakers/amps. Am op-amp buffered PLLXO we made had fewer sonic issues than the full active XOs above but had other issues. Yet to try are PLLXO with Pass B1 style buffers (there is a whole thread on those)
With, even the very best loudspeakers maybe at 10% of where we can take them to, there are so many choices in what compromises can be made, there are no bests, just sets of compromises that suit you best.
dave
- Status
- Not open for further replies.