Here, here, well said Mark!
I agree with your points Mark!I've come to realize it is a mistake to consider an entire system minimum phase.
A system is really a set of minimum phase regions stitched together to become non-minimum phase.
Typical xovers are not minimum phase. (IIR xovers )
Nor are any summations of drivers that don't operate as an acoustic point source together. (such as coaxial, triaxial, synergy etc.)
Individual drivers are minimum phase, but once combined, nope...
Ime, it's why adding EQs to a speaker in xover regions seldom works (other than to a narrow specific listening spot)
It's also why different "Q"s at same xover freq will sound different, due to different polars.
OP question "Will they have the different in sound?"
In my experience, yes.
The why is a complex and as yet not fully understood (and rarely discussed) subject which is beyond the scope of a DIY thread, but here are a few discussion points that might be of interest?
(1) From a snapping twig to a 100 piece orchestra in full swing all airborne sound is 100% "APE" ie Air Pressure Event. A simple increase in air pressure (compression) or decrease in air pressure (rarefaction).
(2) The only two variables are the duration of the APE and intensity (level) of the APE.
(3) APE's are defined in and by TIME not frequency, APE's have a clear and measurable duration, they have a start and a stop time ie when the increase in air pressure starts, and how long it takes for the air pressure to decrease and stabilise back to ambient. (The step response and CSD plots indicate how accurate a loudspeaker driver or complete loudspeaker design is in the time domain)
(4) The frequency of a sound is the direct result of the number of APE's per second.
Obviously there is much more to APE's and the time domain in sound than these 4 "starter for 10" points, but its a fascinating subject which is only just beginning to get some wider acceptance.
IMO, the bottom line to the OP question is this:
The driver/loading/crossover/Eq combination which results in the most accurate Time Domain reproduction of the original source (music/voice etc) APE will sound better.
Any delayed resonance caused by drivers bouncing around on their mechanical suspensions after the initial electrical impulse will distort the sound in a fundamental way. This time domain distortion is single worst type of distortion in the audio chain and 99.99% of it comes from the fundamental failing of all our loudspeaker drivers to react to, and then STOP reacting to(!) an electrical impulse.
The solution?
We need a mechanical loudspeaker driver which can react at the speed of electrical components... Now thats a tall order!!
In my experience, yes.
The why is a complex and as yet not fully understood (and rarely discussed) subject which is beyond the scope of a DIY thread, but here are a few discussion points that might be of interest?
(1) From a snapping twig to a 100 piece orchestra in full swing all airborne sound is 100% "APE" ie Air Pressure Event. A simple increase in air pressure (compression) or decrease in air pressure (rarefaction).
(2) The only two variables are the duration of the APE and intensity (level) of the APE.
(3) APE's are defined in and by TIME not frequency, APE's have a clear and measurable duration, they have a start and a stop time ie when the increase in air pressure starts, and how long it takes for the air pressure to decrease and stabilise back to ambient. (The step response and CSD plots indicate how accurate a loudspeaker driver or complete loudspeaker design is in the time domain)
(4) The frequency of a sound is the direct result of the number of APE's per second.
Obviously there is much more to APE's and the time domain in sound than these 4 "starter for 10" points, but its a fascinating subject which is only just beginning to get some wider acceptance.
IMO, the bottom line to the OP question is this:
The driver/loading/crossover/Eq combination which results in the most accurate Time Domain reproduction of the original source (music/voice etc) APE will sound better.
Any delayed resonance caused by drivers bouncing around on their mechanical suspensions after the initial electrical impulse will distort the sound in a fundamental way. This time domain distortion is single worst type of distortion in the audio chain and 99.99% of it comes from the fundamental failing of all our loudspeaker drivers to react to, and then STOP reacting to(!) an electrical impulse.
The solution?
We need a mechanical loudspeaker driver which can react at the speed of electrical components... Now thats a tall order!!
Let us consider the typical two-way design of two speakers. The drivers are identical for speaker A and B. They are the same crossover point used. Speaker A uses Q at 0.707 on both drivers, the other uses Q at 0.49. Both speakers are measured flat frequency response. Will they have the different in sound?
Given how often Troels Gravensen's name comes up on this forum, I'm surprised this project hasn't been mentioned yet. It's a two-way with switchable crossover configurations allowing first, second or fourth-order operation.
The special feature about this speaker is that you can choose between three crossover topologies, 1st order, 2nd order (LR2) and 4th order (LR4) and provides you with a unique opportunity to pick you favourite crossover, maybe one for one kind of music, maybe another for some other musical genre. The Hypex has three pre-sets and all you have to do is press a button to shift between the three presets. If you're in for further convenience you can buy the Hypex remote and sit in your sofa and pick your flavour.
Intriguingly, Troels doesn't comment on the sound differences.
Alex
Last edited:
I agree with your points Mark!
Thanks for kind words, glad my thoughts ring true to you 🙂
Given how often Troels Gravensen's name comes up on this forum, I'm surprised this project hasn't been mentioned yet. It's a two-way with switchable crossover configurations allowing first, second or fourth-order operation.
The special feature about this speaker is that you can choose between three crossover topologies, 1st order, 2nd order (LR2) and 4th order (LR4) and provides you with a unique opportunity to pick you favourite crossover, maybe one for one kind of music, maybe another for some other musical genre. The Hypex has three pre-sets and all you have to do is press a button to shift between the three presets. If you're in for further convenience you can buy the Hypex remote and sit in your sofa and pick your flavour.
Intriguingly, Troels doesn't comment on the sound differences.
Alex
A much cheaper way to try the same thing is PC with Equalizer APO and a multichannel sound card, or even a surround receiver with HDMI as sound card/amp. RePhase could be useful too. Both programs are free.
So the crossover can be used to blend the effect between the drivers. Clearly something that would be useful in certain situations.I think that one reason a lower Q crossover is desired is with typical cone-dome combinations. The dome tweeter comes with it's very wide dispersion whereas the larger cone mid(woofer) is starting to beam with much narrower dispersion. The lower Q crossover creates a lower total output (power response) above the crossover where the tweeter have very wide dispersion. So creates a more balanced power response
This can be a problem but it isn't always.Isn't banding on the vertical axis due to phase cancellations at crossover? If we increase the crossover bandwidth, won't that intensify?
This doesn't happen just by changing to a different order, unless I misunderstand?The difference in group delay between two different drivers covering the same band doesn't cause phase issues? Comb filtering?
Just to help clarify...
If xovers are fully complementary, as xover order increases there will be no difference between the phase or group delay between the signal sent to the two different drivers. That's the definition of fully complementary.
But as xover order increases, there will be a difference in overall phase rotation and group delay when looking at the combined response curve sent to the two drivers, looking at lowest frequency vs highest.
If xovers are fully complementary, as xover order increases there will be no difference between the phase or group delay between the signal sent to the two different drivers. That's the definition of fully complementary.
But as xover order increases, there will be a difference in overall phase rotation and group delay when looking at the combined response curve sent to the two drivers, looking at lowest frequency vs highest.
"We see lots of us using Xsim Like a video game, Creating a flat sloped response, with good phase tracking and a smooth off axis response, 1st,2nd,3rd,4th order variants plus combos. The various blends are subjectively different."
anyone feel like giving their view on how to each can be viewed sonically.
- Home
- Loudspeakers
- Multi-Way
- Different "Q", same crossover point, different sound?