My guess is that you just don't like that range of frequencies ( when measured "darn flat", as you said ).
Another possibility is the missing frequencies had a high distortion content due to choosing a crossover point that is beyond the "clean" bandwidth of the drivers.
Have you measured the distortion content of your drivers ( in box ) ?
IOW, it's now a guessing game ( triggered by your first post ) that will just about always point back to some form of "design error".
Another possibility is the missing frequencies had a high distortion content due to choosing a crossover point that is beyond the "clean" bandwidth of the drivers.
Have you measured the distortion content of your drivers ( in box ) ?
IOW, it's now a guessing game ( triggered by your first post ) that will just about always point back to some form of "design error".
It makes perfect sense. Certain speakers deliberately introduce a dip at exactly the crossover point to mitigate the 'dispersion gap', where the tweeter's perfect 180 degree dispersion pattern drastically drops to a narrow beam of e.g. a 8" woofer. But you probably won't have such a problem with a 3-way.
There is something unpleasant off-axis at the crossover point, which you "fixed" by deliberately introducing a dip (which is what wiring out of phase essentially amounts to), and which escaped the wrath of your measurement. Narrow bandwidth dips are often ignored by the ear.
There is something unpleasant off-axis at the crossover point, which you "fixed" by deliberately introducing a dip (which is what wiring out of phase essentially amounts to), and which escaped the wrath of your measurement. Narrow bandwidth dips are often ignored by the ear.
Sage12, you have it right, correct acoustic absolute polarity is everything.
Sure the 'phase flipped' midrange measures flat but it don't sound right.
Bose and many other cheap two way speakers run the lows full range and a single series cap to 'bleed in' the highs.
This is well accepted because it sounds 'right' despite measured response errors.
Now that you have 'discovered' the critical importance of acoustic polarity you can further experiment with speaker cable and interconnect cable directions.
Have fun with it.
Dan.
Sure the 'phase flipped' midrange measures flat but it don't sound right.
Bose and many other cheap two way speakers run the lows full range and a single series cap to 'bleed in' the highs.
This is well accepted because it sounds 'right' despite measured response errors.
Now that you have 'discovered' the critical importance of acoustic polarity you can further experiment with speaker cable and interconnect cable directions.
Have fun with it.
Dan.
No scientific evidence supports the notion that absolute phase is audible - meaning, in a blindfolded, true ABX test, no test subject could distinguish between the "correct" polarity versus the polarity flipped.
In any case, the whole argument is a moot one, because all real-world loudspeakers introduce phase changes, and worse still differing phase changes at different frequencies. (In fact, that could very well be the exact reason why absolute polarity is undetectable - because no real-world loudspeaker is "right" at every frequency, meaning, it is possible, in the sense that I will certainly give the benefit of doubt, not because I have any convictions actually believing so, that the phenomenon could very well exist, just that no equipment is able to reproduce it faithfully.)
Simple experiment: mix and match two loudspeakers of different brands or model, pick the highest end best sounding flattest FR you could possibly find, and hear how terrible they sound, because each impacts the phase in a different manner.
The whole phase argument is moot. No evidence indicates absolute phase is detectable, and even if it is, no real world equipment can exploit to that effect.
So, let FR be your guide - but not just on-axis FR. Consider what happens off-axis, too.
In any case, the whole argument is a moot one, because all real-world loudspeakers introduce phase changes, and worse still differing phase changes at different frequencies. (In fact, that could very well be the exact reason why absolute polarity is undetectable - because no real-world loudspeaker is "right" at every frequency, meaning, it is possible, in the sense that I will certainly give the benefit of doubt, not because I have any convictions actually believing so, that the phenomenon could very well exist, just that no equipment is able to reproduce it faithfully.)
Simple experiment: mix and match two loudspeakers of different brands or model, pick the highest end best sounding flattest FR you could possibly find, and hear how terrible they sound, because each impacts the phase in a different manner.
The whole phase argument is moot. No evidence indicates absolute phase is detectable, and even if it is, no real world equipment can exploit to that effect.
So, let FR be your guide - but not just on-axis FR. Consider what happens off-axis, too.
To me it sounds like a irregularity in the amplitude response of the tweeter - call it a knee (like in the attachment but more pronounced) which might be 2db or so higher in amplitude. While depending on woofer output they may sum flat, it will create a sharp sound especially on mezzo-soprano vocals. Most tweeters have rising distortion levels under 2Khz and the knee may invoke a rise in 3rd order distortion. Flipping polarity probably creates a dip there which offsets that little knee.
So...can you post measurements and both driver`s amplitudes?
So...can you post measurements and both driver`s amplitudes?
Attachments
I call BS.
Yes, some/many speakers are phase messy and this will serve to destroy/confuse recorded phase (time) information.
Useless experiment because of loss of L/R coherency.
More fantasy, try saying that to pro-audio guys and they will laugh you out of the venue.
Max Headroom,
True that!
Since the early eighties we ( pro audio ) have consistently wired-up speakers while maintaining absolute polarities.
The market appearance 4-pole ( electronic crossovers ) first using ButterWorth filters ( then LR topologies ) made this possible.
- Adding adjustable delay to these crossovers sealed the deal.
- "Absolute Polarity" very quickly became an adopted "best practice".
- Any company that resisted this trend was severely ostracized ( IME, this proved to be the final nail in Altec's demise ).
This practice was enshrined into the emergent "THX protocol" by Tomlinson Hollman.
As a result of THX, I won't wire-up a speaker without maintaining absolute polarities ( if for pro or my own personal use ).
")
True that!
Since the early eighties we ( pro audio ) have consistently wired-up speakers while maintaining absolute polarities.
The market appearance 4-pole ( electronic crossovers ) first using ButterWorth filters ( then LR topologies ) made this possible.
- Adding adjustable delay to these crossovers sealed the deal.
- "Absolute Polarity" very quickly became an adopted "best practice".
- Any company that resisted this trend was severely ostracized ( IME, this proved to be the final nail in Altec's demise ).
This practice was enshrined into the emergent "THX protocol" by Tomlinson Hollman.
As a result of THX, I won't wire-up a speaker without maintaining absolute polarities ( if for pro or my own personal use ).
")
You all are extrapolating, which is unfair to me.
I said the correct polarity is indistinguishable from the wrong polarity to the ear. I DID NOT say speakers can be wired arbitrarily in a pro audio installation (nor a home theatre installation).
Have you ever seen a pro audio venue with only ONE loudspeaker? If so, you can wire it in reverse polarity and as I claimed, the difference is indistinguishable by the ear. Of course, real installations have MANY speakers, and they have to be wired in phase - either all of them correctly, or all of them incorrectly. To avoid confusion, of course the former is the right way to go.
Why don't you just try? Have someone wire your speakers, both correctly, or both wrongly, without telling you, and you be your own judge whether you can tell which is which. Before doing so, please don't come back and argue on this moot point yet.
The THX mandates and prudent pro audio traditions are to prevent mix-and-match and wiring any speaker out of phase.
I said the correct polarity is indistinguishable from the wrong polarity to the ear. I DID NOT say speakers can be wired arbitrarily in a pro audio installation (nor a home theatre installation).
Have you ever seen a pro audio venue with only ONE loudspeaker? If so, you can wire it in reverse polarity and as I claimed, the difference is indistinguishable by the ear. Of course, real installations have MANY speakers, and they have to be wired in phase - either all of them correctly, or all of them incorrectly. To avoid confusion, of course the former is the right way to go.
Why don't you just try? Have someone wire your speakers, both correctly, or both wrongly, without telling you, and you be your own judge whether you can tell which is which. Before doing so, please don't come back and argue on this moot point yet.
The THX mandates and prudent pro audio traditions are to prevent mix-and-match and wiring any speaker out of phase.
Last edited:
Calling my experiment useless = agreeing with me that it will sound terrible. Why? Let me be more specific and add a keyword to what you correctly said, loss of L/R phase coherency. Which means real-world loudspeakers, even the highest-end, flattest FR mess up the phase, which means even if the ear could hear absolute phase the speakers cannot reproduce absolute phase.
Hi Cyberstudio,
Well i've tried what you purpose and yes i wasn't able to tell which was which. But... i've worked with a violin player which was able to hear the absolute polarity on takes made with his own instrument. Not blind test but it was something which was repeatable as the whole technical team was in disbelief and decided to test this random during the session. Each time he spoted almost instantly.
Spooky.
All i can say is we all hear differently and some may be more prone to hear some things than others.
And yes there was some studio with ONE and only speaker. It was looonnnggg ago but it existed and works from such places wasn't this bad.
Well i've tried what you purpose and yes i wasn't able to tell which was which. But... i've worked with a violin player which was able to hear the absolute polarity on takes made with his own instrument. Not blind test but it was something which was repeatable as the whole technical team was in disbelief and decided to test this random during the session. Each time he spoted almost instantly.
Spooky.
All i can say is we all hear differently and some may be more prone to hear some things than others.
And yes there was some studio with ONE and only speaker. It was looonnnggg ago but it existed and works from such places wasn't this bad.
Maybe, maybe not ( extrapolating and being unfair to you ).
I'm talking about multi-way systems all being wired with a single polarity ( ie; bass, mids, himids, uhf if present all going a single physical direction with a positive or negative electrical pulse ).
Obviously delay and even order filter networks ( based on multiples of 4 ) are typically used to get proper crossover summation.
The above practice obviously contrasts with the standard HiFi practice ( of using whatever polarity it takes to get a smooth crossover region between drivers covering different bandwidths ).
It's my experience that music mixed down on polarity correct monitors sounds different when played back on home speakers with happen-stance polarities ( ie; actual musical parts will go missing, which is hardly a subtle effect to argue over ).
I will grant that it may depend on the test subject. e.g. The great concert violinist can hear it. e.g. I personally have absolute pitch but I don't have absolute polarity.
My personal practice is to always wire the bass unit in a multi-way system in the correct polarity, as I believe most other people do, but upon close examination my choice is illogical (as is the reverse). You noticed I left open the question whether people can HEAR it, but I centered my claim on whether we can DO it in the real world.
First, if my bass unit is further extended downward by a sub, then what should be the polarity of that sub? (If e.g. a second order filter is used which requires flipping one of the polarities.)
Second, again same argument, a second order filter for the tweeter. Isn't the tweeter's polarity "wrong"?
Yes, and yes, but it has always been wrong. Other at 0Hz and at infinity hertz, the phase plots fluctuates wildly, and wraps around several times, such as such that a polarity flip which amounts to only 180 degrees of phase shift is eclipsed by the anomaly from the crossover.
Maybe that's why the full range folks will tell us the virtues of the single driver approach, and I find it likely that the violinist's monitor is a headphone or a full range single driver speaker, but it is not like full range drivers have no phase shifts. Maybe we can DSP until the speaker can reproduce absolute phase, only then will there be a meaningful psychoacoustic experiment be possible.
All I am getting at, is to tell the OP not to get worried about mid-range and tweeter polarities. If the crossover calls for them to be "wrong", it would be wrong to right it. That's all. I myself still religiously wire the bass unit correctly.
My personal practice is to always wire the bass unit in a multi-way system in the correct polarity, as I believe most other people do, but upon close examination my choice is illogical (as is the reverse). You noticed I left open the question whether people can HEAR it, but I centered my claim on whether we can DO it in the real world.
First, if my bass unit is further extended downward by a sub, then what should be the polarity of that sub? (If e.g. a second order filter is used which requires flipping one of the polarities.)
Second, again same argument, a second order filter for the tweeter. Isn't the tweeter's polarity "wrong"?
Yes, and yes, but it has always been wrong. Other at 0Hz and at infinity hertz, the phase plots fluctuates wildly, and wraps around several times, such as such that a polarity flip which amounts to only 180 degrees of phase shift is eclipsed by the anomaly from the crossover.
Maybe that's why the full range folks will tell us the virtues of the single driver approach, and I find it likely that the violinist's monitor is a headphone or a full range single driver speaker, but it is not like full range drivers have no phase shifts. Maybe we can DSP until the speaker can reproduce absolute phase, only then will there be a meaningful psychoacoustic experiment be possible.
All I am getting at, is to tell the OP not to get worried about mid-range and tweeter polarities. If the crossover calls for them to be "wrong", it would be wrong to right it. That's all. I myself still religiously wire the bass unit correctly.
Thank you for your clarification but with all due respect I question the wisdom of this. Consider a simple 2-way 4th-order phase coherent crossover. At the crossover point, the phase is 180 degrees. It is only at 0Hz and at infinity hertz that the phase is correct. (Or, due to bass roll-off in the real world, it could be 0 degrees at 100Hz or so.)
It can't be "correct" at every frequency, period.
If your express goal is to preserve phase, I think FIR is the way to go, but good luck with monitoring with the latency thus introduced.
If for the sake of simplicity we temporarily put aside the issue of the inevitable phase shift associated with bass roll-off, let's say by definition the "correct" phase is 0 degrees at 100Hz. Then, from there, the phase shifts linearly, resulting in a constant group delay. A constant group delay is harmless, and equivalent to a fixed driver offset. Now there is only one filter alignment which satisfy this property - Bessel, and it only does so for the low pass section, except for the first-order (because technically if there is such a thing as a high-pass Bessel it is a made up construct just to sum flat with the low pass section). For the first order filter Butterworth and Bessel are the same and that could be the reason why the first order filter is so sought after.
( Full Quote for emphasis )
The topic here is about maintaining signal polarity > not phase.
The fact that you ramble on about phase ( not once mentioning polarity ) proves to me that you just don't understand the questions raised here.
THX standards exist and they exist for a few reasons.
One reason is to cut through the crap to arrive at audio systems that conform to the THX prescribed wiring scheme ( whether or not the end users understand the actual "whys" ).
The "rules" are there to avoid all the fuzzy thinking that gets associated with phase, like you've just posted. .
Hi Cyberstudio,
the violonist example was just to point that yes there is variations between individuals ability AND habits about listening.
My take on it is that this particular instrumentist ( which wasn't 'great' but a professional musician which play hours everyday in different conditions) was so used to the sound of his instrument that he was able to detect the difference in absolute polarity. I suppose he was able to make distinction in the way the harmonics are related to the fundamental and more broadly the way the wood should resonnate 'correctly'.
About FIR and monitoring well this is what i use and yes i can be an issue but not nescesseraly: it'll depend of the loudspeaker type.
With my three ways yes it can be bothering ( 500hz xover from mid to woofer: 5ms latency) and i've met a ( metal) guitarist which was bothered.
If i had a 2way crossed at 1khz latency would be 1ms and it would be inaudible... or at least usable without too much worry.
I agree with Earlk, you mix two different things: polarity which is an absolute reference and phase which is a deviation from this absolute reference.
There is no such things as an instrument which is 'linear phase' (as you pointed a box will have rotation once you go below fc of the 'box'), but there is an absolute reference about that: once you bow a string on the violin the instrument doesn't shift polarity arbitrarly, it vibrate using an absolute reference which may be identified (well, this is what the example make me think now, but i can be wrong).
the violonist example was just to point that yes there is variations between individuals ability AND habits about listening.
My take on it is that this particular instrumentist ( which wasn't 'great' but a professional musician which play hours everyday in different conditions) was so used to the sound of his instrument that he was able to detect the difference in absolute polarity. I suppose he was able to make distinction in the way the harmonics are related to the fundamental and more broadly the way the wood should resonnate 'correctly'.
About FIR and monitoring well this is what i use and yes i can be an issue but not nescesseraly: it'll depend of the loudspeaker type.
With my three ways yes it can be bothering ( 500hz xover from mid to woofer: 5ms latency) and i've met a ( metal) guitarist which was bothered.
If i had a 2way crossed at 1khz latency would be 1ms and it would be inaudible... or at least usable without too much worry.
I agree with Earlk, you mix two different things: polarity which is an absolute reference and phase which is a deviation from this absolute reference.
There is no such things as an instrument which is 'linear phase' (as you pointed a box will have rotation once you go below fc of the 'box'), but there is an absolute reference about that: once you bow a string on the violin the instrument doesn't shift polarity arbitrarly, it vibrate using an absolute reference which may be identified (well, this is what the example make me think now, but i can be wrong).
Krivium and EarlK: your point is taken with respect to polarity vs phase but doesn't that imply every LR2 is "wrong"? LR2 is addressing a phase difference of 180 degrees by flipping the absolute reference.
Stereophile had an alternative explanation for the audibility of absolute polarity. To me at least, it makes perfect sense. (No real-world driver has perfect symmetry and there is always a difference between the cone moving outward vs inward.)
Absolute Phase: Fact or Fallacy? | Stereophile.com
The rules are there to establish a convention that prevents mis-wiring. My thinking is NOT fuzzy but my ear is. My whole premise, right or wrong, was that the ear was no oscilloscope, and it can't tell apart flipping the absolute reference versus shifting the phase by 180 degrees. I don't see how else LR2 would have even worked. If you do, please enlighten me.
Stereophile had an alternative explanation for the audibility of absolute polarity. To me at least, it makes perfect sense. (No real-world driver has perfect symmetry and there is always a difference between the cone moving outward vs inward.)
Absolute Phase: Fact or Fallacy? | Stereophile.com
The rules are there to establish a convention that prevents mis-wiring. My thinking is NOT fuzzy but my ear is. My whole premise, right or wrong, was that the ear was no oscilloscope, and it can't tell apart flipping the absolute reference versus shifting the phase by 180 degrees. I don't see how else LR2 would have even worked. If you do, please enlighten me.
Last edited:
Hi Cyberstudio,
Not in my view: just think about a guitar for example, it doesn't have a 'perfect' phase behavior ( as it implies a cavity for resonance, you could expect phase variation).
I know this is not a loudspeaker but the point is in my view that in fine we don't really care...
The strategy used for a loudspeaker is a compromise or tradeoff, the fact that an lr2 isn't perfect in regard to phase or absolute polarity doesn't matter as it is just a way to deal with a problem and LR2 does have other advantage.
In the end it depend of what you target or your goal.
What EarlK was explaining is that having absolute polarity fixed is some kind of 'convention' to help in proaudio ( at least this is how i interpret it) especially for 'fixed' specified requirements ( like Thx is)..
In fact how i see things is that it is just an 'ideal' as a point source or a time aligned speaker is... 99% of loudspeaker are not. However this doesn't imply you can't have good or great results. Just that there will be compromise in some area.
As already spotted in this thread with one driver polarity shifted solution some information may be lost... or not!
Not in my view: just think about a guitar for example, it doesn't have a 'perfect' phase behavior ( as it implies a cavity for resonance, you could expect phase variation).
I know this is not a loudspeaker but the point is in my view that in fine we don't really care...
The strategy used for a loudspeaker is a compromise or tradeoff, the fact that an lr2 isn't perfect in regard to phase or absolute polarity doesn't matter as it is just a way to deal with a problem and LR2 does have other advantage.
In the end it depend of what you target or your goal.
What EarlK was explaining is that having absolute polarity fixed is some kind of 'convention' to help in proaudio ( at least this is how i interpret it) especially for 'fixed' specified requirements ( like Thx is)..
In fact how i see things is that it is just an 'ideal' as a point source or a time aligned speaker is... 99% of loudspeaker are not. However this doesn't imply you can't have good or great results. Just that there will be compromise in some area.
As already spotted in this thread with one driver polarity shifted solution some information may be lost... or not!
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.