• 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!

Absolute Polarity - Myth or "Important"?

  • Thread starter Deleted member 23982
  • Start date

René - Acculution.com

Senior Member
Technical Expert
Joined
May 1, 2021
Messages
411
Likes
1,194
Oh no, I certainly wasn't referencing anything from my alma mater, I was thinking of the thread here by @René - Acculution.com . :)

The recording tutorial I read was this one:
https://www.uaudio.com/blog/understanding-audio-phase/

A pure impulse signal doesn't exist in music or in nature (except for Big Bang?). Even in a nuclear explosion, a rarefaction wave follows the compressed detonation front. Music is really about sine waves. And they have an uncanny quality of repeating.:)

Agreed.

Too bad the Clarke fellow is not around to help me with the explaining. Or the @René - Acculution.com fellow.
I am not up to speed (back from vacation). What can I do? :)
 

René - Acculution.com

Senior Member
Technical Expert
Joined
May 1, 2021
Messages
411
Likes
1,194
Oh no, I certainly wasn't referencing anything from my alma mater, I was thinking of the thread here by @René - Acculution.com . :)

The recording tutorial I read was this one:
https://www.uaudio.com/blog/understanding-audio-phase/

A pure impulse signal doesn't exist in music or in nature (except for Big Bang?). Even in a nuclear explosion, a rarefaction wave follows the compressed detonation front. Music is really about sine waves. And they have an uncanny quality of repeating.:)
A pure impulse will have a (Fourier transformed) frequency spectrum with a flat magnitude and zero phase at all frequencies; one phasor per frequency giving the magnitude and phase of that component. Rotate all of these (infinitely many on a continuous spectrum) phasors 180 degrees (this is not a time delay!!!), and the impulse will be upside-down.
 

Frgirard

Major Contributor
Joined
Apr 2, 2021
Messages
1,737
Likes
1,040
My take is that if you play mainly with digital files and notice a difference, you can always flip the recorded file itself, to the one that sounds best. If you have a simple remote switch, the listening session to determine the best can be done fast. Flip the file to the best sounding, and be done. Wrong polarity in bass sessions is at least in my ears audible and also be annoying and harsh in a few instances. But as said before, it is limited to a few records in bass region.

That is why I wish back a polarity switch on DACs. I would buy one, if it is not to expensive.
On my old RME Babyface, the phase can be reversed for each channel.
I guess RME has this capability on all of their products. Perhaps this is common with sound cards in the pro world.

When i was audiophile, the Mmicromega dac Duo pro2 and Duo bs (30 yers ago) had that also.
 

Holmz

Major Contributor
Joined
Oct 3, 2021
Messages
2,017
Likes
1,241
Location
Australia
I am not up to speed (back from vacation). What can I do? :)

Well Sir, some of this stuff in the quotes below… which generally is summed up as:
  • Whether the polarity of two speaker matters
  • Whether having the polarity of every other driver flip 180 with higher order passive XOs.
  • Whether there is a benefit from having phase pretty constant in a speaker system
  • And do mixing engineers invert polarity on tracks that create the whole song in a Willy Nilly fashion?


If there is a signal that is deemed as being correct then the output can only be correct if it matches the input.
One may like to chuck a minus sing in there, but the whole signal is then upside down at that point.

it may be subtle, but I would not want to start with a known flaw in the system… irregardless of whether some percentage of recordings were boned in the studio.

Ok, so then am I to assume that you are just trolling me?



Continuous signals from strings, and speech vowels, are different than pluosive vocalisation, consonants, and drums strikes.
We do not listen exclusively to tones in the majority of music.

A pure Dirac delta function is usually what people use to describe a system’s behaviour in terms of theory.

If you want the woofer to suck back in at the start of an explosion, and the push out when the microphone sees the rare fraction wave, then I suppose you are free to that. but it is not correct… it is the exact inverse of correct.
So it is just not replicating the reality of the SPL that the microphone recorded.

To say that it does, or to say that music and movies sounds are more like tones than like explosive impulses is clearly in error for any action movies, and for most sounds other than monk’s constant humming.


Oh no, I certainly wasn't referencing anything from my alma mater, I was thinking of the thread here by @René - Acculution.com . :)

The recording tutorial I read was this one:
https://www.uaudio.com/blog/understanding-audio-phase/

A pure impulse signal doesn't exist in music or in nature (except for Big Bang?). Even in a nuclear explosion, a rarefaction wave follows the compressed detonation front. Music is really about sine waves. And they have an uncanny quality of repeating.:)
 
Last edited:

René - Acculution.com

Senior Member
Technical Expert
Joined
May 1, 2021
Messages
411
Likes
1,194
Well Sir, some of this stuff in the quotes below… which generally is summed up as:
  • Whether the polarity of two speaker matters
  • Whether having the polarity of every other driver flip 180 with higher order passive XOs.
  • Whether there is a benefit from having phase pretty constant in a speaker system
  • And do mixing engineers invert polarity on tracks that create the whole song in a Willy Nilly fashion?
It is a bit of a tricky topic, but the basics of it tie in with my other thread about Polarity vs Phase: A systems that flips polarity, will invert the sign of any input signal. Positive becomes negative, and vice versa. Whether people are willing to accept this or not, the way to characterise such a system, is by shifting all PHASOR phases related to this system at all frequencies 180 degrees(!). This does not entail any time delay of any kind(!). The delay aspects of the system cannot be seen at one frequency, since if you only know the phase at that one frequency; all you can see is what the steady-state phase will be (and how at each particular frequency the steady-state output will LOOK LIKE a time delayed version of the input; phase delay is not group delay). Study complex theory and signal processing, and this should be apparent. The phasor phase is just an initial phase, initial offset, or however I can explain it in words, if the mathematics is too difficult.

But can you hear this or not?? I do not know. I tried a few listening tests once, and I could not hear any difference, but some people might be able to, and maybe you can train yourself to some degree. But it could be down to non-linearities in the loudspeakers themselves, and if your hearing is also non-linear in that it has some half-wave rectifier functionality in it, then sure, why not. There is certainly something 'uncomfortable' about the whole moving inwards vs moving outwards thing, but the system will have the same magnitude and delay aspects as a one that does not invert polarity, and so one should not expect any difference, if linearity is assumed in the loudspeaker/hearing chain.

And then you of course you have the issue with music being comprised of many inputs from different microphones and different situation.

It is probably best to look at a single driver first and try to define the polarity of this on its own. Is the 'normal' polarity the one that gives you zero phase somewhere in the pass-band, and the 'flipped' one the one that gives you 180 degree in the same frequency interval? And which way should it move when you apply a battery; you need to look at it both transient-wise and steady-state, and consider the behavior of minimum phase vs non-minimum phase systems. I will try and find the time to show this with an example.... (busy, busy). But this page has some nice insights https://www.prosoundtraining.com/2016/11/14/meaningful-loudspeaker-phase-response/

And when you then finally consider a complete loudspeaker system with each driver having its own reponse and cross-over section, and the total output probably being a non-minimum phase all-pass system, then most people lose track of the whole phase thing. Some filter topologies such as Linkwitz-Riley has the same phase for each section as for the entire system (sometimes called phase-coherent, I believe), and so magnitudes can be added directly, but it is not a linear phase so transient behavior is off. The main point for most designers is probably to have a fairly flat magnitude on-axis, and for some filter topologies this just entails a sign change falling out of the mathematics for some of the filter sections, and one way to obtain that is to flip the wiring; if you don't then you don't get the response you are looking for.

I am not sure how important absolute phase is. I think the linearity of the phase, which relates to the temporal behavior of the system, is probably more important, but this is just based on having listened to the Kii3s once in two different filter settings. It is difficult to discuss these topics online, as the mathematics needs to be in focus before all else, and that is not easily digestible for the typical audience.
 
Last edited:

jsrtheta

Addicted to Fun and Learning
Joined
May 20, 2018
Messages
935
Likes
990
Location
Colorado
It is a bit of a tricky topic, but the basics of it tie in with my other thread about Polarity vs Phase: A systems that flips polarity, will invert the sign of any input signal. Positive becomes negative, and vice versa. Whether people are willing to accept this or not, the way to characterise such a system, is by shifting all PHASOR phases related to this system at all frequencies 180 degrees(!). This does not entail any time delay of any kind(!). The delay aspects of the system cannot be seen at one frequency, since if you only know the phase at that one frequency; all you can see is what the steady-state phase will be (and how at each particular frequency the steady-state output will LOOK LIKE a time delayed version of the input; phase delay is not group delay). Study complex theory and signal processing, and this should be apparent. The phasor phase is just an initial phase, initial offset, or however I can explain it in words, if the mathematics is too difficult.

But can you hear this or not?? I do not know. I tried a few listening tests once, and I could not hear any difference, but some people might be able to, and maybe you can train yourself to some degree. But it could be down to non-linearities in the loudspeakers themselves, and if your hearing is also non-linear in that it has some half-wave rectifier functionality in it, then sure, why not. There is certainly something 'uncomfortable' about the whole moving inwards vs moving outwards thing, but the system will have the same magnitude and delay aspects as a one that does not invert polarity, and so one should not expect any difference, if linearity is assumed in the loudspeaker/hearing chain.

And then you of course you have the issue with music being comprised of many inputs from different microphones and different situation.

It is probably best to look at a single driver first and try to define the polarity of this on its own. Is the 'normal' polarity the one that gives you zero phase somewhere in the pass-band, and the 'flipped' one the one that gives you 180 degree in the same frequency interval? And which way should it move when you apply a battery; you need to look at it both transient-wise and steady-state, and consider the behavior of minimum phase vs non-minimum phase systems. I will try and find the time to show this with an example.... (busy, busy). But this page has some nice insights https://www.prosoundtraining.com/2016/11/14/meaningful-loudspeaker-phase-response/

And when you then finally consider a complete loudspeaker system with each driver having its own reponse and cross-over section, and the total output probably being a non-minimum phase all-pass system, then most people lose track of the whole phase thing. Some filter topologies such as Linkwitz-Riley has the same phase for each section as for the entire system (sometimes called phase-coherent, I believe), and so magnitudes can be added directly, but it is not a linear phase so transient behavior is off. The main point for most designers is probably to have a fairly flat magnitude on-axis, and for some filter topologies this just entails a sign change falling out of the mathematics for some of the filter sections, and one way to obtain that is to flip the wiring; if you don't then you don't get the response you are looking for.

I am not sure how important absolute phase is. I think the linearity of the phase, which relates to the temporal behavior of the system, is probably more important, but this is just based on having listened to the Kii3s once in two different filter settings. It is difficult to discuss these topics online, as the mathematics needs to be in focus before all else, and that is not easily digestible for the typical audience.

It is a bit of a tricky topic, but the basics of it tie in with my other thread about Polarity vs Phase: A systems that flips polarity, will invert the sign of any input signal. Positive becomes negative, and vice versa. Whether people are willing to accept this or not, the way to characterise such a system, is by shifting all PHASOR phases related to this system at all frequencies 180 degrees(!). This does not entail any time delay of any kind(!). The delay aspects of the system cannot be seen at one frequency, since if you only know the phase at that one frequency; all you can see is what the steady-state phase will be (and how at each particular frequency the steady-state output will LOOK LIKE a time delayed version of the input; phase delay is not group delay). Study complex theory and signal processing, and this should be apparent. The phasor phase is just an initial phase, initial offset, or however I can explain it in words, if the mathematics is too difficult.

But can you hear this or not?? I do not know. I tried a few listening tests once, and I could not hear any difference, but some people might be able to, and maybe you can train yourself to some degree. But it could be down to non-linearities in the loudspeakers themselves, and if your hearing is also non-linear in that it has some half-wave rectifier functionality in it, then sure, why not. There is certainly something 'uncomfortable' about the whole moving inwards vs moving outwards thing, but the system will have the same magnitude and delay aspects as a one that does not invert polarity, and so one should not expect any difference, if linearity is assumed in the loudspeaker/hearing chain.

And then you of course you have the issue with music being comprised of many inputs from different microphones and different situation.

It is probably best to look at a single driver first and try to define the polarity of this on its own. Is the 'normal' polarity the one that gives you zero phase somewhere in the pass-band, and the 'flipped' one the one that gives you 180 degree in the same frequency interval? And which way should it move when you apply a battery; you need to look at it both transient-wise and steady-state, and consider the behavior of minimum phase vs non-minimum phase systems. I will try and find the time to show this with an example.... (busy, busy). But this page has some nice insights https://www.prosoundtraining.com/2016/11/14/meaningful-loudspeaker-phase-response/

And when you then finally consider a complete loudspeaker system with each driver having its own reponse and cross-over section, and the total output probably being a non-minimum phase all-pass system, then most people lose track of the whole phase thing. Some filter topologies such as Linkwitz-Riley has the same phase for each section as for the entire system (sometimes called phase-coherent, I believe), and so magnitudes can be added directly, but it is not a linear phase so transient behavior is off. The main point for most designers is probably to have a fairly flat magnitude on-axis, and for some filter topologies this just entails a sign change falling out of the mathematics for some of the filter sections, and one way to obtain that is to flip the wiring; if you don't then you don't get the response you are looking for.

I am not sure how important absolute phase is. I think the linearity of the phase, which relates to the temporal behavior of the system, is probably more important, but this is just based on having listened to the Kii3s once in two different filter settings. It is difficult to discuss these topics online, as the mathematics needs to be in focus before all else, and that is not easily digestible for the typical audience.
Thank you for the link.
 

DimitryZ

Addicted to Fun and Learning
Forum Donor
Joined
May 30, 2021
Messages
667
Likes
342
Location
Waltham, MA, USA
I can now happily report to be in possession of an excellent remote audio balanced phase switch, graciously created by @solderdude .

This device allows me to switch absolute phase from my listening seat while the music is playing. 2 XLR in and out plus 5V supply and a green/red LED on the front.

So far I can report that on lossy source - satellite radio - I don't hear any difference.

With CD and SACD on some older recordings I seem to have a preference, but it's very subtle. Will report more as I learn.

It's a really fun box to have in your system!

Thank you @solderdude !
 

jsrtheta

Addicted to Fun and Learning
Joined
May 20, 2018
Messages
935
Likes
990
Location
Colorado
I can now happily report to be in possession of an excellent remote audio balanced phase switch, graciously created by @solderdude .

This device allows me to switch absolute phase from my listening seat while the music is playing. 2 XLR in and out plus 5V supply and a green/red LED on the front.

So far I can report that on lossy source - satellite radio - I don't hear any difference.

With CD and SACD on some older recordings I seem to have a preference, but it's very subtle. Will report more as I learn.

It's a really fun box to have in your system!

Thank you @solderdude !
I can die now.
 

Holmz

Major Contributor
Joined
Oct 3, 2021
Messages
2,017
Likes
1,241
Location
Australia
Clearly, this is a hard room...

Perhaps the Z in DimitryZ is for the complex impedance?

I think it is good work you are dong.
And with recordings which are largely in phase it might be interesting to see what you find.
For multi-track rock, which can be god only knows what phase, it might be harder.
 

DimitryZ

Addicted to Fun and Learning
Forum Donor
Joined
May 30, 2021
Messages
667
Likes
342
Location
Waltham, MA, USA
Perhaps the Z in DimitryZ is for the complex impedance?

I think it is good work you are dong.
And with recordings which are largely in phase it might be interesting to see what you find.
For multi-track rock, which can be god only knows what phase, it might be harder.
Z seems like a bad letter, given the vicious Russian attack on Ukraine. And my last name starts with Z, so I can't win.

I have to do more clicking with my fabulous new remote to get any sense if all new recordings are 'phase neutral. '
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,843
Likes
35,809
Location
The Neitherlands
Self education/experimentation can bring more to some people than reading a thousand opinions and papers.
 
F

freemansteve

Guest
I think there are still amps out there that can have a slight DC offset at the output. This may make a difference.
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,843
Likes
35,809
Location
The Neitherlands
But that has nothing to do with phase polarity though.
In this case only the signal going to the amps has the polarity reversed.
 

René - Acculution.com

Senior Member
Technical Expert
Joined
May 1, 2021
Messages
411
Likes
1,194
But that has nothing to do with phase polarity though.
In this case only the signal going to the amps has the polarity reversed.
I remember reading years ago that some people could hear a difference between doing the polarity flip on the input side vs doing it on the loudspeaker side. Perhaps it could be related to this offset.
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,843
Likes
35,809
Location
The Neitherlands
I highly doubt it and see no 'mechanism' for that.

On the speaker side only the woofer would end up in a slightly different '0 position' but be static after that. All audio would still be produced normally. DC offset would have to be extreme to cause timing differences with the tweeter/midrange.
They would have heard a 'click' when the amp switches on and would have seen the woofer take on a noticeable different position.
On the amplifier side the amp would either remove the DC offset (AC coupled amp at the input) or go into protection when DC coupled and the DC offset would be extreme or the DC on the amps output would vary with the volume control.
In the latter case they would have complained about a scratchy volpot when adjusting the volume as well.

Personally I would chalk this up to 'misinformation' based on sighted testing.
 
Last edited:

René - Acculution.com

Senior Member
Technical Expert
Joined
May 1, 2021
Messages
411
Likes
1,194
I highly doubt it and see no 'mechanism' for that.

On the speaker side only the woofer would end up in a slightly different '0 position' but be static after that. All audio would still be produced normally. DC offset would have to be extreme to cause timing differences with the tweeter/midrange.
They would have heard a 'click' when the amp switches on and would have seen the woofer take on a noticeable different position.
On the amplifier side the amp would either remove the DC offset (AC coupled amp at the input) or go into protection when DC coupled and the DC offset would be extreme or the DC on the amps output would vary with the volume control.
In the latter case they would have complained about a scratchy volpot when adjusting the volume as well.

Personally I would chalk this up to 'misinformation' based on sighted testing.
Oh it was just self reported stuff in a HiFi magazine. Not anything of much scientific value. But the nonlinear effects are important in general, so for example Kms(x) would be different as the work point is different. Anyway, just chiming in.
 

600_OHM

Active Member
Joined
May 19, 2022
Messages
162
Likes
119
For decades I have wandered down this rabbithole myself. And always wondered about the placebo effect. Fertile ground for that.

For a few decades I could hear it, and *see* it - at least in one form, particularly percussion and high-impulse sounds, and couldn't put my finger on it. Maybe, just maybe, those trying to hear any difference can better do so if they focus their attention in one area, rather than listen to the entire spectrum.

Without trying to hijack this thread on something model specific, Solderdude (total respect!) and I have been having spirited discussion about the venerable old AKG K240. It is the only set of cans I know of that are PURPOSELY wired "Inverse In Absolute Phase".

WHY would they do this on purpose and not follow convention like all the rest of the cans I know of?

Forget about how poor the K240 is for absolute sonic purity when measured for the time being. My contention is that there is a specific purpose to the unconventional driver wiring. Because to me, it is not just a placebo, or simple FR, it is far too dramatic to me.

A psycho-acoustic trick if you will. A modern reference sums it up pretty well for me, especially in the words of Douglas Self where he points out this can be more easily discerned with percussion. Second paragraph of the sound reproduction section:


What he describes is *exactly* what apparently I am sensitive to when the weird driver wiring of the K240 is used.

What I hear (and even see on well-timed video) is the loss of a percussionist's "strike" of an instrument, and only hear the "ring". Drum solos are especially disturbing to watch. Before that when I could only hear this, I always thought there was something "odd" about a performance.

Other high-energy instruments like triangle-strikes, cowbells, anything metallic (or glass even) - that first strike is gone.

I can't put this down to solely being the responsibility of FR and poor overall musical tonality coming from the cans. Look beyond that - it is practically a physical sensation to me.

I think that's the problem - some people may be sensitive to it, and it occurs in the brain, where we'd like to put objective measuring devices into it, but most of us can't. :)

I guess I now need to use my conventionally wired cans, and buy an amp where I can invert the absolute phase, and see if I once again notice the loss of a high-energy instrument strike, and only hear the ring like I do with the K240's. I think it was wired on purpose this way - as a true studio can - and not hi fidelity - to reduce desensitization due to high energy impulses when doing long term monitoring.

YMMV.
 
Top Bottom