[jon_010101]

Polarity makes a bigger difference for amplifiers and speakers with asymmetric transfer functions - i.e., those with high even-order harmonic distortion.  When your "hifi" system amplifies the positive and negative halves of the signal differently, the effects of reversing phase will be greater - and not due entirely to the music, but to the reproduction

[andyr]

Polarity makes a bigger difference for amplifiers and speakers with asymmetric transfer functions - i.e., those with high even-order harmonic distortion.  When your "hifi" system amplifies the positive and negative halves of the signal differently, the effects of reversing phase will be greater - and not due entirely to the music, but to the reproduction

Jon, can you make your statement a bit easier for non EEs to comprehend? 

I (think I) understand the gist of what you are saying but I have some Qs:

* is a tube amp typically one with "high even-order harmonic distortion"?  So reversed polarity will be easier to hear with a tube amp than a ss amp?

* so what is the situation with an amp which has "high odd-order harmonic distortion"?  I believe this is what many ss amps show?

Thanks,

Andy

[Wayner]

Yes, could you make it even clearer for EEs to understand.

[opaqueice]

Polarity makes a bigger difference for amplifiers and speakers with asymmetric transfer functions - i.e., those with high even-order harmonic distortion.  When your "hifi" system amplifies the positive and negative halves of the signal differently, the effects of reversing phase will be greater - and not due entirely to the music, but to the reproduction Thumbs up

Right - that's exactly what I just said (although I couldn't be bothered to figure out if it was even or odd order).   Incidentally, not all distortion is harmonic - what really matters is the asymmetry.

Well said Andy. As one who is sensitive to correct polarity, it makes a big difference for me. Couldn't live without that switch. Mine is in the digital domain. no invert switch on preamp. The phono stage has one as well. I have found that LPs are the worst culprits. Some with the bass out of phase some just all out of phase. Love that switch.

That fits perfectly with this.  Vinyl setups are notorious for high levels of asymmetric distortion.

Yes, could you make it even clearer for EEs to understand.

Take an extreme case - imagine an amp that only amplifies positive input voltages, and sets negative ones to zero.  Of course it would sound horrendous, but bear with me.  Input a signal.  Unless it's symmetric (meaning, if you flip negative and positive voltages it looks the same) it's going to sound different coming out of that amp depending on whether or not you reverse the input wires - because in one case one half of the signal gets amplified, and in the other case the other half gets amplified, and the two halves aren't the same.

[jon_010101]

Amplifiers that produce no distortion don't exist - and they have a perfectly smooth straight-line transfer function, such that the output signal is equal to the input signal multiplied by a constant (Out=Gain*In).  Amplifiers that produce distortion have a gain function which depends on input signal amplitude (Out=Gain(In)*In).  (Of course, there are a lot more variables... but let us oversimplify for now )

Amplifiers that produce even order (like, 2nd order, 4th order...) harmonic distortion do so because they amplify asymmetrically - with a curved transfer function that leads to waveforms with differing distortions to the tops and bottoms.  Amplifiers that produce odd order (like, 3rd order, 5th order...) harmonic distortion exhibit symmetric distortion - with a curved transfer function that leads to the same form of distortion on the peaks and troughs of the signal waveform.  The audio perception of the asymmetrically-distorted signal varies more dramatically with polarity, since the waveform now looks quite different if you turn it upside down.

An example of this would be single-ended triode amplifiers, which tend to produce even harmonic distortion, due to inherent asymmetry.  Push-pull amplifiers (tube and solid state) are able to cancel this asymmetric distortion, and thus tend to contribute symmetric distortion (3rd harmonic) - although in much lower levels.  When people claim that they can hear absolute polarity more easily with a SET amplifier than solid state or push-pull tube, it is because the amplifier enhances to the asymmetry of the signal (especially at higher output levels), leading to greater discernibility of signal phase.  Of course there are many variables... all amplifiers exhibit distortion, and usually both odd and even order components are present.  And, some amps do distort more than others. 

[andyr]

Amplifiers that produce even order (like, 2nd order, 4th order...) harmonic distortion do so because they amplify asymmetrically - with a curved transfer function that leads to waveforms with differing distortions to the tops and bottoms.  Amplifiers that produce odd order (like, 3rd order, 5th order...) harmonic distortion exhibit symmetric distortion - with a curved transfer function that leads to the same form of distortion on the peaks and troughs of the signal waveform.  The audio perception of the asymmetrically-distorted signal varies more dramatically with polarity, since the waveform now looks quite different if you turn it upside down.

An example of this would be single-ended triode amplifiers, which tend to produce even harmonic distortion, due to inherent asymmetry.  Push-pull amplifiers (tube and solid state) are able to cancel this asymmetric distortion, and thus tend to contribute symmetric distortion (3rd harmonic) - although in much lower levels.  When people claim that they can hear absolute polarity more easily with a SET amplifier than solid state or push-pull tube, it is because the amplifier enhances to the asymmetry of the signal (especially at higher output levels), leading to greater discernibility of signal phase.  Of course there are many variables... all amplifiers exhibit distortion, and usually both odd and even order components are present.  And, some amps do distort more than others. 

Thanks, Jon - great explanation. 

So if "Push-pull amplifiers (tube and solid state) are able to cancel this asymmetric distortion, and thus tend to contribute symmetric distortion (3rd harmonic) - although in much lower levels", does that mean that push-pull tube amps produce 3rd harmonic distrortion just like ss amps do?  In fact they may well produce higher levels?

And what about 5th / 7th HD etc?

Thanks,

Andy

[jon_010101]

Thanks, Jon - great explanation. 

So if "Push-pull amplifiers (tube and solid state) are able to cancel this asymmetric distortion, and thus tend to contribute symmetric distortion (3rd harmonic) - although in much lower levels", does that mean that push-pull tube amps produce 3rd harmonic distrortion just like ss amps do?  In fact they may well produce higher levels?

And what about 5th / 7th HD etc?

Thanks,

Andy

Happy to help 

Push-pull amps generally produce third harmonic as dominant *if* they don't include any single-ended gain stages, and *if* they don't have imbalance between the signals in each of the push-pull "halves", and *if* the output transformer is not saturating or experiencing DC offset, along with a few more qualifications.  Push-pull signal imbalance (one of the biggest design concerns) contributes to increased 2nd harmonic due to - you probably can guess it - increased asymmetry.  Many push-pull tube amp circuits also include a single ended triode or pentode input stage, which can also contribute some 2nd HD.  Higher harmonics can be produced, too, depending on the tubes being used and how they are implemented, but generally at lower magnitudes - and things do get weirder when various types of feedback are used.  You should really expect to see a mix in almost any tube amplifier, with 2nd and 3rd dominating in most cases for SE and push-pull amps, respectively.

BTW, some of the single-ended amp assumptions I make do not necessarily apply to pentodes, where it is possible to specify an operating point where the 2nd harmonic nearly vanishes - With careful tuning, you can make a pentode amp's distortion signature look like almost anything you want (although I'd usually prefer it to be as weak and low-order as possible ). 

Also, comparing push-pull solid state to push-pull tubes - the magnitudes really can be quite comparable, and the tubes often produce "cleaner" (lower-order) distortion components.  My buddy and I are getting about 0.02% THD, mostly 3rd order and some noise, at 1 kHz / 1 Watt for our latest amp projects.  It may be a bit lower in the final version after some tuning, targeting the 0.005-0.015% range. 

[clarkjohnsen]

"Music normally creates compression waves. Electronics, however, often invert that natural, positive polarity to unnatural, negative rarefaction, thus diminishing physical and aesthetic impact. The term Absolute Polarity uniquely describes the correct arrival to the ear of wavefronts from loudspeakers, with respect to actual musical instruments.

That part is wrong.  It's just not true that "music creates compression waves". 

Regrettably I cut that from an earlier, vaguely-worded version. The correct expression is, "Musical instruments normally create compression transients."

First off, all sound waves consist of an alternating series of rarefactions (lower air pressure) and compressions (higher air pressure) - that's what a sound wave IS.  When the wavefront arrives at your ears/mic, the first deviation from average pressure can be either of the two.  Now probably what you mean is that the leading edge of the wave generated by a musical instrument is a compression.  But that's not correct either - vibrating objects can either create compression OR rarefaction wavefronts, depending on how you strike or pluck them and where you're standing. 

Quite so, but that does NOT obviate the need for whichever way it's hit, to be reproduced the same way out of the loudspeaker(s).

I've done the experiment myself to verify that - I took a string, plucked it, and recorded the result with a mic.  As expected, whether the leading edge of the waveform was compression or rarefaction depended on which direction I pulled the string before releasing it.  Same goes for percussion - you can flick a pen with your fingernail, strike a gong, hit a cymbal - the leading edge will be either rarefaction or compression depending on where you are standing relative to the direction the pen/gong/cymbal was struck from.  For something more complex like a violin, I suspect the leading edge will still depend to some extent on the direction of the bow stroke and the position of the listener, but I haven't tried to verify that.

Ditto my previous remark.

If you don't believe me, with a cheap mic and a laptop you verify it yourself in five minutes.  Just record the sound of flicking a pen (or paper or probably anything else).  Do it once flicking towards the mic, and once away, and see for yourself.

Oh I believe you. But that's not the point!

[clarkjohnsen]

But if you then flicked a phase reversal switch on the tape/CD-R machine, so that the recording was now playing back at you with reverse absolute polarity, it would sound different - strange in fact.  This is because you are no longer hearing the true leading edge of the notes.

Actually the difference is quite subtle.  I happen to have an audiophile test CD (from Chesky Records) in which there's a recording of trumpet note and then the polarity-reversed version.  I haven't tried to ABX it, but the difference is really small even over phase-coherent headphones.

You're making several uncharted assumptions.

1) The CD (or the recording) clearly portrays polarity.

2) The method used to switch polarity was right.

3) The playback system used, is capable of polarity portrayal.

4) You yourself know what to listen for.

If you say you can't hear absolute phase change then IMO there are two possibilities:

1.  you have speakers which use higher-order crossovers which require one driver to be connected in reverse ... this makes it very difficult to pick up absolute polarity issues.  (Mine are like that.   )

2.  you are one of those people who are not particularly sensitive to it (which is perhaps a blessing!   ).

There is a third possibility which IMHO is far more probable.  Many audio systems induce significant distortion.  If the distortion is asymmetric with respect to zero voltage - which is usually the case - reversing the polarity will make an audible difference on any recording which is itself asymmetric (and that's almost everything other than pure single test tones).  So my hypothesis is that people that hear clear differences are listening through systems that induce lots of harmonic distortion, hence making the effect clearly audible.  If instead you listen through a low-distortion system, the effect is extremely subtle (generally it's audible only under artificial circumstances).

[Laughter] So many, many people over the decades, listening over widerange low-distortion systems, have managed to hear polarity that the phenomenon is unchallengeable; nor do the vast majority of them think it "subtle".

[clarkjohnsen]

Vinyl setups are notorious for high levels of asymmetric distortion.

LOL! And digital setups are even more notorious for their punishing edginess.

But my experiments have found that analog yields better results for detection of polarity.

[Wayner]

If you don't believe me, with a cheap mic and a laptop you verify it yourself in five minutes.  Just record the sound of flicking a pen (or paper or probably anything else).  Do it once flicking towards the mic, and once away, and see for yourself.

That doesn't prove a thing except that the microphone has probably has a cardioid type pickup pattern. It's just like recording live with an audience. after the bands done playing and your hear the audience, it sounds different Sure, it sounds different, because the polar response of the microphone is usually piss-poor from the rear and you were actually recording secondary reflected sound waves that entered the front or sides of the microphone, because the cardioid microphone still picks up the best from the front and gets worse the farther off axis you go.

Wayner

[Russell Dawkins]

I think you missed the point because you miss-interpreted "towards" and "away". I think what was meant was flicking the paper towards and away from the mic, so the initial transient is a compression (when flicking towards) or a rarifaction if flicking the paper away from the mic.
In these cases it doesn't matter what the mic pattern is if the mic is facing the paper, and presuming the mic has the correct polarity, meaning that a positive-going pressure on the front of the diaphragm will result in a positive-going voltage on pin 2 (and a negative-going voltage on pin 3) of the microphone.
In the interest of correct absolute polarity, the positive going pressure on the microphone diaphragm should result, ultimately, in an outward-going movement of the speaker cone.

[clarkjohnsen]

It's just like recording live with an audience. after the bands done playing and your hear the audience, it sounds different Sure, it sounds different, because the polar response of the microphone is usually piss-poor from the rear and you were actually recording secondary reflected sound waves that entered the front or sides of the microphone, because the cardioid microphone still picks up the best from the front and gets worse the farther off axis you go.

Quite right, which is why audience noise is an unreliable indicator of correct polarity.

[Wayner]

I still don't see the connection. Let's take a body of water for example. If I drop anything into it be a stone, a boat or a duck, they are going to create a positive wave as the first wave because the object made a depression in the water because the object displaced it.

If we then take that analogy to air, no mater what sound we make it will be in a positive manner, wheather I click the pen facing forwards or backwards, shake the sheet of paper forwards or backwards. it all created positive pressure.

Movement always causes a positive air displacement. (pressure).

Wayner

[andyr]

If we then take that analogy to air, no mater what sound we make it will be in a positive manner, wheather I click the pen facing forwards or backwards, shake the sheet of paper forwards or backwards. it all created positive pressure.

Movement always causes a positive air displacement. (pressure).

Wayner

Are you shure? 

If I stand in front of a mic (let's assume it's connected up with absolute polarity) and blow towards it, it's creating positive pressure (diaphragm displacement).

But if I suck instead of blow ... that's negative presssure.  IE. rarefaction - the reverse! 

Regards,

Andy

[jon_010101]

Not a good idea to compare surface waves with compressional acoustic waves   

Your speaker can produce a positive or negative initial displacement - think of the cone "sucking" in as the pressure in the cabinet increases and the room pressure decreases - or the cone pushing out and increasing room pressure while decreasing cabinet pressure. 

Surface waves have some other key differences, too, which complicate the analogy further - specifically, they exhibit dispersion, where the phase fronts move faster than the overall wave packet, so that the initial perturbation can be "up" or "down" depending when you measure it.  At audio frequencies, this effect is negligible for acoustic waves (but important at ridiculously low frequencies <0.1Hz).

[Wayner]

OK, what musical instrument sucks 

Sucking in air didn't make the diaphragm move, unless you were "eating the microphone" which in itself is an un-natural (and in some states illeagle) act 

The sound waves you made opening your mouth and sucking wind still initially made a positive wave. The speed of sound traveled a lot faster than the speed of the wind you sucked, so the microphone respond to the sound way before it responded to the actual air pressure change. Your argument doesn't hold wind. 

Wayner

[jon_010101]

The sound is the pressure change that counts - Ignore any overall flow, the small pressure oscillations are what matter.  Depending on where your microphone is, and what instrument is involved, the initial perturbation can appear positive or negative.  When your speaker excites the same waveform, that will begin with an appropriate initial compression or rarefaction.

And, the blowing or sucking of AndyR does not contribute meaningfully to the signal 

[Wayner]

But Jon,

If I have a microphone facing into a hurricane and I whisper from the back, who is going to win? I hear what you are saying, but I think there is a physics problem here.

We all know that microphones are hit from all sides from direct and reflected sound simultaneously. The diaphragm can't move in both directions (at one time) so one is going to win. That winner is the strongest wave.

I do own a pair of speakers that you can listen to the back side. They are the mighty MartinLogan reQuests. But, because of their curved panels, one side disperses, the otherside focuses.

I guess you are arguing that whatever started the wave (positive or negative pressure) painted the picture and I just have some physics problems with that. I think what everyone is hearing (or thinks they're hearing) is something else.

[jon_010101]

We all know that microphones are hit from all sides from direct and reflected sound simultaneously. The diaphragm can't move in both directions (at one time) so one is going to win. That winner is the strongest wave.

A really loud sound still consists of alternating compressions and rarefactions, there is no near-"DC" pressure component to worry about in music (we leave that task to the barometers ).  All of the incident waves modulate the diaphragm at different frequencies and different intensities.  Whether or not you can detect one or the other depends on the dynamic range of the system, and the linearity of the microphone.