Reversing Polarity -- Voodoo or Easy Tweak?

In a recent thread I noticed a comment about reversing polarity of speaker wires on both speakers which sparked one of my earliest audiophile memories.

On the liner or cover notes of Dave Grusin: Discovered Again on direct to disc vinyl, circa 1977, it too recommended reversing the polarity on BOTH speakers, for best sound.

Although my first system was a 25 WPC Technics receiver with Infinity Qa's and lousy speaker wire, I still remember getting very enthusiastic about reversing the polarity and wondering if it did anything.

Can anyone explain this and/or recommend if this is even worth the experiment?
If you reverse the leads on only one speaker, you will be out of phase, and your imaging will be wrong. This can be easily checked with any available test disc. It's pretty easy to hear. A single vocal will not have a centered image, but will instead be split across the soundstage.

Absolute phase is something different. To reverse absolute phase, you reverse the +/- leads on both speakers. Absolute phase is more difficult to hear, but it's easy to experiment. BTW, you can reverse leads at the amp instead of the speakers with the same end result.
Some manufacturers (Bat for example) allow you to easily swap absolute phase with the remote & others (Hagerman for example) give you manual switches to play with.
My take from playing with a BAT integrated I owned was that on most recordings keeping absolute phase 'correct' sounded subtly better and on one or two recordings (The Blues Brothers being a prime example)swapping absolute phase made a nice improvement.
I think a lot has to do with how well the recording studio does at keeping track of absolute phase in the mixing & recording processes.
FWIW, a dealer I know has a classical album that was recorded out of phase and you have to swap wires on one speaker only to get it to sound good, & I suspect that on many recordings some of the mixing is 'right' and some of the rest is 'wrong'.
Jeff_jones...Speaking of what I will call "relative" phase (one speaker with respect to the other) the "proper" phase is not always clear. Out of phase signal has a diffuse and directionless quality, and that might be what the recording engineer intended. In a matrix multichannel setup out of phase signal localizes to the rear channels, so reversing one set of speaker wires just swaps front and back speakers. When played back on a two-channel system a source that is intended to be well defined center front (typically a soloist) is a monaural signal (same in both channels). All recordings have some degree of out-of-phase signal, often nothing but ambience. However, there are some exceptions. One good example is the recording by Buffy Sainte-Marie, "The Angel" where out-of-phase signal is used with stunning effect.

I do not suggest that anyone deliberately connect stereo speakers out of phase, but a signal being out of phase is not inherently "wrong", and can be used to great effect when a recording is mastered.
Whoops! I should have said that reversing phase of one channel BEFORE MULTICHANNEL DECODING will swap front and rear. Doing speaker wires in a 2-channel setup will localize ambience, and make the soloist diffuse and directionless.
Tvad and others

Didnt know it was called "absolute phase" - this is what was recommended on the first edition Dave Grusin vinyl.

So I guess the answer is like biwiring: it depends?!
Absolute phase, and phase (what Eldartford calls "relative" phase) are two different things, thus the difference in approaches to wiring the speakers to correct the problem.
The correct term when talking about wiring speakers is polarity as stated in the original question, not phase.

Phase has to do with time, polarity has to do with whether a signal is going positive or negative. If a speaker is wired with the wrong polarity it will be going out when it should be going in, and vice versa. I have friend who claims he can tell when it is reversed because a vocalist will sound like they are gulping air instead of exhaling.

When 2 signals are out of phase it means that one or the other has been shifted in time, it occurs earlier or later in time than originaly produced. One example is a crossover network. They not only attenuate the signals outside their passband, they also phase shift them so they occur at a different time relative to those in the passband.

The fact that many manufacturers use the terms interchangeably adds to the confusion, just as Eldartford did in his post. In the first part where he is talking about wiring the speakers the correct term is polarity. In the second when he talks about the Buffy Sainte-Marie recording the correct term is phase. This deliberate phase shifting is the same thing Hendrix did to get that swirling effect with some of his recordings.
Well, I thought I understood this topic, but I'm as confused as I've ever been.

Herman, your explanation makes sense, however I want to understand the effect of polarity, and the solution.

I own Stereophile's Test CD3, on which track 2 is titled "Stereo Channel Phasing". The description, and effect appears to be what you're referring to as polarity. Quoting from the Stereophile Test CD3 text:
What you should hear: In a conventional stereo system, a centrally placed listener should hear the image of the in-phase noise occupying a very narrow space centered between the loudspeakers, If the sound "splashes" to the sides at some frequencies, or the image is broadened at all frequencies, then there is something suspect in your system-most probably a loudspeaker or room-acoustic problem. The image of the out of phase noise should not be centered; in fact, it should generally be very hard for you to point to where it's coming from. With some loudspeakers, however, the out-of-phase noise might appear to come from a point to the outside edge of either the left or right loudspeaker. If you don't hear these clear distinctions between in and out of phase noise, or if you hear them reversed, try inverting the connection to one of your speakers.

The way I understand this is simply...from amp to speaker, connect plus to plus and minus to minus on both speakers. Connecting plus to minus on only one speaker usually results in the out of phase (polarity?) phenomenon mentioned in the quote from Stereophile.

Do I have this right? Is Stereophile referring to polarity when they speak of phase?

Cripes, my head is spinning...maybe it's out of absolute phase.
Herman...Agreed that terminology is commonly sloppy. However, "OUT" of phase means a shift of 180 degrees, which is the same thing as a polarity reversal.

Audiophiles like to worry a lot about the phase shift introduced by filters, as in a crossover network, but, as the phase shift increases beyond the break frequency, the amplitude decreases, so that phase becomes irrelevant.

Excellent post!


At the risk of oversimplifying things here:

If ONE speaker has + and - (it's polarity) reversed, then the system will be "out of phase".

And no offense , but no self respecting audiophile should need a test disc to hear this -- it creates a weird, vague image, and to my ears at least, a slightly disorienting, almost uncomfortable to the inner ear wandering sound.

When I sold hi fi during college, the owner of the store thought it would be amusing to demonstrate my lack of experience by asking my opinion on a system where he had wired the Acoustat speakers "out of phase". It took me about 15 seconds to hear this.

During the recording process, however, certain things might be recording about of phase to create weird imaging effects. In addition to Jimi Hendrix, I think Roger Waters Amused to Death is known for this type of demo.

My original question of course was

what happens?
and why do some recommend?

changing the polarity on BOTH speakers.

Although the "polarity" in this case changes, it does so on both speakers, the polarity remains consistent so the signal remains "in phase".

I think I am using these terms correctly now?
Cwlondon, I understood you in your first post. Few are stupid enough to wire one speaker out of phase with the other. Absolute polarity is a different matter. I wish I had an easy way to have switchable polarity, so that I had a choice, but I don't have. Even cheap digital players used to have this, but no longer. I am not about to change speaker leads for each cd or record.

In the past, on perhaps 10 percent of my records changing the absolute polarity made a great difference. On about 30 percent it made no difference.
The Grusin LP was recorded with the absolute polarity reversed. Shefield realized it after the fact and this is the reason for noting it on the cover. About half of the LPs and CDs have the polarity reversed. Some of the offenders are Mercury Living Presence, RCA Living Stereo, Capitol, MCA, Decca/London, DG and Warner. When your ear is trained to hear this, it will take about 3 second to pick up on it. I don't do vinyl anymore so it's easy for me to correct with the polarity switch on my Theta DAC. Then I mark the CD insert with a green (correct)or red (180 degrees)sharpie. A polarity switch on the preamp's remote would be even easier. Listen to a recording of Nat King Cole on Capitol both ways. There is a huge difference in these recordings and it's easy to hear. If you can't hear the difference, well...
Cwlondon wrote:
what happens?
and why do some recommend?

changing the polarity on BOTH speakers.

Although the "polarity" in this case changes, it does so on both speakers, the polarity remains consistent so the signal remains "in phase".

This is reversing Absolute Polarity (I mistakenly referred to this as Absolute Phase in an earlier post). In my experience, when the signal is in correct Absolute Polarity, the sound is somewhat fuller, and more complete. The stereo image will remain focused and centered just as when the polarity is maintained plus to plus, and minus to minus.

It's interesting to note that one manufacturer of equipment I own refers to this as Absolute Polarity, and another manufacturer refers to this as Absolute Phase. No wonder there's confusion!

Regarding your statement about the necessity of a test disc to hear correct phase, I don't believe the use of one makes the listener a less-than-self-respecting audiophile any more than the use of a tape measure makes a less-than-self-respecting carpenter. Frankly, for someone new to the hobby, which we all were at one time, a test disc can be an extremely useful tool. Futhermore, the Phase track on the Stereophile Test CD is a quick and easy way to double check for correct polarity (phase) without uncertainty or guesswork.

If ONE speaker has + and - (it's polarity) reversed, then the system will be "out of phase".

Their polarities are reversed, they are not out of phase. However, this description (out of phase) is so deeply ingrained that I should just give up on it. As pointed out, Stereophile uses it as do many manufacturers. I had a Conrad Johnson CD player with a "phase" switch when it was really a polarity switch.

Herman...Agreed that terminology is commonly sloppy. However, "OUT" of phase means a shift of 180 degrees, which is the same thing as a polarity reversal.

No it's not. With a symetrical waveform it is impossible to tell by looking on a scope whether a signal is inverted or 180 degrees out of phase, but musical signals are asymetrical so there is a difference. The usual example is a kick drum that is struck once and then damped. The intial wave that strikes your ear is a compression followed by a rarefaction. If you reverse polarity then it is a rarefaction followed by a compression. This is a big difference. If there was a phase shift then you would still get compression then rarefaction but earlier or later than what it should be.

Waveform A is of equal but opposite polarity to B. They start at the same time but one starts out going positive and the other negative. Waveform C is 1/2 cycle(180 degrees) out of phase with A because it starts 1/2 cycle later. As Eldartford points out, there is no difference between B and C once they get started, but this is because they are symetrical.

Waveform D and E are also equal but opposite polarity while F is phase shifted from D. There is a big difference in E and F whereas B and C appear to be the same. The kick drum (point kd) on waveform D will initially push the speaker cone out while on E it will suck it in.

Some say this all makes no difference in sound, but they are different.
When I said "equal but opposite polarity" I meant equal amplitude but opposite polarity.

If the picture doesn't show up here is the link

To be a total stickler for terminology, speakers CANNOT have their polarity reversed, since they are driven by AC, which has no + or - poles! Only DC has a + or - polarity. Speakers can have only an absolute phase inversion [both speakers have the wires reversed on the + and - terminals], out of phase [only ONE speaker wire will have the + and - terminals reversed...very bad indeed!], or have a phase shift. A phase shift occurs when different frequency ranges are shifted [delayed] by milliseconds. This can be due to crossovers, physical driver alignment, speaker wires, or the amplifier powering the speakers. Sort of like a prism splitting white light into the various frequencies of colored effect, a rainbow. Again, not good. But phase shift is pretty much a fact of audiophile life. The amount of phase shifting will determine the coherency [or not] of the sound.

FYI, here's a neat trick to get raw drivers in phase when the terminals are not marked or marked incorrectly [it does happen!], or for car installations where multiple long concealed speaker wire runs can easily be installed out of phase. Take a "D" cell battery, and hold the speaker wires to the battery. Note whether the speaker cone moves out or in. I consider "out" to be the + terminal and "in" to be -, this being a good memory aid.
Herman, your next to last post was great. Back in the early 70s I owned a pair of 3way Wharfedale loud speakers. The woofers were 12". The bass always seemed to be a little light, lacking authority. I had read somewhere to check the polarity of speakers and make sure the drivers were wired correctly, in reference to polarity, was to connect a 1.5V D cell battery, momentarily, across the speaker terminals + to +, - to -. If the speaker polarity was correct the speaker cones moved forward. On the Wharfedale's, the cones pulled back, just the opposite as they should have. I reversed the speaker cable leads on both speakers and the bass was improved.
Herman...You are a bit hasty in saying that the drum sound will be compression followed by rarefaction. It depends on which side of the drum the mic was on. Also, as a disturbance propagates the wave shape changes. Consider, for example a Tsunami, where the initial disturbance is always that the sea recedes, followed by the flood. I have observed a similar effect of a ship bow wave while going through a canal. There must be an explanation, but I don't know it.

Do you have any scope pictures of a mic signal from a drum?
I see we are still confused, and I don't blame anybody for being that way since it is a bit confusing even without the years of improper use.

AC, which has no + or - poles! Only DC has a + or - polarity. Speakers can have only an absolute phase inversion

This is incorrect. Absolute phase inversion makes no sense. I know what you mean, but it is an incorrect use of the term phase. Phase has to do ONLY with time, and you cannot invert time.

AC does indeed have a polarity at any given given point in time. It may be incorrect to say that a given AC signal is overall positive or negative, but at some points in time it is positive and sometimes it is negative.

A clear example is a balanced system. You have 2 signals. One is the same as what a single ended system would have, and the other is inverted. This second signal is also commonly described as being "180 degrees out of phase" with the other one, but it is not. It's polarity is inverted, not it's phase. In a sense they are mirror images of each other.

When one is going more positive the other is going more negative.
If one has a value at a given point in time of +2 volts and headed more negative, the other one will be -2 volts and headed more positive. As an aside, I see this as a major problem with balanced systems. They must take 2 signals of equal amplitude but opposite polarity and amplify them through 2 different chains of amplifiers with the exact same gain and phase shift in both chains.

A phase shift occurs when different frequency ranges are shifted [delayed] by milliseconds

This is correct. Speaking to Eldartford's earlier point about the phase shift in crossovers being inconsequential, I disagree. At the -3dB (half power) point of a first order filter the phase shift is 45 degrees. This is a significant amount, especially if the fundamental frequency from an instrument falls in the passband and the first harmonic falls outside.
Am I the only one who thinks this thread reads like Abbott & Costello's "Who's On First"?
I thought Who was on second.

Eldartford, perhaps you are correct about mic placement, but if you are listening to a live kick drum you will surely be out in front of it unless you are the drummer and the sequence will always be one way or the other, and if you reverse it when playing it back it will sound different.
Herman...Both Hi and Lo signals are shifted 45 degrees at X/O, so the woofer/tweeter discrepancy is 90 degrees (for first order). In fact the discrepancy is constant at 90 degrees for all frequencies, but causes most problems around the X/O frequency where both drivers are emitting sound. With a second order crossover, 180 degrees of phase shift, there will be a deep sharp notch in frequency response unless one driver is connected with inverted polarity (or, as some might say, out of phase).That's why I like 24dB crossovers, where the shift is 360 degrees, so that sine waves are back in "phase" if you will pardon the term. Sine waves are a pretty good approximation of musical sounds, at least for a few cycles.
What's on second. Who's on first, and Polarity's in the bullpen.
so that sine waves are back in "phase" if you will pardon the term.

I know what you mean, in and out of phase just seems to sound right, and any use of the term polarity with a modifier seems to sound clunky and incorrect.

Both Hi and Lo signals are shifted 45 degrees at X/O, so the woofer/tweeter discrepancy is 90 degrees (for first order).

I think we are talking about 2 different things here. My point was that all musical instruments produce harmonics in addition to the fundamental tone. So if the fundamental falls inside the passband it will not be phase shifted and if a harmonic falls outside the passband it will be, or vice versa. I think this phase shift in the harmonic structure will be audible. I haven't conducted any experiments to confirm this hypothesis.
Well I am totally confused now. I just spent about 20 min listening to the CD of Jennifer Warnes, "The Hunter," track #8, "Way Down Deep". I listened to the first min and half at least 10 times then I reversed the speaker Cables at each of my ProAc studio 200 speakers. I then listened again several times. There is a definitely a difference in sound. I switched them back and forth several times.

On this particular track of this CD with the speakers cables reversed imo bass sounded tighter and Jennifer's voice sound more focused. Is it all in my head...

IF any of you have this CD check for yourself and post back.

Trust your ears, Jea48.
I listened to the entire Jennifer Warnes Cd, I spoke of in my earlier post, with the speaker cables reversed on my speakers. After a few tracks the sound just didn't seem involving, became kind of boring if you know what I mean. After the last track played, I reversed the speaker cables back the way the were originally, Amp+ speaker+, Amp- speaker-. Pushed play on the CDp, to cut to the chase, I prefer the sound better this way...Biggest thing I noticed Jennifer's voice was fuller, more involving.

Herman notes:
I haven't conducted any experiments to confirm this hypothesis
Someone else has -- and I tried desperately & unsuccessfully to find the ref. Apparently you WILL notice a discrepancy. This, (afa I remember) is variably termed as "slight loss of detail" or "microdynamics", etc. This "discrepancy" was particularly obvious when the same musical passage was played with fundamental & 2nd & 4th harmonic shifted into the pass band. ANyway, if I find the ref, I post it.
I wish I had access to a storage scope and a mic. It would be very interesting to examine the shape of transient waveforms (like the drum) both close mic'd and distant. Of course, once the signal is played back through a loudspeaker and rerecorded I suspect that any slight difference due to polarity/phase will be unimportant. Very few speakers can reproduce a square wave (original Ohms were one), and those only at relatively low frequency like 1000m Hz.
Eldartford, I suspect that you are right that would see little difference on a scope. This is always my concern as in some cases I heard a decidedly different sound with a reverse in polarity. I suspect that recording engineers indifference to polarity or multi-miking may contribute to why some records and disks seem unresponsive to such changes. But if recordings in which I hear differences fail to show a difference on a scope, I suspect that you and I would come to different conclusions.

At any rate since I now have no easy way to reverse polarity and some recordings sound much better than others, I can only suffer along. I did have a parafeed output line stage from Exemplar which could easily have had a polarity inverter as it was transformer coupled, but it is not available in the H-Cat. Other tube units that I have had, notably the Siltech Millennium did provide a useless inverter which introduced another tube stage to get the inversion.
Unlike most audiophile tweeks, the absolute polarity effect, at least superficially, seems to make scientific sense. Also, unlike most tweeks, it is easy and cheap. However, I do suspect that the science might fall apart on closer examination, and the real world usefulness of the tweek is minimal because of what you describe as "recording engineer indifference".
I guess I would say that in the past when I had the capability on digital easily compare that about 1 disk in 10 made a difference. I should say that Clark Johnsen also once visited when I had the capability on vinyl and digital. He heard it on more recordings, and when he pointed out what differences he was hearing, others did indeed also hear them.
I found a Sheffield Lab-5 vinyl LP of Dave Grusin, "Discovered Again" 1976, I had not listened to for years. I cleaned both sides put it on the TT platter for a listen. I first played both sides of the Lp to warm up the cartridge. Then I sat down for a listen. For 1976 it still sounds very good. I listened to both sides of the Lp and then focused in on Grusin's piano mainly. I then reversed the speaker cables at the speakers. Their was a difference in sound, but not overwhelming. Imo Grusin's piano did sound cleaner and more open, with a tad more air.

So what do you conclude?
Herman posted,
"Phase has to do with time, polarity has to do with whether a signal is going positive or negative. If a speaker is wired with the wrong polarity it will be going out when it should be going in, and vice versa. I have friend who claims he can tell when it is reversed because a vocalist will sound like they are gulping air instead of exhaling."

Herman puts it into words better than I can. I also agree with every thing he has said in this thread.

Recorded music is converted to electrical signal and that signal has polarity. With no signal being a base line with positive signal above the base line and negative signal below the base line. Positive signal will cause the coil of the speaker to move foward from a no signal state and negative signal will cause the speaker coil to pull back from a no signal state. So imo if the source signal for a given time is positive then the speaker, if the polarity is the same as the signal for the same given time, should move forward. Sorry if I do not say it as eloquent as Herman.

I was talking with my son the other day and I mentioned this thread. My son was the one that reminded me I had the Grusin LP. He told me that Frank Zappa recorded his music with reverse signal polarity.

Question why would any one record music reversing the signal polarity deliberately?
Question why would any one record music reversing the signal polarity deliberately?

Perhaps to give their music a distinctive and uniquely different sound.
"Perhaps to give their music a distinctive and uniquely different sound."

I thought of that but in the case Of Grusin LP there is this statement;

Audiophile Note: "For optimum transient response and spartial clarity, we recommend that the polarity of BOTH channels be reversed at the speaker terminals(+ output terminal on power amplifier to - terminal on speaker and vice versa), however this procedure is not necessary for for perfectly satisfactory playback."
There has been a entire book published about phase inversion called "The Wood Effect" by Clark Johnsen [who, I believe, is a long term member of Audio Asylum]. This book has many documented studies concerning listeners' abilities to hear a difference when phase is inverted [both speakers polarity reversed]. Some pretty heady reading, buy people do seem to hear a difference in almost all of studies.

That's it!

....the exact text which inspired this post. I was just working from memory as my LPs have been in storage for 12 years.
Here's an idea. Why not write to Dave Grusin and ask him?
Tvad, did you wink?
Jea48, no, I didn't wink. Why? I was serious. Write him and ask him...unless he's dead, and I didn't know it?

Perhaps one issue with this recording is that it was cut direct-to-disc. Maybe doing so eliminated one link in the signal chain that would have otherwise corrected the polarity? Just guessing here...
Tvad, I thought maybe you were pulling my chain...LOL...
I did do a quick look for a possible email address for Grusin on Google. No luck yet.

I did Run across a thread where Clark Johnsen gave this link;
Phil Spector introduced his "wall of sound" back in the 50's. Basically he had the vocalists in correct phase and the backing instruments out ophase. What this did was to make the intruments sound 'big' as the out of phase recording smears the the fine detail and creates a bigger subjective image. The voices remain in sharp focus however, and give the overall tune a nice presence.
In the film commentary of the Commitments, Alan Parker does the same thing. In order to showcase the vocalists he plays back the backing instruments out of phase, while keeping the voice in correct phase.
In the case of the Sheffield labs disc. I think they just screwed up....

I ran across some excellent reading material on the subject of this thread.
Check out the links provided in the article.
Audiodir, are you sure of the reversed polarity in Phil Spector's "Wall of Sound"? I was told that the effect was done by cramming everyone into a way too small studio. The instruments and vocals were not isolated to the individual mics, but bled into surrounding mics. Then he did his final mix down on a pair of 6" x 9" car speakers. Not much for high fidelity, but a kick ass mix for the AM mono car radios and cheap ass dinky transistor radios of the day. But I never remember anything about a polarity reverse.

BTW, if you want to hear how bad the "Wall of Sound" plays on an audiophile system, listen to the Ramones song "Rock 'n' Roll High School" produced by Spector using his W.O.S. technique.
Below is a link of a question I asked over on Audio Asylum.
I thought it may be of some interest here for this tread.
Tvad, I believe I have the answer....

The David I credit at the end is David Aiken an AA member.
Jea48, Bingo! That's what I thought...adding another amp into the signal chain would have degraded the sound.

Nice research. Now you can sleep at night.