Thursday, May 3, 2012

Showdown at the Perfect Polarity Pundit Corral

Above is a picture of a copy of my official Polarity Test 2012 disc (having only the 20 minutes of positive polarity pulses) played back on an Adcom CD player in default polarity mode.  Despite claims by my longtime friend George S. Louis, the Perfect Polarity Pundit, that the Adcom GCD-575 and the NEC CD-730 have opposite output polarity in their respective default normal polarity modes, results identical to the above were obtained in all the tests I performed on Wednesday and Thursday on both players.  And despite his claims that the EZDup Cool Copy CD-6198 inverts audio polarity in copied discs at 16x, in our test on Wednesday night, the copied disc also showed identical results.

The display program was SignalScope Pro demo version (I have mixed feelings about this product, but the polarity test I created was deliberately created to work even with scopes that don't trigger well) running on a 13" Macbook Pro.

We also used the polarity test track on the CBS Test Disc CD-1 and found that the first pulse was positive, and the second pulse negative, on both players in their default polarity modes.  George and others had claimed that the CBS Test Disc was recorded with a negative pulse first.  Presented with this information, Stan Ricker later claimed the intial disc was incorrect but that CBS corrected the disc in later production.  We also made a copy of that disc on the EZDup, and we agreed that it showed the correct polarity on the Adcom and NEC players.  The pulse on the CBS test disc is extremely hard to use.  The track lasts 60 seconds and has merely one up pulse (at 17 seconds in) and one down pulse near the end.   Often, the pulse occurred in the interval between one scan and the next.  But we did these tests over and over until we saw at least one of the two pulses, and on most of the critical tests we repeated the track until we had seen both pulses.

Thanks to George who let me run these tests on his equipment and publish the results, even if they appear contrary to his claims.  He says he only seeks the truth by any and all means.

In the past few years, George, aka the Perfect Polarity Pundit, has generated a lot of online text and emails for his polarity obsession.  He claims to have found that lots of CD players and digital audio recordings themselves vary with respect to a relatively well understood factor known as audio polarity.  Any given audio system can reproduce sound either with correct or incorrect polarity as compared with the original sound.  Correct polarity is generally worth preserving, it is well known that humans can detect the effect of incorrect polarity in at least some simplified and extreme cases according to research published by leading audio scientists in the Journal of the Audio Engineering Society.  But that same published research suggests that polarity is not generally audible over loudspeakers (it's most audible with headphones) and not generally audible with musical content having more than a few instruments playing.  George, on the other hand, believes polarity to be among the most important factors in audio reproduction, and furthermore, that it is often reproduced incorrectly due to widespread polarity errors in CD playback devices, CD duplicators, and recordings themselves.  Many audiophiles have come to believe the same thing, either because of independent discovery or George's influence through his website, emails, and personal interactions with leading figures in audio.  Most of George's polarity testing has involved some kind of listening test, ears and brain being required to perform the test, though he claims some of the tests are essentially objective anyway.  I have disputed the objectivity and reliability of his polarity test methods, and have called for what I consider objective tests using unambiguous test signals and an oscilloscope.

Confronted with scope images like the above, which have so far showed no variation among several component level CD players he claimed had different output polarity, and no changes caused by CD duplication, George says he can't explain my results, and criticizes me for having a closed mind in this regard.  I do indeed see these results as being hard proof that the Adcom and NEC (and, in previous test Oppo BDP-95) players have identical polarity, and that George's finding them as having different polarity was incorrect.  And that the EZDup Cool Copy does not invert audio polarity.  George believes many if not most digital CDR duplicators invert audio polarity based on his results.  I find that idea laughable.  But he has to continue to believe that to support his whole edifice of sighted testing and  anecdotal evidence.  If particular CD duplicators do not invert as he has claimed, many of his findings would no longer fit his polarity story.

Further he claims that the mere fact that I have not yet found a polarity difference among all the players I have tested so far (7 or so, including several George believed to have different polarity) is evidence my methods may not be fully reliable.  [Update: As of Saturday night, we have now found two Memorex portable CD players that do invert audio polarity on all tests.  I have never claimed that all CD players, especially cheap portable CD players, are correct.  A few years ago I found an iPod music player to have incorrect output polarity using an earlier louder version of my test signal and a scope.]

George has years of anecdotal evidence (which he is constantly recounting) and listening tests supporting his views on the polarity of the Adcom and NEC players and many others, as well as the polarity inversion caused by the EZDup and other duplicators.  Not surprisingly even faced with exacting measurements which show the opposite, he is not (yet) willing to change his views.

Also, George has one measurement device which has often, it seems, agreed with his polarity calls.  That is the Galaxy Audio Cricket-R.  The Cricket creates a chirpy electronic pulse in it's line output which it can also read back (either through line input or microphone) to determine speaker polarity.  George has recorded a long duration of Cricket pulses on a CDR to test CD players.  He connects the output of the CD player to a line input he has created on the Cricket by plugging a XLR to RCA adapter into the Cricket XLR microphone input.  It then may show either a green light indicating correct polarity (no inversion), red indicating incorrect polarity, or neither light.  During our use of this device, the "neither light" condition seemed to be the most common.  Faced with neither light lighting up, or a light not as George expected, he would twist the gain control until it showed the light he expected.  On several occasions, it could show a red light at one gain setting, and a green light at another.

I'm not exactly sure how the Cricket works, but even from our limited use of this device on Wednesday night, I would not consider it to be 100% consistent.  If it can be used to obtained different results than my series of positive pulses viewed on an oscilloscope, it is entirely clear to me that it is because the Cricket is inconsistent.  I suspect it was not designed to test CD players.  As it uses a low frequency pulse, it is possible that variations in low frequency response may be the cause of inconsistent results.  I suspect it has been designed to work well with testing speakers but not electronic products like CD players.  It may be confused by the extended low frequency response that CD players have compared with most speakers.  Here is what the Cricket electronic chirp looks like:

But George remains puzzled.  He never stops thinking or arguing how his polarity calls must be correct, implying my tests must be wrong, but he won't say that my tests must be wrong, he just doesn't know.  He cannot believe the Cricket would show incorrect results, especially when those results are consistent with his listening tests.  He continued arguing with me about these results continually from Wednesday until Saturday night, when after some tests on a second polarity inverting Memorex portable player I told him I did not want to hear any more of it.


  1. Hello.. I do believe what your article states is that you do not detect polarity reversal. Is that correct? It might be that in some systems where the tweeters and mid/woofers at the crossover are wired out of phase to begin with, that testing may be a bit of a problem. Best, it requires a phase coherent speaker to begin with. Now, if you have a phase coherent system? Then I wonder why it can not be detected. I am a musician and can readily hear a difference on a system with a 1st order crossover. If the subwoofer is out of polarity with the main speakers, that also may cause a blurring of what's to be looked for.

  2. For years I have heard a distinct difference on most discs when flipping the polarity switch on my DAC. I took the trouble to mark each jewel case with a zero or inverted insignia. Then my transport died after 12 years. The new transport is apparently of opposite polarity (unless it's the way its American XLR mates with the British DAC) because all the discs now sound better the other way. Maybe it's just some sonic effect that the switch circuitry has and isn't polarity at all, but it is interesting that it's now reversed.

  3. As one of the commenters stated above... You need phase coherent speakers to make a just evaluation. Its really silly to argue over a factor that depends upon a speaker that plays all frequencies moving in harmony. To me, reverse polarity is like what happens when looking through binoculars from the wrong end. Not as magnified in effect, but the image is either moving you into the music, or pulls you away. Its a silly argument for those hear what is obvious to them. I use Audience 1+1 speakers. You can not get more phase coherent than that. And, it was always obvious on any speaker I listened to with a first order crossover. Many speakers wire the tweeter 180 degrees out of phase with the mid-woofer.