Absolute polarity discrepancy in digital media

[Grayson Dere]

I recently came across a VERY interesting article from the website: www.audiogeorge.com

There is an article called, 'How to play most, if not all, digital media in absolute polarity.'
http://www.audiogeorge.com/how-to-play-most-if-not-all-digital-media-in-absolute-polarity/


In the article the author claims that:

"All tracks on approximately 92% of compact discs (CDs) play in inverted polarity on approximately 92% of compact disc players (CD players or servers that they’ve been ripped to) which means polarity mistakes aren’t random."

Further down the article it reads:

"Playing music in inverted polarity rather than in absolute polarity (aka absolute phase) makes the sound brighter, harsher, more congested, less 3-dimentional sounding, and in general less musically and emotionally involving."

Quickly after reading this portion of the article a lightbulb went off in my head...is THIS the reason why so many people prefer vinyl playback compared to digital media!!?? Have I actually found the scientific explanation to this decades long argument in vinyl vs. digital sound quality?

I am very curious as to what others have to comment on regarding this polarity phenomenon in recorded music. I spoke at length with a friend of mine that is a recording engineer and he claims with all the technology we have today at our disposal for recording purposes polarity issues should never come up...but obviously it still does.

As an example: Miles Davis' Kind of Blue supposedly plays in reversed polarity on 92% of CD players.
(at the bottom of mentioned article there is a list of recordings with polarity notes beside them)

Just to further open the can of worms, what about our blu-ray discs/players? are we not listening to our movies as they were originally intended?


As always, thanks for any and all comments!

 

[dc2bluelight]

I recently came across a VERY interesting article from the website: www.audiogeorge.com

There is an article called, 'How to play most, if not all, digital media in absolute polarity.'
http://www.audiogeorge.com/how-to-play-most-if-not-all-digital-media-in-absolute-polarity/


In the article the author claims that:

"All tracks on approximately 92% of compact discs (CDs) play in inverted polarity on approximately 92% of compact disc players (CD players or servers that they’ve been ripped to) which means polarity mistakes aren’t random."

The article is 92% hogwash. His sole determination of "polarity" is subjective. In fact, he has absolutely no idea what's correct. The CD production chain does not include a polarity inversion by default. In fact, in the early days this is something we did look at. The first (and even later) test CDs had polarity test tracks based on the established standards. There's no way 92% of all CD tracks play inverted, unless there's an inversion elsewhere in the system.

Further down the article it reads:

"Playing music in inverted polarity rather than in absolute polarity (aka absolute phase) makes the sound brighter, harsher, more congested, less 3-dimentional sounding, and in general less musically and emotionally involving."

Again, hogwash.

The audibility of polarity demands a rather significant asymmetry to the signal. The first assumption in error is that all music, or at least most music is fundamentally asymmetrical in nature. Simply not true. There are asymmetrical instruments of course, some fairly significantly, certainly not all, or even most.

So there are big issues just determining what's inverted and what's not. As an example, here are two samples of trumpets:

Note how the peak asymmetry is strongly negative.

Note in this sample how the highest amplitude peaks are asymmetrically positive.

The thing is, these are both samples from the same instrument played by the same musician in the same recording less than seconds apart. What's right? What's inverted? You can't tell by looking, and you sure can't tell by listening.

The only way to determine absolute polarity is with a very tightly controlled and known test signal. Sorry, listening to instruments doesn't count...at all.

Quickly after reading this portion of the article a lightbulb went off in my head...is THIS the reason why so many people prefer vinyl playback compared to digital media!!?? Have I actually found the scientific explanation to this decades long argument in vinyl vs. digital sound quality?

Nope. The reason is simple: whereas there actually IS a standard for polarity in microphones, mixing desks, and digital recording systems, there was no default standard for analog recording and reproducing devices, and electronics. (edit: rethinking this, not totally true, there is a standard for record grove modulation, so the lathe is probably correct and consistent. Not so true of tape recorders, especially with multi-generational mix-downs.)The chances of vinyl being "right" are far fewer than in the digital chain. Then we get to the multi-driver speaker in a room...good grief.

I am very curious as to what others have to comment on regarding this polarity phenomenon in recorded music. I spoke at length with a friend of mine that is a recording engineer and he claims with all the technology we have today at our disposal for recording purposes polarity issues should never come up...but obviously it still does.

But it only comes up in the context of someone's opinion of what "right" or "inverted" is based on subjective listening. Doubtful a large group of listeners on a large cross-section of systems would arrive at anything other than random results.

As an example: Miles Davis' Kind of Blue supposedly plays in reversed polarity on 92% of CD players.
(at the bottom of mentioned article there is a list of recordings with polarity notes beside them)

But there's no proof of it. There's no reference, only subjective opinion. Nobody making these claims was in that studio...and there goes the reference.

Just to further open the can of worms, what about our blu-ray discs/players? are we not listening to our movies as they were originally intended?

Yes, we are. And it doesn't matter anyway.

As always, thanks for any and all comments!

The polarity discussion keeps coming up, always introduced by people with no idea of what that actually means and how the entire audio chain, analog or digital, works. It's always someone with a polarity switch who thinks their preference is absolutely correct. What if their speakers are inverted? And don't think THAT hasn't happened before.

 

[Grayson Dere]

dc2bluelight,

Thank you for your insight! You have knowledge of recording processes and that helps tremendously. Your two graphs that show the trumpet peak asymmetry waveforms is interesting...I don't think I would be able to listen to them and tell which was inverted or not : ) That also begs the question, how is he able to tell what songs are being played inverted? Perhaps I need to have a one on one discussion with the article's author. He would be up for some serious scientific awards if he actually could pick out 100% by ear which songs on various albums were indeed played in inverted polarity.

Some audio experts I know say that when the polarity is correct the stereo imaging is more 'focused' in the middle between the speakers, and when it is inverted the imaging becomes less precise as if the singer was no longer right there in front of you in the phantom image. Would this be a good possible indication of absolute polarity vs. inverted polarity in playback as now we are not listening for individual instruments but sound perception characteristics?

Thanks!

 

[dc2bluelight]

dc2bluelight,

Thank you for your insight! You have knowledge of recording processes and that helps tremendously. Your two graphs that show the trumpet peak asymmetry waveforms is interesting...I don't think I would be able to listen to them and tell which was inverted or not : ) That also begs the question, how is he able to tell what songs are being played inverted?

An instantaneous polarity flip is often (but not always) easy to hear. But hearing a difference is not the same as determining what is correct (which is subjective, and likely random).

Perhaps I need to have a one on one discussion with the article's author. He would be up for some serious scientific awards if he actually could pick out 100% by ear which songs on various albums were indeed played in inverted polarity.

Oh, I wouldn't. He's passionately biased, there's nothing scientific about his opinions. Please be aware that there's no way to confirm absolute polarity without a reference, which is almost universally unavailable.

Some audio experts I know say that when the polarity is correct the stereo imaging is more 'focused' in the middle between the speakers, and when it is inverted the imaging becomes less precise as if the singer was no longer right there in front of you in the phantom image. Would this be a good possible indication of absolute polarity vs. inverted polarity in playback as now we are not listening for individual instruments but sound perception characteristics?

Be very careful about opinionated "experts" who use subjective opinion as their sole basis for judgement. There are scientific ways to research this, but you will typically not find any referenced.

The way spatial hearing works is independent of polarity. Though a polarity inversion of a recording may alter certain aspects of spatial perception (again, remember there's no "correct" without an absolute reference), absolute polarity is not part of the spatial hearing mechanism, nor could it be given the random asymmetry in a acoustic stimulus.

Remember too that two-channel stereo does not present any accurate spatial information, it's all a "fake", made to satisfy the need for an acceptable representation of a sonic event which may never have existed at all. Even minimalist acoustic recordings do not (and cannot) represent the event well spatially, though the result may be completely satisfying to the listener. The acoustic signal, in that case, is likely to be of completely random asymmetry, with the inclusion of a multitude of acoustic reflections and reverberation.

The ability to reliably discern polarity inversions depends on the degree of waveform asymmetry that is not masked by random asymmetry of acoustic reflections (real or simulated). A very dry, very asymmetrical signal would be the most reliably detectable, like certain (but not all) low male voices in a dry voice booth. A male voice with a more harsh "edge" will likely contain more asymmetry than a smooth round full one.

 

[dc2bluelight]

I'll just add this anecdote regarding absolute polarity. Several decades ago I worked with a radio personality with whom I recorded many interviews. The basic waveform of his voice was quite asymmetrical. His voice was edgy, almost raspy in person and in recordings. If you listened to a polarity-correct recording then inverted polarity, his voice took on a much warmer, more pleasant character. It was more pleasant to listen to inverted than correct. Now, if you never met the guy and listened, you'd likely probably select the inverted version because he sounded better that way. But in absolute polarity terms, you'd be making the wrong choice.

 

[Grayson Dere]

I'll just add this anecdote regarding absolute polarity. Several decades ago I worked with a radio personality with whom I recorded many interviews. The basic waveform of his voice was quite asymmetrical. His voice was edgy, almost raspy in person and in recordings. If you listened to a polarity-correct recording then inverted polarity, his voice took on a much warmer, more pleasant character. It was more pleasant to listen to inverted than correct. Now, if you never met the guy and listened, you'd likely probably select the inverted version because he sounded better that way. But in absolute polarity terms, you'd be making the wrong choice.

I really like what you said here. Even though something sounds more pleasant depending on the polarity of the recording it doesn't mean it's the 'right' way. We really would need a reference to determine what is correct, as you said previously.
I guess, for now, it's all subjective listening that this whole polarity discrepancy is based off of...which unfortunately doesn't do us any good.

If I find anymore interesting thoughts on this topic I'll continue to post : )

 

[AudiocRaver]

dc2bluelight covered the topic pretty thoroughly. I agree that reversing the phase of a recording makes no audible difference. It might be true that certain recordings contain enough asymmetrical information that a difference could be discernible, but I have not heard it myself. It is easy to set up an experiment that allows instant phase switching of both channels while listening to a recording. Having tried this, there is simply no difference to be heard.

It would be an embarrassingly amateurish mistake to only change the phase of one channel and not the other, and that would certainly be audible in any properly set up system. So many systems are poorly set up, however, that it is actually not so easy to tell if one channel is out of phase relative to the other. Still, we should assume for this discussion that the phase of both channels together is in question, not one channel relative to the other.

"Playing music in inverted polarity rather than in absolute polarity (aka absolute phase) makes the sound brighter, harsher, more congested, less 3-dimentional sounding, and in general less musically and emotionally involving."

I am a believer that a well-trained ear can sometimes pick out details in audio reproduction that most others overlook. But in this case, I am highly doubtful that the experience described is at all common. Other than a very limited number of recordings that contain and preserve asymmetry that might be audible to a very limited number of well-trained listeners, anyone who thinks inverted phase of both channels together makes the sound brighter, harsher, more congested, less 3-dimensional, is fooling themselves, or being fooled. Sadly the number of listeners and even "experts" who are into the fooling/being-fooled game is very large.

 

[dc2bluelight]

I don't mean to imply that a polarity flip in both channels is always inaudible. That's just not true, and I'll be happy to post examples where the flip is audible if you like, and under what listening conditions it is clearly detectable. But to imply, as the quoted article does, that nearly all CDs and/or CD players are by default inverted is ludicrous because there's no way to tell what is and is not inverted, even if a difference is heard when an inversion is forced. That's my point. There's no way to tell what is inverted and what is not, there is certainly no reference. And the quoted article implies that non-inverted always results in better sound in some abstract way. That's wrong too, and easily demonstrable.

There are some strange psychoacoustic effects involved in a polarity flip, though. As an example, I have an asymmetric test tone signal where flipping polarity sounds like a change in pitch, though that clearly is not happening. So polarity can occasionally be audible in odd ways. But that same signal, tested and flipped under different conditions, reveals nearly zero audible difference. So I do have a problem with the idea of polarity being clearly audible on some very high percentage of recordings. That's just not possible. To add to the circle of confusion, there are many aspects of sound reproduction that tend to effectively scramble asymmetry, and that's not being factored into the article either.

However, preference is just that, and if someone prefers an inversion for some reason, I have no problem with them choosing it over non-inverted or vice-versa. Imposing one's preference or opinions on everyone else by stating they are definitive and correct, especially without supporting evidence, is something I do have an issue with, and that's what's going on in that article. That's why I recommend it be ignored.

edit: oh yeah, flipping polarity on one channel...that's way different, and with a very simple test signal, is clearly audible to anyone. Try a recording with a center-panned bass line. I hear out of phase speakers in stores and restaurants all the time, bugs me silly.

 

[AudiocRaver]

Good points. I will add to my previous comments that the INSTANTANEOUS switch of polarity could very well be audible for a number of reasons. What I should have been clearer about is that I doubt it is likely to hear a difference between the two beyond that instantaneous switching moment, that I doubt that one polarity sounds brighter or harsher or whatever, as asserted in the article referenced. Beg pardon for any confusion.

I agree with your other points.

 

[dc2bluelight]

This gets to the heart of the matter. We know his tests were fully sighted because he derives a preference. We can't assume the switch was instant because he refers to it as being in the digital domain, so it could be either instant or delayed. But no matter, the test is fully sighted and biased, and expresses his subjective preference for a particular polarity. It's about as unscientific as it gets. His speakers are of his own design, stated as being "minimum phase", but that's a whole separate discussion, and must include the room unless he is completely in the near field, none of which is defined in the article.

You are completely correct that in instant polarity reversal can be highly audible for a number of reasons. In fact, that phenomenon can be used to demonstrate the extremely short auditory memory involved in an ABX/DBT. Instant switching can reveal a clear change, but a switching delay of more than a few hundred milliseconds reduces the resolution of such a comparative test very significantly.

But that's not what he's doing anyway. It's just a fully sighted/biased tally of opinion and preference. There are no controls in place, and no effort at all to remove or control tester bias. You could just as well reject the entire database as being fully biased and therefore useless for the purpose stated (which is my recommendation).

All he had to do was state his list as his personal preference for his unique listening conditions, and NOT an absolute statement of correct polarity, and he'd have been ok.

And shouldn't this thread be in Tech Talk instead of Movies/Music/TV/Streaming? I don't even know how I found it in the first place.

Edit: OK, now I understand. Just looked at some of his "products". And that's why this isn't in Tech Talk! There's no "tech" to talk about, just Blue Smoke and mirrors. It explains a lot. A direct quote: "Some products are simply enigmatic." Yup, they sure are!

 

[Grayson Dere]

I know this is an old post but I recently found out that the Audirvana streaming app/software actually has an audio setting you can manually click on and off to hear for yourself if you prefer absolute/reversed polarity tracks. Again, it beats me as to how in the world one would know which polarity it's 'supposed' to be played in but nevertheless this should be a fun exercise : )

Just toggle Signal Polarity -> Invert Globally

 

[dc2bluelight]

I know this is an old post but I recently found out that the Audirvana streaming app/software actually has an audio setting you can manually click on and off to hear for yourself if you prefer absolute/reversed polarity tracks.

It will probably seem like picking nits, but it's not a switch that selects between absolute and reversed polarity, it just one polarity then the inverted version of it. There can be no absolute because the original polarity and possible inversions in the chain remain unknown.