argh! it's polarity not phase!

[Mich]

A lot of people seem to be confused.
When you talk about a phase shift it refers to shifting the phase of the individual sine components of a signal and this is NOT a time shift (like many people here think) in which you simple shift the signal in time.
While a phase shift on a _single_ sine wave is equal to a time shift, this is not true for combinations of sine waves.
To cut a long story short: http://en.wikipedia.org/wiki/Phase_inversion: "Phase inversion means the swapping of the two poles of an alternating current source. A phase inversion is neither a time shift nor a phase shift, but simply a swap of plus and minus."

[timlloyd]

Originally Posted by Mich

A lot of people seem to be confused.
When you talk about a phase shift it refers to shifting the phase of the individual sine components of a signal and this is NOT a time shift (like many people here think) in which you simple shift the signal in time.
While a phase shift on a _single_ sine wave is equal to a time shift, this is not true for combinations of sine waves.
To cut a long story short: http://en.wikipedia.org/wiki/Phase_inversion: "Phase inversion means the swapping of the two poles of an alternating current source. A phase inversion is neither a time shift nor a phase shift, but simply a swap of plus and minus."

Not confused here. A phase shift is a time shift relative to frequency. The difference between phase shift and a simple time shift, is that phase is related to frequency, while a simple time shift is not.

For example: a time delay of 0.5milliseconds would equate to a 180degree phase shift only at 1kHz.

What you might be thinking of is a "phase shifter" effect. This is created by using an LFO to control the time delay of one out of two parallel identical signals via an all-pass filter. This creates a comb filter effect, with the variable constructive and destructive nodes of interference - this has absolutely nothing to do with polarity.

It is a misnomer to say that swapping of the two poles of an alternating current source is a phase inversion. This process is actually called polarity inversion, and does not affect phase.

This isn't a complicated thing at all. I keep having to LINK to articles that explain explicitly because it seems I'm being misunderstood.

[Ollie]

Originally Posted by Mich

A lot of people seem to be confused.

Isn't "bi-curious" the proper term in this context?

[timlloyd]

Is this the "evolution of language" manifesting? So phase and polarity now mean something different when referring to single-frequency periodic signals, than when referring to complex signals with multiple partials? Why should that be?

Now a physicist and an audio engineer can't understand one-another without wasting time in a clarification of terminology that should and did and does actually have standard definitions.

It's confusing and unnecessary.

[ubergod]

To help myself better understand this, I applied a 180 degree phase shift (using the included JS plugin) to a 100Hz sines wave. This is what came out... not sure if it's what's *suppose* to come out, maybe it's a bug/glitch in the software? Anyways, interesting nevertheless, as I was expecting it to be a perfect "flip" so-to-speak:



zoomed out, notice the "glitch" is only at the beginning:


..and again, I'm not against changing the button name in reaper, just trying to understand this stuff better.

[ubergod]

Here it is again with a 20Hz sinewave. The "glitch" happens at the exact same time as 1000Hz tone, between 1664 and 1665 samples (or 0.017 seconds) into the start of the clip.



again, not sure what that means if anything, but apart from that, the phase shift of 180 degrees results in the same thing as a polarity reversal *ducks*

[ubergod]

OK, nevermind, I figured it out, seems to have to do with the FTT size. Here is is when a size of 1024, where before it was 4096:

[timlloyd]

Originally Posted by ubergod

the phase shift of 180 degrees results in the same thing as a polarity reversal *ducks*

lol

You're still missing the point slightly. It only results in the same thing when you're dealing with a single frequency, and you don't look at where the signal starts and ends.

So they only produce the same result when referring to ONE frequency of infinite length..........in other words, they are not the same and do not produce the same results

This might help you out.

[JHughes]

Originally Posted by Mich

"Phase inversion means the swapping of the two poles of an alternating current source. A phase inversion is neither a time shift nor a phase shift, but simply a swap of plus and minus."

This is correct for AC generators. The generator rotates through 360 degrees, and power is tapped at 0, 120, and 240 degrees. Each of these taps is called a phase, and this is called 3 phase power. By tapping the generator out of phase, the requirements of the ground return are greatly reduced, which reduces cost and weight enormously. When you set up a large concert sound system you have to balance the load across these phases or things commence to melting. In a nightclub, everything "should" already be on the same phase.

The term phase inversion was obviously borrowed from the power generation industry. In their context the term is correct, because each tap of the generator is called a phase by definition, and when they are swapped they could be thought of as being inverted. In the rest of electronics including the audio field, it certainly is not correct and is an obsolete convention that should be eliminated. Do smart people use the term incorrectly? Absolutely. Maybe even Rupert Neve says phase instead of polarity. But if he does it's only to avoid discussions like this one! He knows which is correct.

I don't find the subject complicated at all, but then I've actually seen this work with my oscilloscope and my background is in electronics. Regardless, at least ONE of the articles linked by timlloyd should make this perfectly clear to anyone.

This could be the "evolution of language" manifesting. In the same way that saying "I could care less" actually means you couldn't. Arrgh.

[timlloyd]

Originally Posted by JHughes

In the same way that saying "I could care less" actually means you couldn't. Arrgh.

I wrote almost that exact thing a while ago in a similar phase/polarity argument on gearslutz............I need to get out more.

[JHughes]

Let's start a support group...you can come up with the catchy title whose acronym spells something clever.

[Mich]

Originally Posted by timlloyd

Not confused here. A phase shift is a time shift relative to frequency. The difference between phase shift and a simple time shift, is that phase is related to frequency, while a simple time shift is not.
For example: a time delay of 0.5milliseconds would equate to a 180degree phase shift only at 1kHz.

A phase shift is not a time shift relative to a frequency, that is only true in the special case where you only have one sine wave.

Every signal can be represented by a number of sine wave (for digital this number is finite for analog infinite).
Let's take the signal S1 with f=1kHz:
S1 = sin(2*pi*1000*t)
Now let's time shift it by 5e-4s:
S1' = sin(2*pi*1000*(t+5e-4)) = sin(2*pi*1000*t + 2*pi*0.5)
Now let's phase shift it by p = 180 deg = 2*pi*0.5:
S1'' = sin(2*pi*1000*t + 2*pi*0.5)
... see they both turn out the same way.

Now lets say we have a signal S1 with f1 = 1kHz and f2 = 2kHz:
S2 = sin(2*pi*1000*t) + sin(2*pi*2000*t)
Now time shift that by 5e-4s:
S2' = sin(2*pi*1000*(t+5e-4)) + sin(2*pi*2000*(t+5e-4)) = sin(2*pi*1000*t + 2*pi*0.5) + sin(2*pi*2000*t + 2*pi*1)
Notice how in this case the phase shift for the 2kHz wave is 2*pi = 360 deg?
Now lets do the phase shift of 180 deg = 2*pi*0.5 just for completeness sake:
S2'' = sin(2*pi*1000*t + 2*pi*0.5) + sin(2*pi*2000*t + 2*pi*0.5)

Also -S2 = S2'' and -S1 = S1''. Just to show that it is identical to polar inversion, flipping the waveform, switching the sign of the waveform, etc...

It is a misnomer to say that swapping of the two poles of an alternating current source is a phase inversion. This process is actually called polarity inversion, and does not affect phase.

By my above example I beg to disagree. Until proven otherwise.

[ubergod]

I realize I may be veering off topic with this post, but it was fun to put together and thought someone else might find it interesting. The second track is the PhaseBug plugin, which can be found here:

http://www.betabugsaudio.com/plugs.php

compared to the included JS phase_adjust plugin, rotating around and around. I was going to also do the same thing with actual content (becasue a sine wave isn't the best example) but, whatever.... now I'm bored with the topic all together :P

[img]http://img145.**************/img145/1775/wham.gif[/img]

timlloyd, thanks for your patients, I read all your links and understand them all. I guess my questions are more to do with these plugins then the actual theory behind them.

edit: here's the 180 degree rotation of actual content (not just sine waves) with both the PhaseBug and phase_adjust plugins:



In the case of PhaseBug, it doesn't null completely, but with phase_adjust, it does except for the very beginning. Fun stuff

[JHughes]

@mich, Yes, we know that pi radians=180 degrees, but 2pi radians for 1khz and 2khz are not equivalent in the time domain.

[Mich]

Originally Posted by JHughes

@mich, Yes, we know that pi radians=180 degrees, but 2pi radians for 1khz and 2khz are not equivalent in the time domain.

Yes a phase shift of 2*pi for 1kHz and 2kHz results in different time domain shifts for each frequency. (I already said that a phase shift is not a time shift multiple times.) So what is your point.

See in the old analog days people might have been stuck with a time shift and the link Tim posted shows this it totally disregards the following which in practice is impossible in analog but totally possible in theory and in digital.
Without wanting to overly repeat myself any signal can be represented as a sum of a number of sine waves. For analog this number if infinite, thus the following is impractical, but for digital this number is finite. Now with that knowledge all you have to do is split the signal in its sine components then add the phase offset of 2*pi to each sine wave and recombine them. In fact this is what the plugins that ubergod posted so many pictures of do. They take the DFT (decompose the signal in its sine wave components) thus they get a magnitude and phase of each of the sine wave components. They then add 2*pi to each sine waves phase component. After that they use the inverse DFT to recombine the sine waves resulting in a phase inverted version or phase reversed or polarity reversed or whatever you want to call it version.
So where exactly do you think that my reasoning is wrong? Do you need mathematical prove?

[Dstruct]

Let's just rename the button(tooltips) and get rid of the discussion. Would make REAPER the first DAW which has it labeled properly.

[Mercado_Negro]

name it whatever you like, I know what it does so I wouldn't care if it's changed lol

EDIT: hmmm actually, now that I think about this... I never read that label before today

[Lawrence]

Tim said it way back, audio is bi-polar. There are unipolar digital formats however, but they would have no polarity to switch in the digital domain anyway.

You can still adjust the relative phase of a unipolar signal, the timing, but you can't reverse the polarity since it has no opposite polarity voltage.

Unipolar:



Bi-polar: The analogy relates more to how audio waves behave. Not sure which they use in daws - or across a firewire cable or similar to streams 0's and 1' - but it's certainly bi-polar at the ADC/DAC, the voltages. At some point it has to be.



Our audio editors reflect that. The mid point is "zero voltage".

[EpicSounds]

I would also like to see it labeled polarity.

[karbomusic]

Just put a "P" on it, call it what you want and make everybody happy.

[Fabian]

Originally Posted by karbomusic

Just put a "P" on it, call it what you want and make everybody happy.

Ingenious! The Gordian knot untied...
I vote for this

[timlloyd]

@Mich, we're now almost talking about different concepts, and missing the point somewhat in doing so. The point is, the polarity invert button on a mixing console, mic preamp, DI, DAW channel......etc......is not using an analysis/re-synthesis Fourier method to alter the phase angle of all the input frequencies.

All it's doing is swapping the poles, thus inverting the polarity of the signal. Since that is what it's doing, that is what it should be labelled as, end of discussion imho.

I'm persnickety, and as such am not "down" with just labelling it "P" and letting these confusions remain.

As long as it's not represented by the letter phi or theta...............maybe a yin-yang symbol is more appropriate

[karbomusic]

I'm persnickety

I never would have picked up on that.

I really don't see the fascination with arguing what the button does vs what the button appears to fix. If I push it, it's probably because something sounds out of PHASE with something else... relatively speaking of course. How it fixes what sounds OUT OF PHASE really doesn't matter to the end user.

I have a volume dial on my guitar, not an variable resistance dial. I turn it to make the guitar sound louder/softer. Making minutia out of it is silly, not just picky per se.

Karbo

[timlloyd]

I can't be bothered to discuss this anymore

Cockos will do whatever they want, which in this case will probably be to ignore this thread completely and keep on conforming to the industry-standard-misnomer.

[JHughes]

Well said again timlloyd.

karbomusic, your analogy does not hold. A volume dial does in fact adjust volume, but a "phase reverse" switch does not reverse phase. That's the whole point!

(to no one in particular) My god, this discussion sounds like a talk with some dumbass junior high school dropout. "I don't need to know no english. What good is math, it never made me no money." Sheesh. If someone can't be convinced about the need to understand nuance then fine, but to argue that definitions don't matter is really putting that ignorance out there for all the world to see.

Words have meaning, and they do matter. Otherwise every sentence we say could be a grunt followed by a "you know?". True there are those who talk that way already but they are not saying anything meaningful.

This is not only an artistic but a technical pursuit and should be treated as such. There is no arguing with those who think what the hell, it's only a word. Those people say that because, to them, it doesn't matter, after all, they know what button to push even if they don't really know what it does. Well fine, be an ignorant grunting button pusher. I will not. And neither will Bruce Swedein or Al Schmitt or Alan Sides or any of the other greats. I guarantee they know what the difference is between polarity and phase. And if they want to reverse polarity, they do it with the polarity button. If they want to adjust phase, they do that too, but with a proper tool. http://www.littlelabs.com/ibp.html.

Hmm, what does that page say? "Puts you in phase when the polarity button won't". Looks like someone else thinks there's a difference too. The misguided soul should read this topic so he can learn that there is no difference, polarity and phase are the same, even if you have to resort to mathematical tricks to prove it. Reminds me of the mathematical proof that 1+1 does not equal two.

When someone asks me "do you know what all those buttons do?" I can honestly say yes.

Can you?

[karbomusic]

Originally Posted by JHughes

Well said again timlloyd.

karbomusic, your analogy does not hold. A volume dial does in fact adjust volume, but a "phase reverse" switch does not reverse phase. That's the whole point!

I could really care less what it gets called and I don't disagree with what happens when it is pushed. Being willing to waste this much screen space over the term in lieu of the usage because you prefer to call those who use the term a dumb ass instead of giving credit for why gets called what it does. Call it the pos/neg bit flipper for all I care. It's minutia for many end-users!

Most everyone knows what it does (or should that be called "how it operates"), the habit of the name phase is due to what it solves not how it does it. Again, I don't care what it get's called but Tim's comment that even a "p" is not good enough for the polarity international standard gods was a little over the top so I responded in kind. You know better, people don't hear a polarity problem and that's why it gets called what it does regardless of its standards legality... I agree but too uptight in this thread if you ask me



Karbo

[timlloyd]

My nitpicky-ness forces me to point out that what you mean is, in fact "I couldn't really care less." Saying "I could care less" means that you do care to some extent at least, which is an oddly round-about way of expressing that sentiment. Minutia isn't it? Maybe, but it's completely changed what you meant to express into almost the opposite statement.



Waste of screen space? lol If you want to look at it that way I can't stop you. You could just not participate if you really don't care for the discussion.

Evidently not everybody knows what it does........because it's labelled incorrectly and that inevitably causes confusion.

I'm with JHughes, I like knowing exactly what I'm doing when I make a decision; and I didn't call anyone a dumb-ass for not sharing that personality trait.

[karbomusic]

Originally Posted by timlloyd

My nitpicky-ness forces me to point out that what you mean is, in fact "I couldn't really care less." Saying "I could care less" means that you do care to some extent at least.



Waste of screen space? lol If you want to look at it that way I can't stop you. You could just not participate if you really don't care for the discussion.

Evidently not everybody knows what it does........because it's labelled incorrectly and that inevitably causes confusion.

Well I made it a good while didn't I? Yes, I could care less and if so I would never have posted. Then again, there comes a point where ya know what the other means and it's OK to just get the darn job done.



Karbo

[Mich]

Originally Posted by JHughes

If they want to adjust phase, they do that too, but with a proper tool. http://www.littlelabs.com/ibp.html.

... and what does the phase invert button do on this proper tool you linked to?

Also what exactly does this tool do? Is it a simple delay? Or does it really adjust the phase like the plugins in the numerous posts by ubergod?

Originally Posted by JHughes

Reminds me of the mathematical proof that 1+1 does not equal two.

And that would do like ... ?
Here is the prove for 1+1=2: http://mathforum.org/library/drmath/view/51551.html
Prove it wrong

[karbomusic]

Evidently not everybody knows what it does........because it's labelled incorrectly and that inevitably causes confusion.

I missed that part of the thread I guess. What confusion? Something sounds out of phase, I push the button. In practical use, how it achieves what achieves it may not always be most important thing to the user. I might be bothered more if it were called the mute button; now that's confusion. That's what I mean by minutia, no malice intended; call it what you want. Personally, I think it would be most helpful to those who use it to label it polarity/phase to cover both its tecnical doings and its percieved result.

Karbo

[Mich]

Originally Posted by JHughes

The misguided soul should read this topic so he can learn that there is no difference, polarity and phase are the same[...]

Err... if polarity and phase are the same than it shouldn't matter whether the button is labeled polarity invert or phase invert.

[timlloyd]

Originally Posted by karbomusic

Something sounds out of phase, I push the button.

Yes, and while it might sound better with the button pressed, don't be fooled into thinking that it's now in phase!! It might not be!

Of course no malice intended on my end either, it's just cyber-banter really

Originally Posted by karbomusic

Then again, there comes a point where ya know what the other means and it's OK to just get the darn job done.

Yep, which is precisely why I don't try and have this discussion with people in the middle of a session!

[karbomusic]

Yep, which is precisely why I don't try and have this discussion with people in the middle of a session!

LOL, good one.

[Kundalinguist]

To me, it's like sexual intercourse - call it what you want, as long as it works as expected.

[timlloyd]

Originally Posted by Mich

... and what does the phase invert button do on this proper tool you linked to?

Maybe it inverts polarity as it states in the manual. For some reason I cannot fully reconcile, it then references degrees, which as I've said before, kind of doesn't make sense because polarity doesn't require a coefficient of degrees.

Whoever wrote that manual, also makes the golden "could/couldn't care less" mistake Maybe I'm the only one who consistently finds that amusing.

[karbomusic]

Originally Posted by timlloyd

Whoever wrote that manual, also makes the golden "could/couldn't care less" mistake Maybe I'm the only one who consistently finds that amusing.

Well...

“I could care less” is one of those idiomatic expressions, particularly in American English, that doesn’t necessarily mean what it says. There are numerous suggestions for the origin of the phrase. The most recent of these is that “I could care less” is a corruption of the term “I couldn’t care less,” possibly first used in the UK in the 1940s. By the 1960s, Americans had adopted “I could care less.” Was it laziness, poor hearing or deliberate irony?

http://www.wisegeek.com/what-does-i-...-less-mean.htm

[timlloyd]

Ahh, it's the evolution of language thing again; and I use the word evolution quite wrongly, because evolution implies that the new is better than the old

(hey, we are in the nitpicks sub-forum after all )

[karbomusic]

Originally Posted by timlloyd

(hey, we are in the nitpicks sub-forum after all )

Well slap me like a red-headed stepchild... I did not know that... dying hair now.. Damn New Posts link...

Karbo

[beatbybit]

ok.. fix that tooltip, but



leave these as is..





(nuances, heh)

[timlloyd]

lol good catch!

obviously needs fixing too