argh! it's polarity not phase!

[beatbybit]

Originally Posted by timlloyd

lol good catch!

obviously needs fixing too

which one, the "invert" or the "flip"

[timlloyd]

lol........the "phase" part

In this case, "flip", "invert", "reverse" etc..... refer to something that has only two states, ie. positive and negative, ie polarity.

[beatbybit]

but, but, but...

if doing something, do it properly.

consistency is cool.

mixed terms for the same action are 'brrrrr' (which is also 'cool', but in a reversed/flipped/inverted way )


j/k.

[timlloyd]

agreed

While potentially tricky to enforce, there does need to be standardised terminology for REAPER's actions and all the actions included in the SWS bundle from different contributors.

[Mich]

Originally Posted by timlloyd

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.

Ah, good catch on the manual, according to which it seems they use allpass filters to adjust the phase (see you can also do it in analog don't need a Fourier analysis at all). So it is really adjusting the phase (in contrast to all the articles linked so far which seemed to be doing a time shift and it call it a phase shift (talking about misnomers!)).

Anyway the reason they mention degrees there is because a phase shift of 180 degrees (and I mean a phase shift and not a time delay shift) is equal to a polarity inversion. So since you turn the phase all the way round 180 degrees one could say he is inverting it. And just because polarity only has to states doesn't mean that the same thing that polarity inversion is doing can't be achieved by other means.
Take for example the mute button. It has two states on/off. However the volume fader has more than two states but it can also achieve the same thing as the mute button if you turn it all the way to down.

So you argument that the term is technically wrong isn't really plausible to me, because there is enough prove that phase inversion that is a phase shift of 180 degrees is the same thing as polarity inversion. Thus both terms should be technically correct. However one of them has the advantage to be a industrial standard the other doesn't. So why should you adopt the one that doesn't? ... Unless you still want to argue that a phase shift of 180 degrees isn't the same thing as polarity inversion, in which case you can do that ... on your own.

[timlloyd]

Originally Posted by Mich

they use allpass filters to adjust the phase (see you can also do it in analog don't need a Fourier analysis at all).

Using all-pass filters isn't the same as using Fourier analysis/re-synthesis to adjust the phase of every frequency at once.

An all-pass filter still has a centre-frequency, and cannot escape the fact that without Fourier, phase shift is still dependent on frequency. That is what the "phase center lo/hi" button on the IBP is for - it changes the frequency bandwidth where the phase adjust is primarily focussed. You have to change it depending on the type of signal you're dealing with, otherwise it wouldn't work.

//////

So now, we're discussing whether or not the term "phase invert" can make sense. I'll concede that, with a Fourier analysis/re-synthesis it can be done, but that's not what's happening with a polarity invert button is it. no.

[Mich]

Originally Posted by timlloyd

[...]with a Fourier analysis/re-synthesis it can be done, that's not what's happening with a polarity invert button is it. no.

If you want to go with what is actually happening then I think REAPER isn't inverting any polarity at all either.
One would have to investigate the code to see exactly what it does. Is it simply multiplying the signal by -1? Should the button then be named "multiply by -1". Or is it flipping the sign bit? "Flip sign" button? In fact Cockos might actually be so smart and multiply the volume by -1 to save calculations ... so "volume invert" button?

Anyway I see no rational reason to change it.

[-kiff-]

hopefully this industry misnomer will get phased out over time :lol:

I don't know if it's been mentioned, but an easy way to settle this would be to "phase invert" a signal with a DC bias. If it's really an fft with several time shifts it would keep the same bias, or otherwise (as I would predict) it would "invert" the bias

edit: now that I think about it, the fft would get rid of the bias while a polarity swap would "invert" it.

[JHughes]

Originally Posted by Mich

Anyway I see no rational reason to change it.

Then you've not read this topic. There's no rational reason not to change it, but for some reason there are always a multitude eager to defend the status quo. Those who "don't care" spend a hell of a lot of time acting like they do. It's actually those against this change picking the nits.

[Sound asleep]

Originally Posted by beatbybit

Leave the Phase invert button alone!!

lol. or i'll send the cyber police, cause i back traced it.

[Dannii]

Originally Posted by JHughes

For an XLR swap pins 2 and 3, for an unbalanced signal swap tip and sleeve...

On balanced XLR, you're correct. On unbalanced connectors this only works if the audio grounds are isolated. Swapping the hot and cold connectors on unisolated grounds will result in a short circuit and no signal.

[ubergod]

Originally Posted by timlloyd

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

So a null isn't enough proof for you?

[timlloyd]

Originally Posted by ubergod

So a null isn't enough proof for you?

It should be shouldn't it! We were talking about slightly different things; I've already agreed that I was mistaken on some points, and that a "wavelength-relative 180 degree phase shift of all frequencies simultaneously" will produce the same result as a polarity invert. But that's not quite what I was talking about when replying to you

Mich has thankfully stuck with the thread and pointed out that the term "phase invert" can make sense in the digital domain with the help of a Fourier analysis/re-synthesis (which incidentally is what you were showing with your screenshots, while I was pre-occupied with arguing).

But I'm still of the opinion that since the button is actually reversing the positive/negative poles of the input signal (as it is also on a piece of analogue gear), that it makes the most sense to call it a polarity inversion.

Originally Posted by Mich

Is it simply multiplying the signal by -1? Should the button then be named "multiply by -1". Or is it flipping the sign bit? "Flip sign" button? In fact Cockos might actually be so smart and multiply the volume by -1 to save calculations ... so "volume invert" button?

Well, if any of those things is exactly what is going on, then yes, the polarity of the signal is being inverted, so yes I do think it should be labelled polarity.

That is the last I'm going to say on this lol. I'll stop worrying about the damn button now

[SSG]

There *is* a difference between a polarity inversion and a "phase inversion" or phase change of 180º. The difference is that a phase inversion (more aptly, phase rotation) rotates the signal around it's rest voltage, whereas a polarity inversion inverts the signal around 0DC. Now, this difference is not detectable if the waveform's rest voltage is 0DC, but the minute you introduce any DC offset to the signal, the difference becomes obvious. Additionally, summing the original with it's inverted form will wipe out the offset in a polarity inversion, whereas the offset will stay after summing a wave with it's phase inversion.

The "phase inversion" button is usually indeed performing a polarity inversion, as all that required to perform it is a multiplication of the input value by -1, whereas a phase rotation of 180° requires a bit more math. But for legacy reasons, the word "phase" has become the de facto standard nomenclature for the inversion button, dating back to very early analog boards. It's inaccurate, to be sure, but it's historically consistant (unfortunately).

BTW, in general, phase does not have to involve time. Instantaneous phase rotation can happen and be made to happen. All one needs to do is take the sine function, y = sin(x) and add the amount of phase rotation in radians to the input value, y = sin(x + Θ), where "Θ" = the amount of phase rotation in radians.

But in audio we're used to thinking about phase as being dependent upon time, because we rarely, if ever, deal with instantaneous phase shifts in a single waveform; it almost always involves either a shifting of that waveform in time, or phase cancellations between two waves because of a time differential between the peaks and crests of the two waves.

For you geeks that really want to dive into this, head over to http://www.independentrecording.net/...rces/index.php and click on the "Phase and Polarity" icon for a detailed discussion of this topic, complete with plenty of pretty pictures illustrating the explanations.

G.

[JHughes]

More good stuff, thanks SSG.

[JHughes]

From http://www.moultonlabs.com/more/abou...y_reversal/P1/

If you think about it, it's obvious that there is no such thing as "out-of-phase" for a sound that has more than one frequency. As the figures above illustrate, any given delay time will yield an infinite number of phase shifts at an equally infinite number of different frequencies. So how can a sound be 180° out of phase?

Answer: it can't, unless it's a sine wave.

But, Dave, there's an out-of-phase button on my console that can be used for some killer effects, and we're not talking sine waves.

Answer: what you are referring to isn't "out-of-phase," it is "reversal of polarity."

Yeah, so how come it is called "out-of-phase?"

Answer: because when you reverse the polarity of a sine wave, the sine wave appears to have shifted in time by 180° (even though it hasn’t), and electronics engineers, like everyone else, are lazy and sloppy with their language. And, like, they know what they mean, you know?

[liteon]

a interesting subject i admit, though i am a bit worried about the overall misconception. see bellow.

Originally Posted by JHughes

From http://www.moultonlabs.com/more/abou...y_reversal/P1/

the discussion in the above post is a sort of a joke... the part "it can't, unless it's a sine wave." pretty much sums it up as far as the interview goes in regard of the amount of knowledge the questioned person possess. i'm really sorry to view such post on the internet on this subject especially considering the fact that this thread (yeah on the forum) gives a lot of results in google and is full of simply falsed information or theory that is not really understood well by the posters.

Originally Posted by SSG

...

i've seen similar explanations before.
you are missing the fact that the rest voltage of any circuit is always zero volts.
it cannot be anything else?
an AC signal riding a DC offset can be phase shifted by 180 degrees, while the DC offset is retained, however there are circuits which exist that _can_ phase shift both the DC voltage and the AC carrier by 180 degree, which of course would result the same as a "polarity invertion". i admit though that there are cases when a certain design of a "phase shifter" can act like a _short circuit_ block for DC signals.. with this isolated case in mind it does not mean the its not physically possible to phase shift a DC signal.
while it does not make much sense to integrate something that does not change over time, the DC signal can be viewed as periodic signal with an infinite period...and yes, if we look at DSP, if we run a constant amplitude signal trough a phase shifter that uses the discrete fourier transform (DFT) we will probably get a perfect 180 degree shift for ANY signal.

as far as the therm "instantaneous" goes in the "real world", lets not apply it at all unless it is considered theoretical. :-)


---

[shoyoninja]

A phase shift of 180º electrical degrees IS the same as a polarity inversion. No matter how complex or simple the signal is.

The whole misconception here comes from the idea that phase shift is always equal to a absolute time shift. While it IS true for sine waves,it is so ONLY because of the simetry, and only if they are infinite. If they are not, they will remain with the same absolute start and finish, there will be only an angular difference between them (phase). It is simply impossible to apply the same concept on complex wave forms.

In fact, phase-cancelation (which I think that everyone agrees that is a phase related phenomena :P) does not happen if there is also a time shift, it will only happen when there is a only a phase difference between the signals, if they are in anti-phase the interference will result in cancelation.

Phase is not time alignment.

[SSG]

Originally Posted by liteon

you are missing the fact that the rest voltage of any circuit is always zero volts.
it cannot be anything else?

And yet it often enough is. Without wanting to start a second semantical war, I'm not talking physical rest voltage of a physical circuit so much as what is essentially the virtual "rest voltage" of the signal.

Anyone who has been forced to work with crappy PC sound cards or with VSTi emulations of analog synths knows that virtual DC offset exists in real life signals far more often than we wish it to, and that the "inversion" switch is not changing phase, it's multiplying the voltage (virtual or real) by -1, which is a polarity inversion.

The fact that there are circuits that can manipulate the signal to provide the equivalent of either polarity or phase change is irrelevant to the definition of either one and to the definition of the "flip/invert" switch. It remains true that whether there is a DC offset in the signal or not when it reaches the "invert" switch, the switch will perform the equivalent of a polarity inversion, not a phase rotation.

as far as the therm "instantaneous" goes in the "real world", lets not apply it at all unless it is considered theoretical. :-)

Define "real world". Are you talking the analog or digital "real world"? For in the digital "real world", instantaneous phase shift is easily attainable, as it is a matter of simple mathematics.

And as for the earlier contention that one cannot phase rotate a signal more complex than a simple sine wave, that is untrue. By running the signal through a Fourier transform to break it down into the sines of the individual component frequencies, and then adding the same degree of phase rotation to each of the individual sine functions, the entire complex wave can be phase rotated by that set number of degrees.

OK, in reality, it may take some time for the computer to execute those calculations, and in that respect it's not an "instantaneous" process. But there is an impression that phase rotation *has to* be defined by an introduction of time shift into the equations and definition. While it is true that is how it manifests itself to us audio engineers 99.9% of the time, that does not mean that is the only way in which phase can be considered, and it certainly is not how it is defined.

And finally, if given that the "invert" switch is indeed a phase inversion, or that it is really only a semantic argument because the'yre really the same thing, and that an instantaneous phase rotation of a complex signal is only theoretical and not a real world phenomenon, I might point out that since the "invert" switch does do it's thing instantaneously, that at least one of the "givens" in that argument has to be wrong.

G.

[shoyoninja]

Originally Posted by SSG

And yet it often enough is. Without wanting to start a second semantical war, I'm not talking physical rest voltage of a physical circuit so much as what is essentially the virtual "rest voltage" of the signal.

Anyone who has been forced to work with crappy PC sound cards or with VSTi emulations of analog synths knows that virtual DC offset exists in real life signals far more often than we wish it to, and that the "inversion" switch is not changing phase, it's multiplying the voltage (virtual or real) by -1, which is a polarity inversion.

The fact that there are circuits that can manipulate the signal to provide the equivalent of either polarity or phase change is irrelevant to the definition of either one and to the definition of the "flip/invert" switch. It remains true that whether there is a DC offset in the signal or not when it reaches the "invert" switch, the switch will perform the equivalent of a polarity inversion, not a phase rotation.Define "real world". Are you talking the analog or digital "real world"? For in the digital "real world", instantaneous phase shift is easily attainable, as it is a matter of simple mathematics.

And as for the earlier contention that one cannot phase rotate a signal more complex than a simple sine wave, that is untrue. By running the signal through a Fourier transform to break it down into the sines of the individual component frequencies, and then adding the same degree of phase rotation to each of the individual sine functions, the entire complex wave can be phase rotated by that set number of degrees.

OK, in reality, it may take some time for the computer to execute those calculations, and in that respect it's not an "instantaneous" process. But there is an impression that phase rotation *has to* be defined by an introduction of time shift into the equations and definition. While it is true that is how it manifests itself to us audio engineers 99.9% of the time, that does not mean that is the only way in which phase can be considered, and it certainly is not how it is defined.

And finally, if given that the "invert" switch is indeed a phase inversion, or that it is really only a semantic argument because the'yre really the same thing, and that an instantaneous phase rotation of a complex signal is only theoretical and not a real world phenomenon, I might point out that since the "invert" switch does do it's thing instantaneously, that at least one of the "givens" in that argument has to be wrong.

G.

About real world examples: You can have phase rotation simply changing the orientation of the signal source, be it audio, light, whatever.

And in that regard, the phase will shift without any perceived time delay.

You see, there is no point to mix timming and phase togheter, because they are not the same thing.

Delays CAN and WILL lead to phase issues, in any kind of signal, be it audio, microwaves or light. But that does not mean that phase is a timming related quality.

To make it even clearer, lets say you have 2 loudspeakers, you plug both to the same amplifier, with the same signal. BUT in one of those, you misplaced +/- and inverted its polarity.

Sure, the electric signal is now polarity "flipped" (really? supposing our system has a DC offset of 0, what polarity did we have and what do we have now?). But what about the sound (the mechanical movement)? Does it make any sense to say that its "polarity" is flipped? What sense do you have to make of "polarity" so that the answer to that question becomes "yes"?

Anyone familiar with polar coordinates and math understands how close the terms polarity and phase are. BUT, in electronics, polarity relates opposing charges while phase is used when a signal angular alignment is concerned. Polarity is a binary property, there is no midway.

The "phase flip" button is generaly used to deal with phase problems, and its only logical to call it that way.

In electronics, polarity is used when DC signals or static charges are the subject in question. When AC signals are the subject, phase is used.

[JHughes]

Originally Posted by shoyoninja

But what about the sound (the mechanical movement)? Does it make any sense to say that its "polarity" is flipped? What sense do you have to make of "polarity" so that the answer to that question becomes "yes"?

When the microphone diaphragm senses a pressure wave, then the speaker cone should also cause a pressure wave in the room. That is called absolute polarity.

[White Tie]

A useful discussion on the name of the thing... what about the icon? Can I just check if anyone has a problem with the icon we use?

[Dstruct]

Originally Posted by White Tie

A useful discussion on the name of the thing... what about the icon? Can I just check if anyone has a problem with the icon we use?

I like it. Fine with me.

[timlloyd]

Originally Posted by White Tie

A useful discussion on the name of the thing... what about the icon? Can I just check if anyone has a problem with the icon we use?

haha

well

if it's called "phase invert" then using the Greek letter "phi" makes sense

if it's called polarity, then the same does not make sense

[karbomusic]

Originally Posted by White Tie
A useful discussion on the name of the thing... what about the icon? Can I just check if anyone has a problem with the icon we use?

I suggested a simple "P" but that was shot down as well. I did call it phase twice last night by mistake in the privacy of my own studio and the building wasn't surrounded by snipers so I suppose I got away with it.

KW

[ubergod]

Because it's probably almost never used, I say get rid of it completely and let the phase thing be handled as a plugin where needed. You could even have two plugins: one for polarity flip, and one for phase rotate. Done.

[JHughes]

The current icon is fine.

[robg]

Well, I was going to chime in here, but fortunately I saw SSG said everything I was going to say almost two months ago

Besides, I forgot to take into account signals with a DC offset like he did, so this all just be comes a 'kudos to SSG' comment!

[medicine tactic]

I'm sure this has been discussed around here at length somewhere, but I'm not finding it.

Basically, what Reaper calls "Phase" is more correctly called "Polarity".

For anyone who's not aware of the difference, this is a pretty good summary. Here's the crux of it:

What's confusing about the '180 degrees out of phase' term is that it is sometimes used to refer to a situation where the second channel has not been delayed, but has had its waveform flipped upside down, so that the peaks become troughs and vice versa — a process more unambiguously referred to as polarity reversal. This scenario also results in silence at the combined output, hence the common confusion in terminology, but it's very important to realise that the total phase cancellation here is brought about by inverting one of the waveforms, not by delaying it. In this example, it might seem like we're splitting hairs, but in practice the distinction between time delays and polarity becomes much more important.

This is a bit of a pet peeve of mine, since I was myself mislead for years by the misinformation in the audio world. So, in the interest of correctness, and of not contributing to the confusion around this subject, I would very much like Reaper to replace "Phase" with "Polarity".

[hamish]

You're right you know.

And I wonder what Bob Katz thinks of the term 'Volume' on our Track Gain sliders?

I've come to live with it. Truth be told, I'm generally having so much fun in REAPER that I don't want to whinge. Not that you're whinging of course.

I'd like to think that REAPER users are a bit more informed than some other systems that shall remain nameless. It would be good if the labels showed that. But then it might be boring.

[EpicSounds]

It bugs me too

There was a thread or feature request around here somewhere for this a couple years back.

[Ollie]

Originally Posted by medicine tactic

I'm sure this has been discussed around here at length somewhere, but I'm not finding it.

Fixed that for ya.

[medicine tactic]

Originally Posted by Ollie

Fixed that for ya.

Ah, thanks Ollie

[chip mcdonald]

I would suggest that both "phase" and "polarity" are incorrect, because they are terms originally meant to describe an analog electrical process, and not a digital signal process.

The button effectively changes the polarity *at the conversion output*, but in the digital medium that really has no direct relationship.

So, if one is wanting to be pedantically argumentative on the basis of "what is the most accurate term", then you probably should be saying something about signed magnitude or bit wise operations, since in reality you don't have an actual "negative" value represented in a binary system.

[JHughes]

I think once we get it called "polarity" we can worry about bitwise operations. Well, someone else can anyway.

Arguing semantics is fun and all, when you want to show folks the magnitude of your critical thinking skills. But in the real world, it's polarity and most of us know and agree on why!

[j79]

Amplitude reversal? Mirror amplitude? Turn waveform upside down? a "what everyone else is calling flip phase"-button?

Its just words...but i agree that technically, calling it phase is incorrect.

We could label the button "*(-1)" because thats exactly what it does (well, technically its not, but its the same thing), changing the sign of each sample...but that would not exactly make things any clearer.

In some sense its a misnomer everyone agrees on...so its not really a problem the way i see it.

Changing it to polarity would be cool, and i am all for it. But it wont really make a difference to anyone will it?

I dont have a problem with the icon...but i dont really understand it either. What does it mean? Empty set? Diameter? I dont get it in the context of polarity. But i still think it is fine

[chip mcdonald]

Guess what? I know "polarity" is technically "better" in the electrical sense, but in *reality*...


... it's easier and quicker to say "flip phase" than "flip polarity". It rolls off the tongue easier, and likewise - that's what I'll probably say verbally in the future.

I also say "trem bar", or "tremolo bar" referring to a guitar, and - I also know that's wrong.

In a world where people can't figure out the difference between "lose" and "loose" - this semantics argument is in turn not a good utilization of time.

[karbomusic]

this semantics argument is in turn not a good utilization of time.

It isn't a good utilization of time, it's a bit of a persnickety OCD like focus on something that really in the grand scheme of things doesn't matter. In 30 years I've never seen someone hunched over a console staring at the button, shaking and not knowing what to do with it. If you are taking an exam, I get it, if you are using the gear, its irrelevant because if pushing it sounds good then we're done.

[medicine tactic]

Part of me agrees with those who are saying it's pedantic. But here's the case for why it isn't:

The problem is, a proper understanding and mastery of polarity is one of the things that separate the men from the boys when it comes to mixing. It's incredibly important.

So this isn't just another unimportant misnomer of convenience. There are noticeable real-world repercussions, in the form of worse mixes. Particularly for those who know just enough about audio to know what phase means, but not enough to know that in this case it's a misnomer. These people -- and there are a lot of them -- actually think the "phase" button delays the signal so it's 180 degrees out of phase. You see it all the time. This notion is so wrong (and impossible in most cases), that it's cancerous, spreading to other areas of one's understanding of the physics of audio.

The thing is, once you join the club of people who actually know the difference, it stops being a problem for you, so you stop caring. This is an instance of the broader class of technical fields flaunting their impenetrable and contradictory terminology as a sort of membership card, or artificial barrier to entry, or linguistic hazing.

[j79]

^yes...thats a good point.