Farview said:
You might be missing one small point about the phase button: It only acts on one track. In order to affect phase, there has to be (at least) two waveforms. The degree of phase would depend on the frequency you are measuring in a complex waveform. Flipping the polarity of one waveform against another (different waveform) will result in comb filtering, because the frequency content is not exactly the same in each waveform. Only certain frequencies will be 180 degrees out.
The only time a phase switch will actually throw something 180 degrees out of phase is when you have two of the exact waveforms playing against each other. This really doesn't happen in real life, so polarity is more accurate.
Once again, the GIGO machine is in effect.
I didn't want ot have to do this, but since you all seem to think that "OMFG Phase = Time" and "You ned two waves" then I'm going to have to point this litte fact out to you:
THe wave equation, in it's most correct form, is:
d^2u/dt^2 = c^2(delta u) - or, in other words, the acceleration of a particle is relevant to it's displacement. This, of course, will usually result in the sinusoidal movement of the particle, however when we intergrate the double derivitave, you end up with a few constants that cannot be forgotten about.
One of these is the phase shift.
Performing hte double intergration with the Fourier Transform, and over simplify, you end up with:
u = cis(omega*t + phi), where u is sthe displacement as a function of time, cis is the complex sinusoidal term ( cos(X) + i Sin(X)), omega is the angular velocity (complex word for fresuency, represented by the "velocity" of the phase angle on the complex plane- ie f/2pi), t is the time and phi is the constant of intergration.
This is where the major "Phase = time!" fallicy comes into play, because a change in phase will ahve the same effect on u as a change in time would have (as they are both in the same set of brackets)
And that's the simplest form. Once you start looking at other intergration paths, you'll notice that they all have some constant, which, when intergrated out, become the initial phase of the wave.
In other words, every wave has a phase releative to "the great cosmos" or whatever you want to call it. Push your hand forward through the air- you've jsut started a spherical wave with an arbitrary phase. Now, push your hand backwards. You've now just created a spherical wave with a reversed "phase". Whilst it may seem insignifigant until you comapare two waveforms together, every wave has its phase.
What Farview is thinking of in interference, in which you do need two waveforms.
However, if you have two identical waveforms that are totally out of phase, then they will cancel TOTALLY. If you want to start thiking about complex interferance, you've got to shake the sinusoidal notion of waves and start thinking of complex and aperiodic waves. A phase reversal will always result in total destructive interferance. That's basically how you work out the resultant amplitude- you compare the vectors on the complex plane, add them together, and the resultant vector's x component will give you the amplitude. Then, as those vectors spin around, your amplitude will change accordingly (unless, of course, the vectors are exactly opposite, in which case the resultant x component will alwys be 0).
You're also totally wrong on the polarity front- take a shape- any shape, be it a wave form, square... whatever.
Mirror it (ie change it's "polarity")
Add the two together. In the simplest case, the mirror changes the sign of every "part" of the shape, and when you add +x and -x, you get 0 in all cases.
Draw an aperiodic waveform, then draw the same waveform where the amplitude is the negative of the first waveform for all t (ie, siwtch the instantaneous polarity of the wave).
Then, for all t, add the amplutides of the first and second wave together- no matter what the instantenous amplitude, the result will always be 0, so by switching the poliarty of the instantaneous amplitude, then you will get total cancellation, not comb filtering.
In reference to the Katz Quote:
Firstly, I'll say it one more time:
Polatiry is a static vector term
Phase is a dynamic vector term.
If you have a look at at a wave on the complex plane, you will see it's vector spinning around in a circle. It is this behaviour that excites nerdy types, as expressing a wave as a circle makes things so much easier than trying to draw sin waves ad infinitium.
If you express a swith in polatiy on the complex plane, you have a straight line passing through the origin, with equal lengths both sides of the origin.
a "Physical" phase shift is one that you can "do" with your hands- like switching pins 2 and 3, pressing a phase button... whatever.
What this does to the electronics is a poliarty reverse, as it switches two bits of metal, which are static.
However, what it does to the wave (as seen by the next downstream instrument) is switch it's phase.
Once again I reffer to the the cis(omega*t + phi) part of the above equation- whilst a change in time or phase can manifest itself with the same effect as a change in the other, that isn't what's happeneing.
The point I was trying to make earlier in reference to the Katz quote is that whilst a time shift and a phase shift look similar, they aren't the same thing, so by saying that a phase shift is a shift in time, you're creating the false environment where a phase shift of pi will create a differnt delay for different frequencies.
But, we know this is false, becuase phase is time independant, hence a phase shift of pi will leave the original wave intact, just with a phase shift...
One final request... please, please start using your heads people.
If you want to know more about this subjetct, DON'T keep reading this forum, go out and get a basic wave theory textbook and try working through it.
Take everything I've said with a grain of salt, as I haven't verified any of this (although I will on request). However, make sure that you take what everyone else says with that same salt- even if they have a huge reputation and a "big name".
Strive to know your industry, don't just copy others.
)
