Why does a phaser work the way it does?

[ItsGiusto]

I understand the basic theory of how a phaser like the Phase 90 works - each phase stage shifts one particular frequency by 90 degrees. Therefore by combining two phase-shift stages, and then combining that with the original dry signal, you'll cancel out the frequency and get a notch. A 4 stage-phaser like the Phase 90 creates two notches at two separate frequencies. However, I have two followup questions:

1. Looking at the phase-shifting stage of the Phase 90 here: https://www.electrosmash.com/images/tech/phase-90/mxr-phase-90-script-logo-schematic-parts.png
We can see that each of the phase-shifting stages is completely identical. Wouldn't that result in 1 frequency getting shifted 360 degrees instead of 2 frequencies getting shifted 180 degrees?

2. When you clip R16 in the output mixer stage, to kill the dry-blend, and thus get rid of the phasing notch, you still end up hearing a vibrato sound. Why would the phaser stages be bending the pitch at all? I though it's just changing the phase of a frequency? If you don't blend with the dry, shouldn't the phasing sound be inaudible?

[Rob Strand]

We can see that each of the phase-shifting stages is completely identical. Wouldn't that result in 1 frequency getting shifted 360 degrees instead of 2 frequencies getting shifted 180 degrees?

It just moves the notches outward from each other.  It tends to make the phasing effect sound weaker but you can also think of it as just being different.

If you have some background noise like the hiss from a laptop fan then cup you hands over your ears and move the cups over and away from you ears you can hear the swirly changing.   This is very much like the swirl you get from a phaser.  It's caused by the peaks and troughs in the frequency response moving.

When you clip R16 in the output mixer stage, to kill the dry-blend, and thus get rid of the phasing notch, you still end up hearing a vibrato sound. Why would the phaser stages be bending the pitch at all? I though it's just changing the phase of a frequency? If you don't blend with the dry, shouldn't the phasing sound be inaudible?

So that's the second aspect of a phaser, Pitch-shifting.  By the way when you make the filters different frequencies it lowers the amount of pitch shift.

The pitch you hear is better thought of as the derivative of the phase.    With a fixed frequency the phase of the waveform continues to increase with time (phase = 2*pi*frequency*time).   When you take the derivative with respect to time you end up with pitch of 2*pi*frequency, which is the constant frequency you hear.  However now consider the phaser case the phase is shifting up and down with the sweep  so now when you take the derivative those time varying phase components contribute to a pitch shift.   The shift is positive or negatiive depending on the sweep direction.   The classical example is the Doppler effect where a siren pitch shifts up when a vehicle approaches you and pitch shifts down when when it moves away ; to the driver the pitch is constant.

[ElectricDruid]

I understand the basic theory of how a phaser like the Phase 90 works - each phase stage shifts one particular frequency by 90 degrees. Therefore by combining two phase-shift stages, and then combining that with the original dry signal, you'll cancel out the frequency and get a notch. A 4 stage-phaser like the Phase 90 creates two notches at two separate frequencies.

Not really. A phase-shift stage doesn't shift *only one* frequency. It's an "allpass filter"  e.g. it shifts them all, pretty much.
I like to think of it as a half-way between an inverting amp and a non-inverting amp - in fact, it goes from being one at low frequencies to being the other at high frequencies. And you can build it either way around too, either 0 to 180 or 180 to 0. We only focus on the frequency where the shift is 90 degrees because that's the point that our notch will be deepest when we add a few stages together.

1. Looking at the phase-shifting stage of the Phase 90 here: https://www.electrosmash.com/images/tech/phase-90/mxr-phase-90-script-logo-schematic-parts.png
We can see that each of the phase-shifting stages is completely identical. Wouldn't that result in 1 frequency getting shifted 360 degrees instead of 2 frequencies getting shifted 180 degrees?

If we've got four stages, we've got 4x180=720 degrees of phase shift in total. At low frequencies we start we 0 degrees, and as the frequency climbs, the phase shift increases until we get to 720 degrees. That means there are two points in the spectrum where the output is inverted w.r.t. the input: 180 degrees and 540 degrees. That's why you get two notches.

2. When you clip R16 in the output mixer stage, to kill the dry-blend, and thus get rid of the phasing notch, you still end up hearing a vibrato sound. Why would the phaser stages be bending the pitch at all? I though it's just changing the phase of a frequency? If you don't blend with the dry, shouldn't the phasing sound be inaudible?

Shifting the phase of a signal is the same as modulating its frequency, so you get a vibrato. For a steady frequency, the phase of the signal increases linearly from 0-360,0-360,0-360 over and over. If you start messing with that linear increase in phase (like by introducing crazy phase shifts using a circuit designed for that purpose!) then you also affect that steady frequency.
In mathematical terms, if we've got a sine wave LFO "lfo = sin(x)" and a signal "output = sin(y)" then we're doing this:

output = sin(y + sin(x))

since we're adding our LFO signal to the phase of the output. That's a frequency modulation.

HTH,
Tom

[StephenGiles]

Don't forget - you can't keep taking away from a positive integer without it turning negative, so how would a dome filter figure in all this???

[pinkjimiphoton]

If you have some background noise like the hiss from a laptop fan then cup you hands over your ears and move the cups over and away from you ears you can hear the swirly changing.

is it wrong i do this all the time now? 

[Rob Strand]

is it wrong i do this all the time now? 

How do you know it doesn't come from within 

[Rob Strand]

is it wrong i do this all the time now? 

The opposite experiment is pretty cool  too.  Sitck your head under a whole heap of blankets so there's no background noise.  Do the cupped hands thing, you hear nothing.  If you have a shell you can listen "to the sea" in the shell but when you go under blankets you hear nothing.   The sound in a shell is actually just a filtered version of the background noise.

[ItsGiusto]

I was always under the impression that the swirly sound you get by cupping your hands and moving them to and from your ear is not due to phasing, but due to flanging. You're causing comb filtering by creating a reflection source that close to your ear. Then by moving your hand, you're oscillating the delay time of the reflection source.

[Rob Strand]

I was always under the impression that the swirly sound you get by cupping your hands and moving them to and from your ear is not due to phasing, but due to flanging. You're causing comb filtering by creating a reflection source that close to your ear. Then by moving your hand, you're oscillating the delay time of the reflection source.

The effect is actually the moving filter.  Doesn't matter how you achieve it.    The cupped hands and shell act more like a resonator with a peak (the effect is stronger when there's only a small gap between your ears/head and the thing)  but there are some weaker multiple reflections thrown in.

Here you go, it's on wiki,
https://en.wikipedia.org/wiki/Seashell_resonance

[ItsGiusto]

I was always under the impression that the swirly sound you get by cupping your hands and moving them to and from your ear is not due to phasing, but due to flanging. You're causing comb filtering by creating a reflection source that close to your ear. Then by moving your hand, you're oscillating the delay time of the reflection source.

The effect is actually the moving filter.  Doesn't matter how you achieve it.    The cupped hands and shell act more like a resonator with a peak (the effect is stronger when there's only a small gap between your ears/head and the thing)  but there are some weaker multiple reflections thrown in.

But the same effect is achieved even if you don't use a resonator. If you either move your head close to and far from a wall, you'll get the effect. Or if you move the noise source close to and away from the wall.

[Rob Strand]

But the same effect is achieved even if you don't use a resonator. If you either move your head close to and far from a wall, you'll get the effect. Or if you move the noise source close to and away from the wall.

Well, the cup of you ear helps.    However, you can get a filtering from the flat surface simply because the you hear a direct sound and a reflected sound.  They interfere with each other and produce peaks and notches.   This effect is weaker than a resonator, a resonator will boost the level.  Another effect is diffraction, when you put your head near a wall it blocks the high frequency but the low frequencies will diffract around your head and make it your ear, that kind of rolls-off the high frequencies.

What you will notice with a cup or a surface is the pitch is the "shell sound" decreases as the gap between the ear and the object becomes small.    If you check out this page,

https://en.wikipedia.org/wiki/Helmholtz_resonance

You can see from the formula that as area of the mouth of the resonator is reduced the resonant frequency *decreases*.   If you have a reflection from a surface bringing the object closer would raise the frequencies (you might be able to hear that when further away).

The direction of movement (reduced area) vs pitch is the thing that shows the main effect.

[pinkjimiphoton]

dare i say comb filter?  lol

[Rob Strand]

dare i say comb filter?  lol

lol, you know what,  I was going say something about the quiet experiment not working on you because your hair would make rustling noises  .


Rustlers disturbing the peace,

[iainpunk]

Don't forget - you can't keep taking away from a positive integer without it turning negative, so how would a dome filter figure in all this???

dome filters are just a bunch of cascaded allpass filters so that there is an band of frequency's that lags or leads 90 degrees. reading up on that subject, and seeing the mathematics involved makes me feel inadequate. 

yes you can keep subtracting from a positive integer and keep it positive, look at linear numbers on a LOG scale, they never even reach 0.

cheers

[PRR]

how would a dome filter figure in all this???

The Dome gives the -difference- of two signal paths. They both have positive delay, but not the same. Take the difference the right way, and you can have negative relative phase between the two paths.

This works great over a long SSB radio link where the listener has no access to the original source. Probably true for e-guitar where the original source is not audible behind the LOUDspeaker.