Phase Settings of Synths

[Sweaty Ear]

Hello all.

Is there anyone here who can explain how the phase setting of a synth changes the sound? It doesn't seem to do anything on my vst synths or at least i'm not hearing it.

[FilterEverything]

Hi,

A lot of synths are based on playing various single cycle waveforms (such as a sine wave, saw wave, square wave, etc...) and giving some controls over how to adjust these basic waveforms and how you want the waveforms to be played back.

A lot of the times when there is some kind of "phase knob" that can be tweaked, it relates to the starting point of the waveform you want to play back.

Here's a somewhat interactive graph that shows it instead of just explaining it. It has 3 sliders, frequency, phase1 and phase2. Phase1 is fixed, not tweakable.
https://www.desmos.com/calculator/ksp2iml9p8

There are two single cycle waveforms on the graph, both are sine waves with the same frequency. The X axis is time, the Y axis is amplitude. If you tune the frequency slider, both graphs change how many times they finish a "full motion" over a period of time. If the frequency is 1Hz, a cycle spans from 0 to 1, from 1 to 2 and so on... If the frequency is 2Hz, a cycle spans from 0 to 0.5 from 0.5 to 1 and so on...

Once you start to tune the phase2 slider, you can see that one of the sine wave begins to "shift". The shape and frequency remains, a shift in the X axis (denoting time) is the only thing the slider does. You can think of the intersection of the axis as the starting point of the playback. For the sine wave if there is no phase change, the playback starts from 0 amplitude, and gradually increases from there, if there is 90deg (1.57 rad on the graph), the playback starts from the maximum amplitude and gradually decreases from there.

After the beginning of the playback, you won't hear a difference between the original and the phase shifted waveform, because the difference only exists at the beginning. As long as there is only 1 layer, changing the starting phase won't really matter especially for sustained sounds.

If the playback does not start from Y= 0, you will hear a small click at the beginning of the sound. The larger the difference from 0, the bigger the click will be. This can be useful if you want to make a somewhat percussive sound like a quick pluck or stab, you can just change a the starting phase a bit to add a nice little attack without relying on other layer.

However, phase becomes extremely important once you start summing (mixing) similar waveforms together. If the waveforms are out of phase, the sum will be smaller than the parts because at parts of one of the waves where the amplitude is positive, the out of phase version's amplitude will be negative, so they will sum to 0 or near to 0.

A way you could test this in a DAW is to use a stereo oscilloscope VST where you could use the same synth patches as input, and once you patch one of them to the left channel and the other one to the right channel, you could start tweaking the phase and see how the phase tweaked channel starts shifting relative to the other channel.

[Sweaty Ear]

thanks for your response. will have to try out these ideas. off the top of your head would there be any other uses for phase other than adding percussiveness?

[planetearth]

As @FilterEverything explained, a "phase" setting on a synth should only affect the phase of one of the oscillators (usually, oscillator 2) with respect to the other oscillator (usually, oscillator 1). In a single-oscillator synth, you wouldn't even have a "phase" adjustment generally speaking—and I can't think of a single-oscillator synth that has a "phase" setting.

And as @FilterEverything also implied, as you adjust the phase of one oscillator with respect to the other, the sound will get thinner and thinner until one of the oscillators is 180 degrees out of phase with the other oscillator, and the sound disappears as the waves cancel each other out. As you go past that point, the sound gets thicker again. So, aside from a percussive "tick" sound, adjusting the phase of one oscillator with respect to the other can make a synth sound "thinner" or "thicker".

You might also see "Phase" on a synth in reference to an oscillator or filter envelope. The correct term would really be "Invert", where the envelope is "flipped" upside-down (or, inverted) with respect to what the ADSR settings show. That would be the only case I can think of where "Phase" might be used on a single-oscillator synth, and again, it wouldn't be the correct term.

Steve

[swilow11]

thanks for your response. will have to try out these ideas. off the top of your head would there be any other uses for phase other than adding percussiveness?

Being able to adjust the starting phase of your bass can help it better align with your kick. If the bass and kick are 180° out of phase, any frequencies they share can cancel each other out. Aligning them so the peaks of each waveform line up will prevent this although stacking waveform peaks like this can create amplitude spikes which can eat up headroom.

Phase is also useful if you're FMing stuff. The phase of the modulator can quite significantly alter the timbre of the modulated waveform. If you have Phase Plant, you can see and hear this in action.

If you modulate phase, you will also hear a sort of pitch shift effect. Rapid modulation can make a chorus type sound.

In general, phase is a very important concept to understand because it extends out to any quality that runs in cycles. Things like LFOs have a phase too which can determine the sort of modulation you get and how it marry up to your track.

It's normal to twiddle a phase knob on a synth and not hear much though. You don't really hear phase most of the time but it changes how the signal interacts with other signals. For example, with a supersaw or unison waveform, if all the voices start at the same phase, it sounds very artificial and gives a muddy sort of attack sound. In this case, you want phase randomisation which means that every voice starts at a different point of the waveform. It is the phase cancellation and reinforcement that gives supersaws their characteristic sound.

[swilow11]

As @FilterEverything explained, a "phase" setting on a synth should only affect the phase of one of the oscillators (usually, oscillator 2) with respect to the other oscillator (usually, oscillator 1). In a single-oscillator synth, you wouldn't even have a "phase" adjustment generally speaking—and I can't think of a single-oscillator synth that has a "phase" setting.

And as @FilterEverything also implied, as you adjust the phase of one oscillator with respect to the other, the sound will get thinner and thinner until one of the oscillators is 180 degrees out of phase with the other oscillator, and the sound disappears as the waves cancel each other out. As you go past that point, the sound gets thicker again. So, aside from a percussive "tick" sound, adjusting the phase of one oscillator with respect to the other can make a synth sound "thinner" or "thicker".

You might also see "Phase" on a synth in reference to an oscillator or filter envelope. The correct term would really be "Invert", where the envelope is "flipped" upside-down (or, inverted) with respect to what the ADSR settings show. That would be the only case I can think of where "Phase" might be used on a single-oscillator synth, and again, it wouldn't be the correct term.

Steve

This isn't correct or at least not completely. A phase knob works per oscillator. It isn't a rekative setting, or not always. Think of a waveform like a sinewave as a perfect circle. A phase knob effects at which degree or point of the circle the waveform starts at. It certainly isn't isolated to oscillator 2 as a matter of convention. Every waveform goes through 360° of oscillation and the phase knob determines the point at which the signal starts.

Your 2 oscillator example will only work for 2 oscillators with precisely the same waveform. But regardless of whether you've got 1 oscillator or 100, all of them can be described by using phase position and all phase positions effect sound independent of any other oscillators.

I've never seen a synth with phase for an envelope but most synths with LFOs will have a phase setting for the LFO although I've also seen this labelled position or simply start. It's most evident with sharper waveforms like a saw where you can get quite different sounds depending on where the cycle begins.

[planetearth]

As @FilterEverything explained, a "phase" setting on a synth should only affect the phase of one of the oscillators (usually, oscillator 2) with respect to the other oscillator (usually, oscillator 1). In a single-oscillator synth, you wouldn't even have a "phase" adjustment generally speaking—and I can't think of a single-oscillator synth that has a "phase" setting.

And as @FilterEverything also implied, as you adjust the phase of one oscillator with respect to the other, the sound will get thinner and thinner until one of the oscillators is 180 degrees out of phase with the other oscillator, and the sound disappears as the waves cancel each other out. As you go past that point, the sound gets thicker again. So, aside from a percussive "tick" sound, adjusting the phase of one oscillator with respect to the other can make a synth sound "thinner" or "thicker".

You might also see "Phase" on a synth in reference to an oscillator or filter envelope. The correct term would really be "Invert", where the envelope is "flipped" upside-down (or, inverted) with respect to what the ADSR settings show. That would be the only case I can think of where "Phase" might be used on a single-oscillator synth, and again, it wouldn't be the correct term.

Steve

This isn't correct or at least not completely. A phase knob works per oscillator. It isn't a rekative setting, or not always. Think of a waveform like a sinewave as a perfect circle. A phase knob effects at which degree or point of the circle the waveform starts at. It certainly isn't isolated to oscillator 2 as a matter of convention. Every waveform goes through 360° of oscillation and the phase knob determines the point at which the signal starts.

Your 2 oscillator example will only work for 2 oscillators with precisely the same waveform. But regardless of whether you've got 1 oscillator or 100, all of them can be described by using phase position and all phase positions effect sound independent of any other oscillators.

I've never seen a synth with phase for an envelope but most synths with LFOs will have a phase setting for the LFO although I've also seen this labelled position or simply start. It's most evident with sharper waveforms like a saw where you can get quite different sounds depending on where the cycle begins.

Well, I tried to keep things simple, since no one knows how much experience the OP has had with "phase" on a synth. I didn't say a phase knob is "isolated to oscillator 2", however. I was only describing a simple implementation of phase, along the lines of what @FilterEverything's link showed. And yes, the phase knob determines the point at which the signal starts.

But as I said, I don't know of any single-oscillator synth that bothers to offer a "phase" knob. The most noticeable changes that "phase" offers in a synth is when you're altering the phase of one oscillator with respect to another. And while an LFO might offer a "phase" adjustment, that's still technically adjusting the phase of a second oscillator with respect to the first. And since the OP didn't ask about LFOs, I didn't want to muddy the water by trying to explain how the phase of an oscillator that is so low it (usually) isn't heard is affecting/modulating the sound of another oscillator.

As you said yourself, "You don't really hear phase most of the time but it changes how the signal interacts with other signals". That's really all I was getting at, while trying to keep it simple. The examples of the effects phase has with respect to FM and supersaws is a good point, But I was only trying to get the basic ideas across in the context of what had already been discussed.

Steve

[zerocrossing]

One thing you can do with a synth that doesn’t have free running oscillators is to give the oscillators a random phase modulation. This can help a digital synthesizer sound a bit more lively.

[swilow11]

As @FilterEverything explained, a "phase" setting on a synth should only affect the phase of one of the oscillators (usually, oscillator 2) with respect to the other oscillator (usually, oscillator 1). In a single-oscillator synth, you wouldn't even have a "phase" adjustment generally speaking—and I can't think of a single-oscillator synth that has a "phase" setting.

And as @FilterEverything also implied, as you adjust the phase of one oscillator with respect to the other, the sound will get thinner and thinner until one of the oscillators is 180 degrees out of phase with the other oscillator, and the sound disappears as the waves cancel each other out. As you go past that point, the sound gets thicker again. So, aside from a percussive "tick" sound, adjusting the phase of one oscillator with respect to the other can make a synth sound "thinner" or "thicker".

You might also see "Phase" on a synth in reference to an oscillator or filter envelope. The correct term would really be "Invert", where the envelope is "flipped" upside-down (or, inverted) with respect to what the ADSR settings show. That would be the only case I can think of where "Phase" might be used on a single-oscillator synth, and again, it wouldn't be the correct term.

Steve

This isn't correct or at least not completely. A phase knob works per oscillator. It isn't a rekative setting, or not always. Think of a waveform like a sinewave as a perfect circle. A phase knob effects at which degree or point of the circle the waveform starts at. It certainly isn't isolated to oscillator 2 as a matter of convention. Every waveform goes through 360° of oscillation and the phase knob determines the point at which the signal starts.

Your 2 oscillator example will only work for 2 oscillators with precisely the same waveform. But regardless of whether you've got 1 oscillator or 100, all of them can be described by using phase position and all phase positions effect sound independent of any other oscillators.

I've never seen a synth with phase for an envelope but most synths with LFOs will have a phase setting for the LFO although I've also seen this labelled position or simply start. It's most evident with sharper waveforms like a saw where you can get quite different sounds depending on where the cycle begins.

Well, I tried to keep things simple, since no one knows how much experience the OP has had with "phase" on a synth. I didn't say a phase knob is "isolated to oscillator 2", however. I was only describing a simple implementation of phase, along the lines of what @FilterEverything's link showed. And yes, the phase knob determines the point at which the signal starts.

But as I said, I don't know of any single-oscillator synth that bothers to offer a "phase" knob. The most noticeable changes that "phase" offers in a synth is when you're altering the phase of one oscillator with respect to another. And while an LFO might offer a "phase" adjustment, that's still technically adjusting the phase of a second oscillator with respect to the first. And since the OP didn't ask about LFOs, I didn't want to muddy the water by trying to explain how the phase of an oscillator that is so low it (usually) isn't heard is affecting/modulating the sound of another oscillator.

As you said yourself, "You don't really hear phase most of the time but it changes how the signal interacts with other signals". That's really all I was getting at, while trying to keep it simple. The examples of the effects phase has with respect to FM and supersaws is a good point, But I was only trying to get the basic ideas across in the context of what had already been discussed.

Steve

Yeah, fair enough. It was good for you to provide an audible way to hear phase via phase offset and cancellation between oscillators. Similarly, I mentioned LFOs because you can hear starting phase and obviously being able to hear stuff takes us a long way down the path to being able to manipulate those properties in a musical way.

It's a funny sort of concept because it is deeply important to synthesisers and yet isn't necessarily audible in most contexts.
Unless we're taking about phase modulation or phase distortion like bending etc. Getting a good grasp on what phase really is helped me unlock a fair bit of sound design goodness.

[Sweaty Ear]

while i do want to learn as much as possible about phase, simple is good. lol. knowing you can make a sound thinner without a hpf for example can come in handy. i get phase position and where it is on the lfo but tend to get lost when it comes to fm synthesis and/or technical math behind it. so being able to see/hear real world applications is helpful. thanks.

[FilterEverything]

I think there are only a few mathematical concepts that are worth diving into as a producer, especially if you don't consider it as the part of hobby/fun. Understanding math rarely if ever translates into writing better music but it sometimes lets you make smarter decisions about mixing and it sometimes can help quite a bit with making synth sounds. With that said, phase is one of the more useful piece of math that's worth understanding. Maybe not to the point you start solving trigonometric functions to practice it but definitely to the point you understand the concept fundamentally and have a strong intuition of how the sound would react to a phase change.

By far the most important thing is to understand the idea behind how a phase shifted (time delayed really) signal summed with the original causes both destructive and constructive interference at the same time, also known as comb filtering. If you understand that, you'll have a very easy time understanding chorus, flanger, phaser plugins, unison settings found in most synths, understand why recording the same vocals multiple times and playing them together actually makes sense (and when it isn't), why a multi mic drum recording somehow sounds horrible even after spending ton of time stacking effects on it and tweaking it. Understanding phase can explain why a channel suddenly starts peaking even though only subtractive filtering is applied, why saturators and distortion plugins, or dynamic processors such as compressors and limiters can sound wildly different if an all-pass filter is applied even though the effect of the all-pass filter by itself is negligible and so on... and this is mostly mixing, not even getting into the modulator/carrier synth stuff like FM or RM synthesis.

knowing you can make a sound thinner without a hpf for example can come in handy.

The thing is, you probably already knew that a sound can be made thinner by applying a comb filter to it but you probably didn't know that all a phase shifted copy (mixed with the original sound) does is creating a static comb filter.

[Sweaty Ear]

this does shed some light on why unison, flangers, comb filters, etc. sound similar.

[ugo]

Returning to the original question/example of a synth with a phase control for oscillators… (I apologize if this was already said and I missed it.)

When mixing two waveforms that are in phase with each other, the shared harmonics are reinforced. When they are out of phase the shared harmonics cancel each other out. FilterEverything provided a good example of this. Mixing two sine waves with the same pitch, amplitude, and phase results in a louder sine wave. As you shift one out of phase with the other they begin to cancel each other and the output gets quieter. Once you hit 180 degrees out of phase, you have silence. Even though you shifted the phase of one of the sines instead of inverting it, when comparing the two the result is the same so they’ve completely cancelled each other out.

Now let’s say we are mixing a saw and a square. A saw has both even and odd harmonics, and a square only has odd harmonics. With the oscillators in phase, the odd harmonics (including the fundamental) are now amplified - resulting in the beefy sound you likely expected. Now put them 180 degrees out of phase from each other and you get…a quieter saw…one octave higher. All the odd harmonics (again including the fundamental) have been cancelled out so you’ve removed a lot of energy from the signal - but the harmonic structure of a saw is such that the result is the same as if you shifted everything up an octave.

If you did the same test with different waveforms, you wouldn’t necessarily get the same results. It all depends on what frequencies the two have in common, as well as the amplitude of each frequency.

[Sweaty Ear]

this is interesting. i wonder if there's a list anywhere of the results of mixing 2 basic waveforms and adjusting the phase of one. if not will just have to put in the hours.

[planetearth]

this is interesting. i wonder if there's a list anywhere of the results of mixing 2 basic waveforms and adjusting the phase of one. if not will just have to put in the hours.

What would you want from such a list? I think @ugo explained it very well, especially if you mix a sawtooth and a square waveform (where the original example only showed sine waves).

There are probably "audio" websites that show what would happen if you mix two waveforms and adjust the phase of one. But you could also do this much quicker in any basic "analog"-type synth you have, as long as it has at least two oscillators, allows you to choose waveforms, and allows you to invert one. You'd also get the immediate benefit of hearing what this sounds like, instead of reading the resulting amplitude and frequency graphs and trying to imagine what that sounds like.

Almost any "analog" or "virtual analog" plug-in or hardware synth can do this, so if you're not sure how to set it up, just tell us which synths you have, and I'm sure someone here could walk you through it. (And for what it's worth, I doubt it would take "hours" for you to start to understand the concepts and start to know what the resulting sound would be. )

Steve