Using a dual delay and quadrature LFO for pitch shifting

[Buttons ARE toys]

This is a patch that I made just to see how it would sound to build a basic pitch shifting algorithm out of modules. It gives you a somewhat old school sounding effect, definitely not clean and precise, but it's a lot of fun to experiment with breaking it in ways another pitch shifter might not let you.



You want to use a ramp wave to modulate the delay time so that the sound is constantly rising or falling in pitch. But, to hide the glitch that occurs when the ramp resets, you need to control the amplitude with a sine wave that is low when the ramp reset occurs. This creates a tremolo effect though, so you need a second delay that's modulated by another ramp, 180 degrees out of phase with the first. Control that delay's output with another sine wave that's 180 degrees out of phase with the other, and you'll get a continuous stream of pitch shifted audio where rising ramps increase the pitch, and falling ramps decrease it.

I'm not sure how this would sound with another delay, but I'm sure it would be possible with many different ones. The main difference would be using an attenuator on the ramp signals to control the pitch shift amount, since you can just do essentially the same thing with the d0's slew. It would be interesting to hear what kind of character different modules produce if anyone wants to try this with some different modules and capture some audio.

[timoka]

i already wrote on lines, this is such a fantastic patch, so clever and it showcases why i love modular synthesis!
indeed it would be quite interesting to hear this patch with a different delay. but don't you think it needs the slew rather than just attenuation of the ramp waves?

[Buttons ARE toys]

i already wrote on lines, this is such a fantastic patch, so clever and it showcases why i love modular synthesis!
indeed it would be quite interesting to hear this patch with a different delay. but don't you think it needs the slew rather than just attenuation of the ramp waves?

I would say that the delay needs some amount of slew applied, but I have the patch rebuilt currently, using Frames to control the amplitude of the ramp waves before hitting the d0, and it does a great job of controlling the shift amount. The slew on the d0 is then free to be used to adjust the character of the sound. Having no slew applied at all though pushes the effect into noise territory, so I think other delays should work as long as they slew the delay time a bit.

[Navs]

Must try this.

[starthief]

Very cool!

[kosmische]

This is rad, thanks for sharing!

[dragulasbruder]

Sounds fantastic. A few questions:

1. Why does this patch need a quadrature LFO when you're just using phase relationships (180deg) that can be attained with a simple inverter? Or did I misunderstand the phase relationship between the sine and ramp waveforms?

2. When you are modulating the amplitude of the two delay lines with the sine waves, what's happening during the negative portion of the cycle? Are the sine waves offset so that only their maximum negative voltage is reducing the VCA's gain to 0? Or are they used without offset, making the VCA gain 0 for the whole negative half of the waveform?

3. How does the LFO speed affect the outcome of the pitch shift? How fast are you running the LFO in your example?

Thanks!

[Buttons ARE toys]

Sounds fantastic. A few questions:

1. Why does this patch need a quadrature LFO when you're just using phase relationships (180deg) that can be attained with a simple inverter? Or did I misunderstand the phase relationship between the sine and ramp waveforms?

2. When you are modulating the amplitude of the two delay lines with the sine waves, what's happening during the negative portion of the cycle? Are the sine waves offset so that only their maximum negative voltage is reducing the VCA's gain to 0? Or are they used without offset, making the VCA gain 0 for the whole negative half of the waveform?

3. How does the LFO speed affect the outcome of the pitch shift? How fast are you running the LFO in your example?

Thanks!

1. For the sine waves you could totally use an inverter. For the ramp though, inverting will reverse its direction, so you would have one delay raising the pitch while the other lowers it. Having a 180deg phase ramp keeps them both moving the same direction.

2. The VCA I'm using has onboard bias and attenuation for the incoming control signal, so I have it tuned as close as I can to zeroing out at the bottom of the LFO cycle. I've tried it without any bias so that you're just getting half of the sines' amplitudes and it still works, but the effect of the AM will be more noticeable.

3. The speed of the LFO changes a few things. It impacts the amount of pitch shift to some degree, with faster LFO being able to produce higher and lower pitch shifts, and it impacts how audible the AM is. That could be a result of Batumi losing fidelity as it goes higher into audio rates, but even at frequencies around 100Hz or so, the AM becomes quite noticeable. I'm not sure of the exact frequencies I ran the patch at in the video, but at the start it sounds to be around 20 or 30 Hz. Then when I mention the tone changing with the speed, it's up near 100Hz.

[dragulasbruder]

Incredible, thank you! I think that a lot of us (or maybe just I) forget sometimes that a 180deg phase shift of an asymmetrical waveform is not the same thing as inversion!