Quadrature oscillator
  • Can you think of any way to use the available tools to produce a quadrature oscillator... that is, an oscillator that produces a pitch/phase-locked sin and cosin wave simultaneously?

    Or does this amount to a feature request?

    (Or is this something I should be able to build myself in code?)
  • You should probably be able to build it yourself, but I don't really understand what it is. "(sin(x)+cos(x))/2"? That would just be a phase shifted sine wave, which isn't very interesting sounding, so I'm obviously missing something. What does it sound like?

    In any event, your best bet is probably to start with a Phasor running into an Expr node, like I have in the attached (obviously incorrect) attempt at a Quadrature. :)
    JJT Non-working Quadrature.audulus
  • A quadrature oscillator is one that generates two signals simultaneously, 90 degrees out of phase. Here's a link to a circuit: http://www.play-hookey.com/oscillators/audio/quadrature_oscillator.html

    It turns out that using a quadrature oscillator is one way to make a frequency shifter without needing brickwall filters:



    The expression node does bring up an idea though. I could split a sin and do a simple trig transform to convert one side to cos. That could be bundled as a quadrature oscillator.

    I'll try that. Thanks :-)
  • Oh, so *that's* what the phasor is for. Now I get it :-)
  • This is a subject I've been working my way through. I failed Algebra in school so I still struggle with anything that is actual hands-on math. One thing I do know from electronics though is that opposite phases cancel each other out. This is true at any frequency for sine waveforms. When I combined the phases in your patch however, they don't cancel. Instead it appears that the frequency triples. Not what I would have expected.
  • You're absolutely right, opposite phases do cancel. And that's the basis of the discussion I was having with biminiroad about ring modulators. In a modular patch you can suppress the carrier of an amplitude modulated signal by mixing an opposite-phase copy of the carrier back in.

    This is a bit different though. Opposite phase is 180 degrees of shift. In waveform (time domain) terms, this means pulling the waveform back one-half cycle. This little patch produces a sin and a cosin, though, and a cosin is 90 degrees - one quarter cycle - out of phase with the sin.

    So yeah, when you mix them they won't cancel. And the waveform you see when you mix then isn't really a tripled frequency. It's just the sort of odd looking thing you get when you mix two signals together. They add, the result is (slightly) more "complex."

    Set the constant to something audible like 400Hz, and attach an output. You'll hear what I mean.
  • I was actually just looking up about this stuff after seeing your other thread about suppressing the carrier and thought about single sideband. To jog my memory I was looking at some of the wikipedia stuff about implementation, and they mentioned phase shift to the signal rather than carrier. It seems easier to generate a carrier and another signal in quadrature than to have a constant phase shift applied across the whole audio frequency.