Delay Node
  • What is the maximum delay time of the node? Also, is there a way to output that info to a value node? Also, when I modulate the time to zero, there is often a clicking sound. Any guesses as to why that may be?
  • 2 seconds - yes, you would do something like the attached patch:

    If you suddenly change something that is processing audio that causes the wave to jump sharply, you'll likely create a clicking noise. Like if you have the attack or release times set to 0 on your ADSR node. This is because the wave is sharply moving away from or returning to zero which, when not bandlimited like the saw oscillator is, causes distortion. It's the same thing that happens in any DAW and you make a loop where the waveform isn't settled to 0 on both ends of the loop - the sudden mismatch in amplitude creates a burst of harmonics, not dissimilar from a finger snap.

    To prevent this from happening, apply a slew limiter to your control signal. Giving that little rise and fall time will eliminate that clicking noise. When you connect and disconnect things with patch cords in Audulus, this is done automatically. You can see it happening when you connect a 1 value to a value node - briefly you'll see little tiny values pop up first instead of transitioning immediately to 1.
    Delay Knob Second Indicator.audulus
  • I did notice those little blips, I always wondered what they were. Many thanks as always for the quick reply.
  • Welp, I'm still having problems when the modulation reaches zero and I'm using a sine wave, so there shouldn't be any transients in the modulation signal. Could it be a phase cancellation artifact? I've attached a patch to illustrate the clicking.
    Clicking Delay.audulus
  • This fixes the clicking - just adding a tiny bit of time to the floor.

    Unlike most normal delays, this delay node goes from 0 seconds to 2 seconds. Most delays go from like 20ms to X seconds. I'm sure it's doing something similar where it's creating a snap from a burst of harmonics from it hitting 0.

    Also, the way you have the delays set up is not technically correct. Although you're going to get a maximum of 4 seconds from a burst of time, you really just have 2 2 second delays running in series.

    If you play one note and have the two delays turned all the way up with the mix all the way to wet, you'll hear the delay come 4 seconds later. But if you want to add feedback, you'll have to do it externally like I did on my analog modelling delay. That is, if you tried to adjust the feedback to greater than 0, it would audibly be two delays in series rather than one long one.
    Screen Shot 2016-11-20 at 6.39.05 PM.png
    574 x 407 - 54K
  • Well, it pains me that my delay will not be accurate to the millisecond, that certainly is a fine workaround for the time being. You are quite right that chaining delays isn't really the same as one actual really long delay. In fact you may have noticed that it sweeps the delay buffer twice from high to low, which is kind of an interesting effect in itself.

    But I am using it more like brian eno long tape delays for ambient soundscaping and sending a delayed modulation signal. A delay with a long time and no color is perfect. I am guessing there is a reason why you can't record a bit of audio and play it back in Audulus? At any rate, thanks for the help and I will be sure to look up your analog modelling delay again to really understand what's going on. All the best :)
  • Is there a constant knob with an input? Where is that hiding in the menu tree?
    Screen Shot 2016-11-20 at 5.07.37 PM.png
    949 x 595 - 80K
  • Your delay can still be accurate - you just need to add that 0.001 to the total as it goes to the value node. If you want it to max out at 2, you'd first multiply the knob by 1.999 and then add .001. That way the range will be .001 - 2.

    Constant knobs don't have inputs, per se. The knob is the input. You should make a habit of leaving all knobs tuned to 0-1 so that modulation signals interface with them without needing translation. If you need a constant knob with a different range, just do the multiply/add thing that I described above, or one of the knob range modules.

    As for the analog delay, what makes it "analog"is a LPF in the feedback path. You can still have a clean digital delay but you just have to tune the feedback the right way. I put a tanh(x) expression inside the delay path to prevent runaway feedback, and it actually overdrives nicely. I'm going to do a tutorial here in a moment on how to use the ES-8 to create your own custom "analog" delay with an analog filter module using the delay node - you'll see how it works then :)

    Also, haven't forgotten about your random thing, I'll do that soon too!
  • Cool! I just copied and pasted the constant with an input because it is handy for slapping a 0-1 multiplier to the signal. It save me typing a "Signal*Knob" expr node and makes things look tidy.

    Looking forward to the vid, are you going to use the gristleizer? I love the name of that thing. Hope you get around to the random thing someday but no rush. Until next time.
  • @robertsyrett - I'm using the Mankato Filter by STG Soundlabs