DS Curves 1
  • Attach a knob, feed a control signal through it, DC-offset audio then reset it back to AC - curves like these are essential for making compressors, or inducing any kind of non-linear response into a patch. Here's the first collection of 8 modes with 4 different taps/pressure settings - that's 32 separate settings from one lil subpatch!

    Between these 3 waveforms, you could easily make a tremolo with 96(!!!) wave settings!

    DS Curves 2 pack coming soon!

    DS = Direct Signal = Control Signal = 0 to 1

    MODE 1: Logarithmic
    image
    image
    image
    image


    MODE 2: Exponential
    image
    image
    image
    image


    MODE 3: Inverse Logarithmic
    image
    image
    image
    image


    MODE 4: Inverse Exponential
    image
    image
    image
    image


    MODE 5: Relaxed Cleave Squeeze
    image
    image
    image
    image


    MODE 6: Aggressive Cleave Spread
    image
    image
    image
    image


    MODE 7: Logarithmic to Exponential Pinch
    image
    image
    image
    image


    MODE 8: Subtle Exponential to Logarithmic S-Curve
    image
    image
    image
    image
    1-1.png
    1000 x 563 - 251K
    1-2.png
    1000 x 563 - 258K
    1-3.png
    1000 x 563 - 259K
    1-4.png
    1000 x 563 - 258K
    2-1.png
    1000 x 563 - 252K
    2-2.png
    1000 x 563 - 258K
    2-3.png
    1000 x 563 - 257K
    2-4.png
    1000 x 563 - 254K
    3-1.png
    1000 x 563 - 262K
    3-2.png
    1000 x 563 - 262K
    3-3.png
    1000 x 563 - 261K
    3-4.png
    1000 x 563 - 257K
    4-1.png
    1000 x 563 - 263K
    4-2.png
    1000 x 563 - 260K
    4-3.png
    1000 x 563 - 257K
    4-4.png
    1000 x 563 - 249K
    5-1.png
    1000 x 563 - 247K
    5-2.png
    1000 x 563 - 248K
    5-3.png
    1000 x 563 - 249K
    5-4.png
    1000 x 563 - 248K
    6-1.png
    1000 x 563 - 247K
    6-2.png
    1000 x 563 - 251K
    6-3.png
    1000 x 563 - 255K
    6-4.png
    1000 x 563 - 259K
    7-1.png
    1000 x 563 - 237K
    7-2.png
    1000 x 563 - 241K
    7-3.png
    1000 x 563 - 245K
    7-4.png
    1000 x 563 - 248K
    8-1.png
    1000 x 563 - 243K
    8-2.png
    1000 x 563 - 244K
    8-3.png
    1000 x 563 - 242K
    8-4.png
    1000 x 563 - 240K
    DS Curves 1 Demo.audulus
    1M
    DS Curves 1.audulus
    479K
  • Here's an example patch - this one's a compressor/expander/gate (the shape of the curve determines which)

    image

    image

    image

    image

    image

    image

    image
    Learning Curve Compressor.audulus
    1M
    Screen Shot 2016-02-18 at 6.07.16 AM.png
    1000 x 563 - 316K
    Screen Shot 2016-02-18 at 6.07.21 AM.png
    1000 x 563 - 364K
    Screen Shot 2016-02-18 at 6.07.26 AM.png
    1000 x 563 - 309K
    Screen Shot 2016-02-18 at 6.07.30 AM.png
    1000 x 563 - 351K
    Screen Shot 2016-02-18 at 6.07.33 AM.png
    1000 x 563 - 213K
    Screen Shot 2016-02-18 at 6.07.37 AM.png
    1000 x 563 - 211K
    Screen Shot 2016-02-18 at 6.07.41 AM.png
    1000 x 563 - 250K
  • That's excellent. I guess this can't work at audio rate right?
  • @SansNom - they can, but you have to DS-offset the audio (add 1, divide by 2) and then retranslate it back into audio (multiply by 2, subtract 1). What you're going to get mostly is distortion, however - you are, after all, distorting the wave by applying a bias to the audio. This is essentially what tube distortion does, and it's what people mean by "non-linear response."

    Curves are also key to making filters that sound more "analog" - later today I will be pairing this patch up with @Afta8's filter bank and showing you how adding curves can really liven up filters.

    To make a proper tube distortion, you'll have to do non-linear distortion + non-linear filtering. It's beyond me to just say "this is how you do it" at the moment, but it's something that I will experiment with. One thing that's sorely missing from Audulus is good, versatile tube-emulating distortion, and I plan to rectify that (pun intended).

    And, again, all you need to do to make a tremolo is feed an LFO into the curve package and draw its output to a level node, then send your audio through that level node.
  • This graph shows some curves of real RCA tubes. This graph also illustrates why you have to run tubes at high voltages to get real tube distortion - any pedal that you have that has a 12AX7 in it that's not running at full heater voltage is basically window dressing. As you can see, the low plate voltage's harmonic distortion (the tube sound you want) ranges from 1-7%, whereas the high voltage tube in the same range goes from 1-~50% (goes off the chart, hard to tell).

    http://www.triodeel.com/7199d.gif
  • I'm not to sure about what you said about DS-offsetting the audio but seing that DS curves 1 is fed with values ranging from 0 to 1, I tried to input a signal coming from a sawtooth wave with the correct range and the resulting sound is heavily aliased/distorted. This is clear above 200 Hz. It shouldn't be since the wave shape chosen is typical of a slightly high-passed sawtooth wave, nothing more.
  • @SansNom Check this out (have to admit, it does sound better on sine than saw, but saw is very tricky anti-aliasing since it's such an instantaneous drop). This shows you how to translate the audio (AS) to DS, through the curve transformer (the virtual "tube"), and then back into AS. You need to pull it back to AS because if you apply DS to a speaker node, you will only be driving the speaker cone halfway through its cycle, causing added, ugly distortion.
    Sansnom Non-Linear AS2DS2AS.audulus
    1M
  • @Sansnom - I'm actually just going to go ahead and make a tube synth - will pop it up here later today - you'll see a variety of techniques at play
  • Thanks, I will check that tomorrow. However, I was not talking about speakers. I checked the waveshapes and harmonics with a signal analyzer and an oscilloscope. And offsetting to -0.5 / 0.5 or not doesn't change anything.
  • @Sansnom - yeah if you start with a sine instead, you can still make many, many tubey sounds that are killer, but I don't believe the mapper and crossfade nodes are anti-aliased, which is what would need to happen for this to be truly anti-aliased. I developed these mostly to modify control signals (which is why they're DS curves and not AS curves), but you made me think hey I could probably shunt audio through them and yes, it actually sounds very killer - it's like a tube synth but you get to choose from imaginary tubes! The number of combinations in one oscillator alone they way I'm building it is in the thousands I believe - some of them are only subtle differences, but I'm sure that once I start stacking them on top of one another, the tonal ranges will be virtually infinite.

    By the way, i call them DS and AS because calling them DC and AC is kind of a misnomer - they're not really currents, they don't really have voltages - they're just signals - so I figured hey why not just think up my own terminology. It's just something I developed for my own library, it's not a standard I expect or even want everyone to adhere to - it does, however have many advantages :) (DGS = digital signal)
  • Hmmm. I wonder how these will work with the Crossfade synth.
  • @biminiroad - since the output audio is heavily aliased I doubt it's usable this way, at least now. I tried to use the mapper node for filtering and waveshaping for my own project and I had to give up. However for LFO this is just excellent. Have to check what you say about sine and tube synths though.
  • Ze Gitter Elektron Tube Synthesizer.audulus
    4M
  • Awesome! Also love your mux /demux sub-patches. Made some nifty noise feedback oscillators with this. Is there a way to share from the iPad to the forum? I have so many things to share now but I'm too lazy to do the whole transfer and post from laptop.
  • @Macromachines - Glad you like those switches - I'm tired of making the same things over and over, and I realized I needed to build out a ton of little building block ingredients. Hopefully these will accelerate people's patch creation.

    Unfortunately, no there isn't a way to share from forum to iPad - thank Apple for their shoddy file management and basically hamstringing their products. We put a man on the moon, and yet there's some reason for iOS to be special and not let you move files around like a normal human being.

    However, you can mail yourself multiple patches at once!

    Also, in the time it took to wonder how to send patches to the forum, then make a post on the forum, then read my reply, and then possibly type out another reply to this, you could have uploaded the patches already ;)

    We are working on a better patch sharing scheme - this is only temporary.
  • @biminiroad @MacroMachines It is possible to upload patches to the forum using only an iOS device. You can email a patch to yourself (make sure you clear history in your patch by hitting the two arrow circle icon to the left of the lock) then use open in from your email to store the patch in one of these: Dropbox, Google Drive, AudioShare, etc. Then choose attach file in your post on the Audulus forum and select it from wherever you stored it.