Complete the voice effect harvesting from 'previous'; other voice effects we should add

[baconpaul]

The only part of 'previous' (the SC2 code) we haven't really scoured for the new code yet is the voice effects. We have most other things ported over.

So this issue is kinda a checklist of both (1) voice effects I want to add which weren't in SC2 and (2) the ones in previous we haven't ported yet. More a note to self and a place to park other ideas

The code in 'previous'

• [x] filters_chorus.cpp
  • Just contains an early version of the existing surge chorus. Not a good voice effect candidate and the bus effect will come from surge back port
• [x] filters_delay_based.cpp
  • COMB1 and COMB3 are simple comb filters. Porting the surge filter bank will obviate the need for these
• [x] filters_destruction.cpp
  • BF is already ported
  • OD is just a hard clip with some filters and polarity. Waveshaper already does better
  • [x] Treemonster. We definitely want a voice- as well as bus- level tree monster so some work here
  • clipper is covered by the wave shaper effect already
  • fdistortion is covered by the wave shaper effect
  • microgram is already ported
  • [x] fslewer seems to be a simple slew and pair of eq filters. Looks kinda neat actually and compact. Port.
• [x] filters_dualdelay.cpp
  • Bus level dual delay. Already ported via surge. Ignore
• [ ] filters_dynamics.cpp
  • [ ] limiter. Definitely investigate and perhaps make this code voice level available
  • [ ] gate. same. what does it do and do we want it
  • [ ] stereo-tools. Unclear. Do we want a MS/SM voice effect option for doing MS/SM processing?
• [x] filters_eq.cpp
  • [x] morph eq is a preset based eq set. I think we can probably skip it especially since I don't think we have the original presets
  • eq2band is ported
  • [x] eq6band graphic eq is neat and we should port it - but its all peaks. do we want low and high shelves instead?
  • LPHP Morph looks super easy to port and but the code it calls in BiquadFilter both in surge and shortcircuit ignores the morph parameter. So instead we should just remove that coefficient method from the Biquad since it is incomplete
  • COMB2 is covered either by surge filters or by the new string-like idea below
• [x] filters_luxury.cpp
  • Basically the original surge filters. Skip
• [x] filters_modulation.cpp
  • [x] RING. yes please. But we can probably do a bit better than the version here when we port
  • [x] FREQSHIFT. Also yes please. A dual ring mod based pitch shifter. Really the pitch ring is better than this. We can skip
  • PMOD already ported
  • rotary speaker: bus effect. Port from surge with our various upgrades
  • phaser same thing
• [x] filters_oscillator.cpp
  • These are all done
• [x] filters_reverb.cpp
  • Basically an early version of reverb1. Already ported
• [x] filters_supersvf.cpp
  • This is already ported, but it is also just one of the modes of set-filters, so when we bring in the sst filter set, we can remove the port
• [x] filters_traditional.cpp
  • This basically exposes all the coefficient makers on the BiquadFilterUnit as user exposed. We probably don't want it. if we want a super lightweight variable filter we can use the simper SVF; if we want better filters we have ssd-filters. So ignore this.
• [ ] filters_v1effects.cpp
  • [x] Faux Stereo looks fun. Delay line based stereo field expansion (MS -> comb for side -> SM)
  • fs_flange better with surge flanger as a bus effect I think (but do we want a voice level flange? I don't think so)
  • [ ] freqshiftdelay also looks fun - basically the freq shift from modulation inside a badly interpolated delay line. we can do the same thing but better for sure (and can probably do it with the pitch ring frequency shifter even better)

New voice effects to add or consider

• [x] The surge filter set, obviously. This implies we can do inter-twined parameters in the metadata in a voice effect since the range and values for the sub-type depends on the type. Probably also requires us to finally clean up some of those subtype name fields in set-filters
• [x] A 'mini-string' generator with the input stream as exciter; basically a feedback comb filter which in a zone can get tuned. Sort of an in-line combulator in the voice effect stance
• [x] A 'short delay'; basically a sub one second delay line which you can top in a zone (sub one second to keep the memory pool bounded) for slap backs and so on.
• [x] As mentioned above perhaps some for of voice-modulation effect (flanger, chorus etc) could be interesting but no commitment now
• [x] Add the simper SVF (including the extended modes) as a super-light-weight filter option
• [x] A voice-level phaser with the Simper SVF APF should be quite nice and easy to add

also

• [x] micro here at ofersamole uses the wrong rate

[mkruselj]

morphEQ presets are in SC1, I have all of them.

[baconpaul]

Oh awesome can you zip them and add them Here?

[mkruselj]

It's all in one XML file, here goes:

<?xml version="1.0" encoding="utf-8" standalone="yes" ?>
<morphEQ>
    <snapshot name="flat" gain="0">
    </snapshot>
    <snapshot name="tone LP 1" gain="-6">
        <band i="1" freq="-2" gain="24" BW=".1" />
        <band i="2" freq="-1" gain="24" BW=".1" />
        <band i="3" freq="0" gain="24" BW=".1" />
        <band i="4" freq="1" gain="24" BW=".1" />
        <band i="5" freq="2.0" gain="24" BW=".1" />
        <band i="7" freq="0"/>
    </snapshot>
    <snapshot name="tone LP 2" gain="-6">
        <band i="1" freq="-.2" gain="24" BW=".01" />
        <band i="2" freq="-.1" gain="24" BW=".01" />
        <band i="3" freq="0" gain="24" BW=".01" />
        <band i="4" freq=".1" gain="24" BW=".01" />
        <band i="5" freq=".2" gain="24" BW=".01" />
        <band i="7" freq="0"/>
    </snapshot>
    <snapshot name="Bell labs" gain="-40">
        <band i="0" freq="-6"/>
        <band i="1" freq="-1" gain="50" BW="0.02" />
        <band i="2" freq="1" gain="40" BW="0.04" />
        <band i="3" freq="1.58" gain="40" BW="0.06" />
        <band i="4" freq="2.07" gain="40" BW="0.07" />
        <band i="5" freq="2.43" gain="40" BW="0.08" />
        <band i="6" freq="2.77" gain="40" BW="0.1" />
        <band i="7" freq="6"/>
    </snapshot>
    <snapshot name="combs" gain="0">
        <band i="1" freq="0" gain="-80" BW="0.01" />
        <band i="2" freq="1" gain="-80" BW="0.01" />
        <band i="3" freq="1.58" gain="-80" BW="0.01" />
        <band i="4" freq="2" gain="-80" BW="0.01" />
        <band i="5" freq="2.32" gain="-80" BW="0.01" />
        <band i="6" freq="2.58" gain="-80" BW="0.01" />
    </snapshot>
    <snapshot name="random dips" gain="0">
        <band i="1" freq="0" gain="-80" BW="0.01" />
        <band i="2" freq="0.7" gain="-80" BW="0.01" />
        <band i="3" freq="1.6" gain="-80" BW="0.01" />
        <band i="4" freq="2.3" gain="-80" BW="0.01" />
        <band i="5" freq="3.1" gain="-80" BW="0.01" />
        <band i="6" freq="4.54" gain="-80" BW="0.01" />
    </snapshot>
    <snapshot name="ups and downs" gain="0">
        <band i="1" freq="0" gain="-80" BW="0.01" />
        <band i="2" freq="1" gain="12" BW="0.1" />
        <band i="3" freq="1.58" gain="-80" BW="0.01" />
        <band i="4" freq="2" gain="12" BW="0.1" />
        <band i="5" freq="2.32" gain="-80" BW="0.01" />
        <band i="6" freq="2.58" gain="12" BW="0.1" />
    </snapshot>
    <snapshot name="formant A1" gain="-40">
        <band i="1" freq="0.686" gain="60" BW="1.6" />
        <band i="2" freq="1.33" gain="30" BW="0.13" />
        <band i="3" freq="2.51" gain="30" BW="0.07" />
        <band i="4" freq="2.77" gain="30" BW="0.05" />
    </snapshot>
    <snapshot name="formant A2" gain="-40">
        <band i="1" freq="0.433" gain="60" BW="1.8" />
        <band i="2" freq="1.01" gain="30" BW="0.15" />
        <band i="3" freq="2.51" gain="30" BW="0.07" />
        <band i="4" freq="2.77" gain="30" BW="0.05" />
    </snapshot>
    <snapshot name="formant E" gain="-40">
        <band i="1" freq="0.35" gain="60" BW="2.0" />
        <band i="2" freq="2.05" gain="30" BW="0.11" />
        <band i="3" freq="2.51" gain="30" BW="0.07" />
        <band i="4" freq="2.77" gain="30" BW="0.05" />
    </snapshot>
    <snapshot name="formant I" gain="-40">
        <band i="1" freq="-0.65" gain="60" BW="2.4" />
        <band i="2" freq="2.357" gain="30" BW="0.1" />
        <band i="3" freq="2.51" gain="30" BW="0.07" />
        <band i="4" freq="2.77" gain="30" BW="0.05" />
    </snapshot>
    <snapshot name="formant O" gain="-40">
        <band i="1" freq="0.04" gain="60" BW="2.0" />
        <band i="2" freq="0.87" gain="30" BW="0.3" />
        <band i="3" freq="2.51" gain="30" BW="0.07" />
        <band i="4" freq="2.77" gain="30" BW="0.05" />
    </snapshot>
    <snapshot name="formant U" gain="-40">
        <band i="1" freq="-0.5" gain="60" BW="2.3" />
        <band i="2" freq="0.99" gain="30" BW="0.4" />
        <band i="3" freq="2.51" gain="30" BW="0.07" />
        <band i="4" freq="2.77" gain="30" BW="0.05" />
    </snapshot>
</morphEQ>

[baconpaul]

OK so here's what morphEQ is doing then

1. A lot of code to load that data table from XML, which I think we shouldn't repeat. Lets just make the data table part of the effect
2. Once you have that you have two int params (preset 1 and 2), a percent morph, and then a frequency bandwidth and gain offset
3. Process is if anything changed calc coeffs input.scale-by-gain for (auto f : bandfilters) if (f.active) f.process
4. calc_coeffs is where the work is. 8 bands of biquad. band 0 is a HP band 7 is a LP, rest are peak each has frequency of (preset0 (morph-1) + preset1 morph + frequency offset) and same for bandwidth LP and HP hardcode Q to 1/root 2
   5. frequency graph is the obvious choice

so looks like an easy and useful voice effect to port indeed!

[baconpaul]

Andreya points out that https://github.com/bogaudio/BogaudioModules/blob/master/src/Pressor.cpp the BOG "Pressor" compressor is a great example of a full featured basic compressor which would be very useful as a voice or group effect.

[Andreya-Autumn]

I do indeed. :) I'd be happy to take a swing at porting it, maybe next week. But if you feel the urge before then, by all means don't hold off!

[baconpaul]

I took a first swing at it but there's a few other bits and bobs i want to borrow and factor (I think I'm going to port their excellent envelope follower to basic blocks, for instance). So I've got it I think.

[Andreya-Autumn]

Nice! And (stating the obvious), also putting it in Surge (not voice-level, granted) for 1.3.2 would meet a very common feature request there.

[baconpaul]

Yeah. Yeah it would.

Right now bus-effects and voice-effects have different APIs (understandably) so I'm thinking about how to code it so we can project it to either.

[baconpaul]

https://github.com/bogaudio/BogaudioModules/issues/238

Just linking that here.

[Andreya-Autumn]

Right! I wonder how common the "both" case is going to be there. If it stays this rare perhaps it's easiest to just incude in both spots? But I don't know, maybeit won't be.

[Andreya-Autumn]

Nice

[baconpaul]

Right! I wonder how common the "both" case is going to be there. If it stays this rare perhaps it's easiest to just incude in both spots? But I don't know, maybeit won't be.

Yes I'm thinking about that right now. It's actually pretty easy to just push down one more level of implementer template. I'll use Pressor as a test case.

[baconpaul]

I'm going to close this issue as soon as tree monster is done, but only if tree monster is done with the bus and voice effect sharing dsp.