Vibe Coding Log - Sharing Journey - Git - Glitch/Time Warping FX - More MIT delays, A dozen types of Lush Supersaws

[Touch The Universe]

Dialing in on a custom verb is perhaps too beyond me right now, skipping this approach, lol. feedback madness humbles quickly. i don't even know enough what to request to make it work right, hmmm.

Spoke too soon, i'm tyring to implement the cloud verb verb into it yet what we had as a starting base is much much too simplisitic so it was exploding....



I find this out so im feeding it the entire open source git zip and testing now, this thing sounds gorgeous. I have implemented freeverb, not bad, but it could be much better

[kerfuffle]

I find this out so im feeding it the entire open source git zip and testing now, this thing sounds gorgeous. I have implemented freeverb, not bad, but it could be much better

So your vibe coded synth will be licensed under the GPL?

[Touch The Universe]

Possibly, right now I'm testing things out. I thought freeverb was MIT? I have custom ones that sound decent but if I can't be happy, I will INDEED open source the synth if I can't be satisfied with the verb. Or I will request the help of people here to fine tune it in because reverb is VERY important to me.

If you meant cloudverb, yes, i I know that is gpl. I am working on a facsimile but it's sounding atricious. Everything else has been smooth sailing but the reverb has giving me the hardest trouble.

[kerfuffle]

If you meant cloudverb, yes, i I know that is gpl. I am working on a facsimile [...]

It sounded to me like you were saying you were giving the AI access to the git repo of CloudReverb and were asking it to make something similar.

[Touch The Universe]

Yes. I have the actual cloudverb embedded in to test a/b and a facsimile and two other from scratch cloud based algorythms before i found the cloudverb.

I test them in this video.




I made a thread on this topic as I hope I get lucky and somehome can help me out how to take such athing. No code necessary just a rough answer would suffice.

kvr_cloud_reverb_isolated.zip

Above is the actual dsp code






HERE IS A GREATER INDEPTH EXPLANATION OF WHAT IS GOING ON FROM OTHER THREAD





The first algorythm is the cloud verb vsti that is open source. The others are from scratch. How do they tame that feedback?



I think I’m treating feedback runaway after the fact, instead of designing the feedback network to be energy-stable from the beginning.

I am experimenting with a couple of these, it can sound lovely for 2 seconds, but then the feedback builds and its madness

I don't know enough what to even say so i have ai generate this for me. Please give me some grace on this, even though I'm just a hack who knows nothing about this stuff


I’m working on a JUCE synth/reverb project and I’m trying to build a lush “cloud” style reverb, somewhat in the territory of CloudSeed / Mutable Clouds-ish diffuse reverb / huge modulated ambient tails.

The problem I’m running into is feedback stability. The reverb can sound lush and wide at lower settings, but when feedback/decay/diffusion get pushed, the tank can build up badly, especially in the low and low-mid range. Sometimes it becomes boomy, sometimes it rings or howls, and sometimes the feedback seems to run away even if the output is clipped or limited afterward.

The current architecture I’ve been experimenting with is roughly:

multiple parallel late delay lines
allpass diffusion stages inside or around the feedback paths
modulation on delay/allpass sections
damping / shelf filters
wet/dry mix
optional freeze
safety limiting / clipping
post high-pass / low-pass / tone shaping

What I tried:

1. Directly clamping the feedback coefficient lower.
This prevents runaway, but it kills the bloom and makes the reverb feel short, flat, or too safe.

2. Coupling feedback and diffusion.
The idea was: as diffusion rises, feedback headroom falls. This helped stability, but it also changes the musical response in a way that can make the reverb less lush.

3. Adding low-frequency buildup detection / a “governor.”
I added smoothed low/low-mid envelope detection so the feedback gets reduced only after buildup is detected, rather than clamping it from the start. This helped somewhat, but it can still feel like a compressor inside the reverb. It can pump, darken, or suddenly close down the tail.

4. Post high-pass / low-pass filtering.
This reduces the audible boom, but it does not truly fix the internal tank if the instability is already happening inside the feedback loop.

5. Dynamic notch / low-mid control.
I tried targeting the area where the reverb piles up, around low-mid resonances. Again, useful, but it feels like treating the symptom after the feedback network has already become poorly behaved.

6. Saturation / tanh clipping / output limiting.
This stops digital clipping, but it does not make the reverb musically stable. It can make the tail gritty, congested, or “stuck” instead of smooth and spacious.

The question: what is the proper way to design this kind of dense cloud reverb so it remains stable and musical at high decay/diffusion settings?

More specifically:

Should every delay/allpass feedback matrix be explicitly energy-normalized?
Should the feedback matrix be unitary/orthonormal like an FDN approach?
Is it better to use a known stable FDN or Dattorro-style topology first, then add cloud/granular/modulated smear around it?
Where should damping filters go: inside every feedback loop, per delay line, or only globally?
How do you preserve bright, lush “cloud” density without low-mid runaway?
Is post-filtering/limiting always the wrong place to solve this?
How do commercial reverbs allow huge decay/freeze-like behavior without the low end accumulating uncontrollably?

I’m trying to avoid just lowering feedback until it is safe, because that removes the whole ambient bloom I’m after. I’m looking for guidance on the correct DSP structure: feedback matrix design, gain normalization, damping placement, modulation safety, and how to keep diffusion dense without the network becoming unstable.

[BertKoor]

Do you now understand why no reverb has a built-in synth?

[Touch The Universe]

Do you now understand why no reverb has a built-in synth?

[Touch The Universe]

I made good progress understanding how reverbs are generated so i feel i now enough to prompt a better sounding verb that doesn't explode. I put together a 200 page plan with gtp covering many ideas I have for the verb. It's enough for it to become its own verb/synthesis engine.

All pass filters need to diffuse the feedback loops in the matrix you choose, there are various, and there are routings to consider, and modulations smoothen out resonances so they don't bloom in one frequency too long, etc. I have half a dozen ways to tame the feedback explosion and i came up with half a dozen ways to even use those explosions as well, so they are tamed and can become useful. A hint to the ideas I'm planning on implementing..

Im now working on some unique ideas for the wavetable, very unique, though not sure if it's good as this point. Need to test if the idea has merit.

[JustinJ]

I find this out so im feeding it the entire open source git zip and testing now, this thing sounds gorgeous. I have implemented freeverb, not bad, but it could be much better

So your vibe coded synth will be licensed under the GPL?

Just so you know, although CloudVerb VST is GPL3 (probably because they're using JUCE for free), the actual CloudSeed reverb code that it uses is MIT licensed.

[Touch The Universe]

I find this out so im feeding it the entire open source git zip and testing now, this thing sounds gorgeous. I have implemented freeverb, not bad, but it could be much better

So your vibe coded synth will be licensed under the GPL?

Just so you know, although CloudVerb VST is GPL3 (probably because they're using JUCE for free), the actual CloudSeed reverb code that it uses is MIT licensed.

Thanks, this is really helpful

[Touch The Universe]

custom gui, implemeting this now in the synth. Though i will use this as just an additional algorythm. I will work with the untameable exploding code too so i can tame it myself if my own methods plus add more features, as a learning exercise and plan fun

Screenshot 2026-05-19 at 8.08.53 PM.jpeg

[Touch The Universe]

Here Is another take..I kind of like the clean look of the first, but what say you guys?

The end result of the gui of the synth might have a similar vibe with the knobs for sure, though ill likely keep the flat style too as that grows on me.

Screenshot 2026-05-19 at 8.41.49 PM.jpg

[Touch The Universe]

Screenshot 2026-05-19 at 10.37.02 PM.jpg

cloud seed now works with this gui/interface

Building an advanced version too of a seperate branch for the CS verb that is the more ambient white one.

[Touch The Universe]

I'm noticing just how much work it is to implement nice gui with layouts, even more work than compiling and test dsp code for filters, distortions etc, I might hold off all work on gui's until the heart of the synth is finished. Though oddly that html gui went up pretty quicly. Maybe it would be quicker to just implement html knobs in the gui? but would the tradeoffs be if any?


DSP stays native C++ / JUCE. HTML becomes a control surface.

That means the HTML GUI sends values like count, diffDelay, lineMod, lowFreq, etc. into the APVTS/native parameters, but it does not own the audio engine. Then we get the speed and beauty of HTML while keeping the audio stable, automatable, preset-safe, and host-friendly.

So the architecture would be:

CloudSeed DSP C++
← controlled by → APVTS parameters
← mirrored by → HTML/WebView GUI

That would let us keep the current JUCE GUI branch and the HTML GUI branch isolated, which is exactly what you wanted: tweak the pretty GUI without breaking the stable CloudSeed MIT verb.

[Touch The Universe]

Screenshot 2026-05-20 at 1.19.35 AM.jpg

after 40 passes back and forth, builds with gtp, reloading studio one, manually copying the dll, replacing dsp files, at this point i might want to use cursor or windsurf to manage these tasks for me, when its tedious like this, the dsp code is not so bad though, as the results are fast adn engaging. This is bland