⚡ Chaos Noise Mix

AudioMixer is the polite way to combine two signals: add them, pan them, done. ChaosNoiseMix is the rude way. It takes two audio inputs and combines them using operations most people have only seen in a programming class - XOR, modulo, ring modulation, even a raw "chaos" mode. If you've ever wanted your noise to sound like a hard drive in distress, this is the node.

What it does

Two required inputs, noise_a and noise_b, combined by a mix_mode into one chaos_audio output. The modes are where the personality lives:

• add / subtract / multiply - familiar mixing math.
• max / min - pick the louder (or quieter) sample at every point. Creates aggressive interleaving between the two sources.
• xor / modulo - bitwise and modular arithmetic on the samples. Pure destruction; samples get scrambled into digital screech. This is the mode that makes people ask what you're doing.
• ring_mod - amplitude modulation between the two signals, the classic metallic bell-tones trick.
• chaos - the mode that leans on the node's chaos_amount to inject unpredictable behavior into the mix.

On top of the mode, you get mix_ratio (how much of B leaks into the result, 0–1), distortion (0–1, applied to the mix), an amplitude master, and - the one that makes this a self-feeding monster - an optional feedback amount up to 0.8. Feed that and the output feeds back into itself, turning a static mix into something that evolves on its own.

How it works

Sample-by-sample combination: each mode is a rule applied to the two waveforms at every point in time. mix_ratio blends the raw combined signal with the input, distortion pushes the result through a waveshaping curve, and feedback mixes a delayed copy of the output back in. The result can get loud and spiky fast, which is the point - but it's also why the amplitude controls exist.

The inputs that matter

• mix_mode - the whole decision. Start with ring_mod or add to stay sane; graduate to xor and modulo when you want to break things.
• mix_ratio and feedback - the two dials that determine whether this is a texture or a full-blown meltdown. Keep feedback under ~0.4 until you know what you're in for.
• amplitude - treat this as a safety knob and check your levels; chaos modes can spike well past 1.0.

Install

Standard pack install:

cd ComfyUI/custom_nodes
git clone https://github.com/eg0pr0xy/comfyui_noisegen.git
cd comfyui_noisegen
pip install -r requirements.txt

Windows embedded Python: ComfyUI\python_embeded\python.exe -m pip install -r requirements.txt. Or ComfyUI Manager, search "NoiseGen" / eg0pr0xy. No models, no keys.

Gotchas

The pack's whole dependency rule applies: don't skip the pip install, since scipy is imported at load and the pack won't register without it. And this node rewards gentle hands - the README frames it as "extreme audio mixing for harsh noise," which is marketing for "turn the mix_ratio down first." The jump from add at 0.5 to xor at 0.5 is not a small jump. Also worth knowing: it needs two inputs, so it's a combiner, not a standalone source - pair it with NoiseGenerators or external samples (VideoHelperSuite's VHS_LoadAudio works fine) and let it glue them into something unrecognizable.