Rescale Float List | Akatz

Most float lists that show up in animation and audio-reactive workflows arrive with an unknown range. Audio amplitudes might live in -0.3..0.9, a keyframe interpolation in 0..100, a brightness curve somewhere in between. Before you can feed any of that into a parameter that expects 0..1 (or -1..1, or 0..360), you have to rescale. AK_RescaleFloatList does that with one convenient twist: you never tell it the original range. It reads the min and max straight off the input list, so you only set the target range.

Why you'd reach for it

This is the workhorse of the pack's float-list utilities. Normalize a schedule for a mask-dilation node, remap audio levels into a CFG or denoise curve, squash a brightness list into a usable band - anywhere a raw measurement needs to become a well-behaved parameter. It's also the natural cleanup partner for the pack's AK_ListToNumpyFloatArray and AK_ShrinkNumSequence, all of which speak the same "float list through a FLOAT socket" language.

How it works

It converts the list to a NumPy array, computes orig_min and orig_max from the data, and applies the standard linear map:

new_min + (value - orig_min) * (new_max - new_min) / (orig_max - orig_min)

That's classic min-max scaling. One guardrail: if every value in the list is identical (orig_max == orig_min), it raises a ValueError rather than dividing by zero - a constant list can't be rescaled meaningfully, and the node tells you so.

Inputs and outputs

• float_list (FLOAT) - the list to rescale. A list, not a scalar.
• new_min (FLOAT, default 0) - low end of the target range.
• new_max (FLOAT, default 1) - high end of the target range. Set this pair to whatever your downstream node wants.
• output_list (FLOAT) - the rescaled list, returned as a plain Python list.

That's the whole surface. Note the docstring mentions orig_min/orig_max inputs, but they don't actually exist in the schema - the node derives them from your data, which is the feature, not a bug.

Installing the pack

Part of Akatz Custom Nodes (akatz-ai/ComfyUI-AKatz-Nodes). ComfyUI Manager: search "Akatz" and install, or clone it:

cd ComfyUI/custom_nodes
git clone https://github.com/akatz-ai/ComfyUI-AKatz-Nodes

Restart ComfyUI; it lives under 💜Akatz Nodes. Only numpy and torch are needed here (both already in ComfyUI); the pack's requirements.txt also lists opencv-python and pydub, which serve its image/audio nodes and install alongside but aren't required for this one. No models. The README is a stub that points at the author's Notion docs and a custom GPT.

Gotchas

• Set new_min and new_max in the order you want them - the node won't sort them, so a reversed pair flips your list upside down.
• A constant list is an error by design. If you're normalizing something that can legitimately be all-equal, guard it upstream.
• Same list-convention warning as the rest of the pack: feed it a genuine list output (from AK_BrightnessToFloatList, AK_KeyframeScheduler-style sources via AK_ConvertListToFloatList, etc.), not a bare scalar.