Normalized Amplitude To Graph

NormalizedAmplitudeToGraph is the debugging half of the pack's "verify before you trust" story. It takes a NORMALIZED_AMPLITUDE series and renders it as an IMAGE you can preview - a line plot of amplitude over frames. It's AmplitudeToGraph's twin, aimed at the normalized 0–1 curve instead of the raw one.

Why you'd reach for it: the pack's whole pitch is "drive an animation with a 0–1 signal," and that signal can lie. Normalize a curve with one giant transient and you'll get a mostly-flat line; gate it at the wrong threshold and it collapses to zeros; invert it and everything flips. This node is how you actually see whether NormalizeAmplitude, GateNormalizedAmplitude, and their friends did what you expected - before you burn a full render discovering the curve was garbage. In the README's example chain it's also just the natural "show me what the music looks like" node.

Mechanically it's simple: matplotlib line plot, frame number on the x-axis, amplitude on the y-axis, rendered to a PNG in memory and passed out as an IMAGE tensor. The plot width scales with the length of your series (capped, so very long tracks don't produce absurd images). Wire the output to a PreviewImage node to see it.

Inputs: just normalized_amp (NORMALIZED_AMPLITUDE). Output: graph_image (IMAGE).

Two honest notes. First, this is a static render of the whole series - a snapshot of your curve, not a live meter, so it won't help you tune in real time. Second, a flat-looking graph is almost never this node's fault; the bug is upstream in your normalization or thresholding, so trace back through the chain. Install comes with the pack: ComfyUI Manager search ComfyUI-AudioScheduler, or git clone https://github.com/a1lazydog/ComfyUI-AudioScheduler into custom_nodes and restart. CPU-only, no models.