Audio Mask Analyzer ⚡🅡🅞🅣🅘

Audio masks in this pack - the kind AudioRegionMask and AudioTemporalMask produce - are invisible in the way image masks aren't. You can preview a black-and-white image mask at a glance; an audio mask is just a tensor of numbers over time, and unless you know exactly what shape and range it should be, you're flying blind about what region of your ACE-Step generation you actually just marked for change (or protection). This node exists to turn that tensor into something you can read.

How it works

Feed it a mask, and it inspects the values and outputs a plain string. Think of it as a sanity-check step: before you wire a hand-built or feature-driven mask into an ACE-Step repaint/extend operation and burn generation time on it, drop this node on the wire and route its output into a text-display node so you can eyeball what the mask is actually describing - which stretch of the timeline is active, roughly how strong, that sort of thing - instead of discovering it was wrong after the render.

The inputs and outputs that matter

There's exactly one input and one output, which is the whole appeal - it's a lightweight diagnostic, not a processing step:

• mask (MASK, required) - the audio mask you want to inspect.
• Output - a single STRING describing it. Route this into any "show text" style node to actually read it.

How to install it

Via ComfyUI Manager: search "RyanOnTheInside," install, restart. By hand:

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

then restart. No model download involved - this is pure inspection of a tensor you already have in your graph.

Common issues & troubleshooting

You wired it up but see nothing. The node only produces a string - it doesn't display anything by itself. You need a separate text-preview node (any "Show Text" style node works) downstream of it, or the analysis just sits there unread.

The mask you're analyzing came from outside this pack. This node expects the standard MASK type, so it'll happily take a mask from anywhere, but it was clearly built and tested against this pack's own audio-mask nodes. If you're feeding it something exotic, treat the output as a rough read rather than gospel.

Timing looks off compared to what you expect. Trace it back to the mask-generating node, not this one - AudioMaskAnalyzer just reports what's in the tensor it was handed. If the numbers look wrong, the bug is almost always upstream in how the mask was built.