Frequency Filter Preset ⚡🅡🅞🅣🅘

Building a frequency filter by hand means picking cutoff frequencies and filter types and hoping you land in the right range - "isolate the kick drum" isn't obvious as a Hz number if you're not an audio engineer. Frequency Filter Preset skips that entirely: pick a named preset, and it builds the underlying FREQUENCY_FILTER for you, tuned to a specific, common goal.

How it works

preset is the whole interface - eight named options covering the two things people actually want a frequency filter for: isolating a specific element (isolate_kick_drum, isolate_snare_drum, isolate_hi_hat, isolate_bass, isolate_vocals) or cleaning up a mix (remove_rumble, brighten_mix, remove_hiss). Each one packages a band-pass or notch configuration tuned to that goal, so you're choosing an intent rather than dialing in numbers. This node doesn't touch your audio directly - it produces a FREQUENCY_FILTER object, the same type this pack's AudioFilter node consumes to actually run the filter over your track. previous_filter is the chaining mechanism: feed in an earlier preset (or any other node that outputs a FREQUENCY_FILTER) and this one appends to it, so you can stack presets - isolate_vocals followed by remove_hiss, say - into one combined filter instead of running multiple separate filter passes.

Inputs and outputs that matter

• preset (required) - one of the eight named goals: isolate kick/snare/hi-hat/bass/vocals, or remove rumble/hiss/brighten mix.
• previous_filter (optional, FREQUENCY_FILTER) - chain onto an earlier filter definition instead of replacing it.
• Output - a single FREQUENCY_FILTER, which feeds into this pack's AudioFilter node (or anywhere else in the pack that accepts a FREQUENCY_FILTER) to actually apply it to audio.

Installing it

Via ComfyUI Manager: search "RyanOnTheInside," install, restart. Manually:

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

then restart. The pack's audio tools lean on librosa and related signal-processing libraries for the underlying filter math - that install step is what pulls those in, no separate model download needed.

Common issues & troubleshooting

Nothing happens to your audio. This node only defines a filter - it doesn't apply one. You still need to wire its output into AudioFilter (or another consumer of a FREQUENCY_FILTER) alongside your actual audio input; a Frequency Filter Preset sitting alone in a graph does nothing audible.

Preset doesn't isolate cleanly. These are general-purpose presets, not tuned per-track - a kick drum sitting unusually high in frequency for a specific mix, or vocals that overlap heavily with another instrument's range, won't separate perfectly with a fixed preset. Treat these as a fast starting point, not a mastering-grade solution.

Chained presets produce a muddier or overly narrow result than expected. previous_filter stacks filters rather than averaging them - two aggressive presets chained together (say, isolate_vocals plus isolate_bass) can end up fighting each other rather than combining cleanly, since one is keeping a narrow band and the other is keeping a different one. Think through what each preset actually keeps before chaining more than one or two.