Music - Reverb

Every piece of generated audio suffers from the same tell: it's too dry. AI music and TTS vocals come out of the model completely un-reverbed, and nothing screams "synthesized" faster than sound with zero sense of space. Music - Reverb from ComfyUI MusicTools is the pack's answer - a deliberately lightweight multi-tap ambience effect with exactly one control.

The one knob

decay (0–1, default 0.5) - how long the reverb tail rings out. That's the entire interface. One AUDIO in, one AUDIO out.

The "multi-tap" in the description is the mechanism: instead of a full convolution reverb with an impulse response (heavy on CPU and requires an IR file), it uses a handful of delayed, decaying taps - a short delay line with feedback that's cheap enough to run near real-time on CPU. For adding a sense of room to a dry vocal or instrumental, that's plenty. The default 0.5 is a moderate ambience; push toward 0.8+ for a long tail, drop below 0.3 for just a touch of space.

Where it fits

The honest use cases:

• A dry AI vocal that sits in a void - a little reverb (0.3–0.4) is the single biggest "this sounds less synthetic" upgrade you can make.
• Blending a processed stem back into a mix - a touch of ambience makes a separately-processed vocal sit in the track instead of on top of it.
• Making a track feel bigger - subtle, not cathedral.

It's also built into Music - Master Audio Enhancement as a reverb_amount control, so the standalone node is for the cases where you want reverb without the full mastering chain - say, on just the vocal stem after separation, or as the last touch on a TTS narration before loudness normalization.

The honest take

This is not a "nice reverb" in the way that, say, a good convolution plate or algorithmic hall is. It's an ambience effect - its job is "add space," and it does that without pretending to be a high-end space designer. If you need lush, characterful reverb, this isn't the node and you should look at real audio tools. If you need "my generated vocal sounds like it's in a room instead of in a vacuum" with zero setup, it's exactly right.

One practical note: reverb before compression sounds different from reverb after. Reverb after compression (or after the limiter) can push the tail back into the mix unevenly; reverb before loudness normalization means the normalization sees the full ambience. For the "add space to a finished track" job, put it near the end - after EQ and compression, before the final loudness step.

Install

Standard pack install:

cd ComfyUI/custom_nodes
git clone https://github.com/jeankassio/ComfyUI_MusicTools.git
cd ComfyUI_MusicTools
python -m pip install -r requirements.txt

Restart, find it under music. No IR files, no models, no GPU - just a short feedback delay, which is exactly why it's instant.