KORE — Vocal-Karaoke Mixing Tool

KORE is a vocal-karaoke mixing pipeline that runs noise reduction, vocal effects (high-pass filter, compressor, reverb), and loudness-normalized mixing from the command line.

Installation

Requires Python >= 3.11 and uv.

git clone https://github.com/MrityunjayRoy/kore && cd kore
uv venv
uv sync

This installs the kore command into the virtual environment. Activate with source .venv/bin/activate.

Usage

kore <vocal> <karaoke> [options]

Positional arguments

Argument	Description
vocal	Path to vocal audio file
karaoke	Path to karaoke/instrumental

General options

Option	Default	Description
-o, --output	kore_output.wav	Output path for final mix
--gain	1.0	Output gain multiplier
--keep-denoised	—	Save intermediate denoised vocal
--keep-processed	—	Save intermediate processed vocal

High-pass filter

Option	Default	Description
--highpass-cutoff	80.0	Cutoff frequency in Hz

Compressor

Option	Default	Description
--compressor-threshold	-20.0	Threshold in dB
--compressor-ratio	3.0	Ratio
--compressor-attack	2.0	Attack in ms
--compressor-release	50.0	Release in ms

Reverb

Option	Default	Description
--reverb-room-size	0.3	Room size 0–1
--reverb-damping	0.5	Damping 0–1
--reverb-wet	0.2	Wet level 0–1
--reverb-dry	0.6	Dry level 0–1
--reverb-width	0.8	Stereo width 0–1

Mixer / loudness

Option	Default	Description
--vocal-target	-9.0	Vocal peak target in dBFS
--karaoke-target	-13.0	Karaoke peak target in dBFS
--loudness-target	-14.0	Integrated loudness target in LUFS
--true-peak	-1.0	True peak ceiling in dB

Example

kore vocal.wav karaoke.mp3 -o mix.wav \
  --highpass-cutoff 100 \
  --compressor-threshold -24 \
  --reverb-wet 0.25 \
  --vocal-target -10 \
  --loudness-target -14

Pipeline

1. Noise reduction — DeepFilterNet removes background noise and reverb from the vocal.
2. Vocal effects — A pedalboard chain applies high-pass filter → compressor → reverb.
3. Mix & master — Peak-normalized vocal and karaoke are summed, then loudness-normalized to the target LUFS and true-peak limited.

Standalone subcommands

Run kore --help for a full list of options.

Standalone subcommands

Each stage can be run independently (also available in PATH after install):

python -m app.cli.noise-reducer <input> -o denoised.wav
python -m app.cli.vocal-effects <input> -o processed.wav [effect options]
python -m app.cli.mixer <vocals> <karaoke> -o mixed.wav [mix options]