Skip to content
Projects

yumiaura/audio8bit has 92 GitHub stars and 7 forks. GitHot measured 0 stars in 24 hours and +3 in 7 days. Last checked 2026-11-15 20:00 UTC. Last push 2026-09-22.

How these counts are measured

What you can do

  • Works with vocal songs and instrumentals - picks the melody source automatically.
  • Polyphonic transcription (basic-pitch) keeps the chords and bass, reduces them to a single lead line, or arranges them as a full chip band (pulse lead, pulse harmony, a triangle bass from the bass stem and noise drums from the drums stem) - either stacked (band) or arpeggiated and beat-tight (nes).
  • Source separation with Demucs, deterministic so the same input always gives the same result.
  • Alias-free chiptune synthesis with loudness dynamics and a smooth limiter.

Measured star growth

in 24h
0
in 7 days
+3
in 15 days
Update delayed
in 1 month
Update delayed

Net star changes over overlapping periods, not daily totals. Missing measurements are left blank. Stars measure attention, not software quality.

About

CLI to convert any audio file into 8-bit / retro game style sound

Creative & Media

README.md

Original

audio8bit - turn any song into 8-bit chiptune music

CI PyPI Downloads Python

English | Español | Português | Français | Deutsch | Italiano | Русский | 中文 | 日本語 | हिन्दी | 한국어

A command-line tool that turns any song into 8-bit, video-game-style music. It finds the melody (from the vocals or the instruments) and replays it with retro "chiptune" sounds, like an old game console. Everything runs locally.

Requirements

  • Python 3.9 or higher
  • ffmpeg on your PATH (it ships ffmpeg and ffprobe). Install it with sudo apt install ffmpeg (Linux) or brew install ffmpeg (macOS).

Installation

pip install audio8bit

Or from GitHub:

pip install git+https://github.com/yumiaura/audio8bit.git

First run is slow: it downloads a small AI model (about 80 MB) and can take a few minutes. That is normal; later runs are faster.

Usage

# Convert a song (auto-detects vocal or instrumental) audio8bit -i song.mp3

# Just the main melody, no chords audio8bit -i song.mp3 -V lead

# Full chip band: pulse lead + pulse harmony + triangle bass + noise drums audio8bit -i song.mp3 -V band

# NES-style: the band with an arpeggiated harmony, snapped to the beat audio8bit -i song.mp3 -V nes

# Take the tune from the singing or from the instruments audio8bit -i song.mp3 -s vocals audio8bit -i song.mp3 -s instrumental

# Save as a different format audio8bit -i song.mp3 -f ogg

# Play it 5 semitones higher audio8bit -i song.mp3 --transpose 5

# Show help and version audio8bit --help audio8bit --version

Command Line Options

  • -i, --input - input audio file, required (any format ffmpeg can read)
  • -o, --output - output path (default: output.<ext>)
  • -f, --format - output format, e.g. ogg, wav (default: same as input)
  • -s, --source - melody source: vocals, instrumental, auto (default: auto)
  • -m, --method - note finding: transcribe or pitch (default: transcribe)
  • -V, --voices - chords (one voice, with harmony), lead (single line), band (full chip band: pulse lead + pulse harmony + triangle bass from the bass stem + noise drums from the drums stem), or nes (the same band with an arpeggiated harmony, vibrato lead and everything snapped to the beat) (default: chords)
  • --transpose - key shift in semitones (default: 0)
  • --bits - bit depth, 1-8, lower is crunchier (default: 8)
  • --rate - sample rate in Hz, lower is more retro (default: 22050)
  • --duty - pulse-wave duty cycle, 0-1 (default: 0.25)
  • --key-snap - band/nes: snap off-key notes to the detected key: on, off (default: on). The key is detected and reported either way, because --arrange builds its chords from it.
  • --arrange - band/nes: chord-progression accompaniment, root bass and looped drums instead of replaying the transcription: on, off (default: on)
  • --echo - band/nes: tempo-synced echo on the melody: on, off (default: on)
  • --dither - band/nes: TPDF dither before bit quantisation: on, off (default: on)
  • --no-cache - do not read or write cached Demucs stems
  • --cache-dir - directory for cached stems (default: ~/.cache/audio8bit, or $AUDIO8BIT_CACHE_DIR)
  • --version - show version

Exit codes: 0 success, 1 conversion error, 2 bad arguments.

Features

  • Works with vocal songs and instrumentals - picks the melody source automatically.
  • Polyphonic transcription (basic-pitch) keeps the chords and bass, reduces them to a single lead line, or arranges them as a full chip band (pulse lead, pulse harmony, a triangle bass from the bass stem and noise drums from the drums stem) - either stacked (band) or arpeggiated and beat-tight (nes).
  • Source separation with Demucs, deterministic so the same input always gives the same result.
  • Alias-free chiptune synthesis with loudness dynamics and a smooth limiter.
  • Key transposition and adjustable bit depth, sample rate and pulse tone.
  • 8-bit PCM output as WAV, or any format ffmpeg can write.
  • A quality report after every run, and everything runs on your own machine.

License

Noncommercial