AudioToMMP

Replaces BasslineAnalyser, which was the simple first iteration.

AudioToMMP is a WIP audio analysis and transcription tool designed to bridge the gap between raw audio files and LMMS. It uses digital signal processing (DSP) to break down mixed audio, detect pitches and rhythms, and translate them directly into LMMS pattern files (.xpt), and full multi-track LMMS projects (.mmp).

Core Features

• Multi-Band Project Mode (Audio to .mmp): Define specific frequency bands (e.g., Bass, Mids, Highs) and let the app scan the entire song. It will automatically transcribe the detected notes for each band and generate a complete, multi-track LMMS project file ready for editing.
• Pattern Mode: Click and drag a selection box over a specific part of the spectrogram to isolate an instrument, melody, or bassline. The app detects the exact MIDI notes or chords (using algorithms like YIN or Harmonic Product Spectrum) and maps them onto a 16-step or 32-step pattern grid based on the BPM. This grid can be exported as an LMMS .xpt pattern.
• DSP Source Separation: A dedicated module to isolate or remove specific elements (like basslines or vocals) from the mix using cascaded, steep Biquad filters. You can play back the separated audio or save it to disk.
• Spectral Sampling: Drag an area on the spectrogram (time × frequency) to apply a bandpass filter, isolating that exact sound, and export it as a ready-to-use .wav sample.
• Automation Editor: An external editor to generate, manipulate, and inject complex automation data directly into LMMS .mmp projects.
  • Algorithmic Macros: Generate LFOs (Sine, Square, Triangle, Saw), ADSR envelopes, Random (Sample & Hold), and rhythmic gates.
  • Interactive Drawing: Double-click to add or delete points, and drag them freely across the interactive canvas just like in LMMS.
  • Math Modifiers: Smooth curves, Quantize Y (snap to grid), Humanise (jitter), Invert, and Scale Amplitude.
  • MIDI-to-CV Conversion: Load an LMMS .xpt pattern and automatically convert the notes into stepped automation (Control Voltage) using highest note priority.
  • Shape Presets: Save and load custom automation shapes using the dedicated .xpa (XML Pattern Automation) file format.
• Visual Data: Features detailed spectrogram and waveform visualisers to help you see exactly what you are extracting.
• Amiga Module Drum Extractor: Extract drum patterns from Amiga modules and export to LMMS.
• 303 Test: Test to see if the code could generate a TB-303 style baseline with cutoff and resonance automation.
• House Beat Generator: Create beats with shuffle off the 1/16 grid.
• House Piano Stabifier Select two chords and it automatically generates a dynamic, 4-bar rhythmic LMMS pattern specifically styled for House music piano stabs.

Current State

Functional note detection, offset/BPM detection, and full .mmp generation is implemented. Chord detection (Template Matching and Peak Picking) is currently active but needs further refinement.

---

Third-Party Libraries

This project relies on the following open-source libraries:

QCustomPlot

• Description: Used for plotting the interactive audio waveforms, spectrogram maps, and visual grid selections.
• Author: Emanuel Eichhammer
• License: GPLv3
• Source: https://www.qcustomplot.com/

KissFFT

• Description: A simple, lightweight Fast Fourier Transform (FFT) library used to generate the spectrogram visuals and power the frequency-based pitch/chord detection algorithms.
• Author: Mark Borgerding
• License: BSD-3-Clause
• Source: https://github.com/mborgerding/kissfft

miniaudio

• Description: A lightweight, single-file C library used for decoding and loading audio files (WAV, MP3, FLAC) into raw PCM float data for DSP analysis, as well as handling audio playback and exporting processed .wav files.
• Author: David Reid (mackron)
• License: Public Domain (Unlicense) / MIT-0
• Source: https://github.com/mackron/miniaudio