AudioAugmenter

Desktop GUI for importing, augmenting, and exporting WAV/MP3 samples.

Features

• Drag-and-drop import into a local workspace/ directory.
• Central table to edit transcriptions, delete rows, and track applied augmentations.
• One-click augmentation (Gaussian noise, time stretch, pitch shift) via audiomentations.
• Export to CSV or JSON with optional audio copies.

Screenshot

Requirements

• Python 3.13+
• Dependencies: PyQt6, audiomentations, pandas, numpy, soundfile, pydub (see pyproject.toml).

Quick start

python -m venv .venv
source .venv/bin/activate
pip install -e .
python app.py

Usage

• Use Add Files or drag .wav/.mp3 files into the left pane. Files are copied into workspace/.
• Edit transcriptions directly in the middle table.
• Select rows, check one or more augmentations (edit parameters in the panel), and click Apply Augmentation to create a single derived file per source with all checked augmentations applied in order.
• Delete selected rows and their workspace files with Delete Selected.
• Export selected rows (or all if none selected) to CSV/JSON; exported rows are reindexed from 0 and filenames reset to rec_<n>.wav; audio copies go to exports/audio/ by default.
• Use Save to workspace to persist the current table to workspace/entries.csv; restarting the app reloads this state.
• Double-click a row to play the audio (requires pydub playback support, e.g., simpleaudio).

Notes

• Augmentations rely on audiomentations and soundfile; ensure their native deps are available on your system.
• Augmented outputs are written as WAV files to maximize compatibility.
• Playback is not included; integrate pydub/sounddevice if needed.