AudioProcessor v2.0

High-performance audio and video transcription for Windows x64 using Faster-Whisper (CTranslate2) with automatic NVIDIA GPU acceleration.

What It Does

Converts audio/video files into professionally formatted DOCX transcripts. Automatically detects and uses NVIDIA GPUs for up to 8x real-time processing, with CPU fallback.

Key Features

• Faster-Whisper Large-v3 — best accuracy, ~4x real-time on GPU (int8)
• Faster-Whisper Large-v3-turbo — 2x faster, half the VRAM, excellent quality
• Native Whisper Large-v3 — OpenAI fallback (GPU or CPU)
• Vintage tape preprocessing — noise reduction, loudness normalisation, dynamic range compression optimised for 1980s–90s recordings
• Single file or recursive batch — process one file or an entire folder tree
• Skip / replace / replace-before-date — resume interrupted batches without re-transcribing
• Per-project settings — model, recursive mode, and replace policy saved per folder
• Domain terms — feed a special_words.txt to improve recognition of specialist vocabulary
• Clean componentised GUI — gui_transcribe.py (main) + gui_components.py (panels)

Quick Start

One-Liner Install (GeForce GPU)

Open PowerShell and paste:

irm https://raw.githubusercontent.com/Rob142857/AudioProcessor-v2.0/main/install_geforce.ps1 | iex

This installs Python, Git, FFmpeg (via winget), clones the repo, sets up a venv, installs PyTorch CUDA 12.4, downloads models, and launches the GUI.

Manual Setup

git clone https://github.com/Rob142857/AudioProcessor-v2.0.git
cd AudioProcessor-v2.0
python -m venv .venv
.\.venv\Scripts\Activate.ps1
pip install -r requirements.txt
# Install PyTorch for your GPU — example for CUDA 12.4:
pip install torch --index-url https://download.pytorch.org/whl/cu124

# Cache the models (~4.7 GB total)
python preload_models.py

# Launch
python gui_transcribe.py

Or use the one-click bootstrap:

.\run.bat

Requirements

• Windows 10/11 x64
• Python 3.11+
• 8 GB+ RAM (16 GB recommended)
• NVIDIA GPU with 8 GB VRAM (GTX 1070 Ti or newer recommended)
• ffmpeg on PATH