📚 Goethe - German Sentence Audio Extractor

Learn German by building your own audio library from videos!

🎯 What is Goethe?

Goethe is a tool that helps you learn German by extracting specific sentences from your videos and saving them as MP3 files.

The Problem

I started learning German and wanted to use Anki for vocabulary practice. Anki is great for learning with audio, but I had a problem:

📹 German learning videos (hours of content)
        ↓
✂️ I needed to cut specific sentences
        ↓
🎧 Convert them to MP3 files
        ↓
📱 Add them to Anki for practice
        ↓
⏰ This took HOURS of manual work!

Every time I wanted to add a new sentence to Anki:

I had to find the exact position in the video
Cut the audio manually
Convert it to MP3
Save it with the right name
Add it to Anki

This was wasting my study time. I wanted to learn German, not become a video editor!

The Solution

So I built Goethe to do all this work automatically:

📹 German Lesson (30 min)
        ↓
🔍 You write: "Ich habe keine Zeit" in sentences.txt
        ↓
🤖 Goethe finds it in the video (2:15 - 2:18)
        ↓
✂️ Automatically cuts the audio
        ↓
🎵 Creates: Ich_habe_keine_Zeit.mp3 (3 seconds)
        ↓
📱 Ready to add to Anki in seconds!

Now I can:

✅ Learn German instead of editing videos
✅ Build my Anki audio library quickly
✅ Practice specific sentences anytime
✅ Save hours of manual work

If you're learning German and using Anki, this tool is for you! 🎧

✨ Features

Feature	Description
🎤 Speech Recognition	Converts video speech to text with timestamps
🔍 Smart Search	Finds sentences even with small differences (fuzzy matching)
🎵 MP3 Extraction	Creates high-quality audio clips (192 kbps)
📁 Organized Output	MP3 files organized by sentence in individual folders
🎨 Beautiful Interface	Colorful, professional output in your terminal
🔄 Resume Support	Won't reprocess videos you've already done
⚡ Fast Processing	Uses optimized whisper models for speed
🚀 Auto-Setup	Everything created automatically on first run
🇩🇪 German Optimized	Handles ä, ö, ü, ß and German grammar

📋 Requirements

Before you start, make sure you have:

Requirement	Version	How to check
Ubuntu	22.04 or later	`lsb_release -a`
Python	3.8 or later	`python3 --version`
FFmpeg	Latest	`ffmpeg -version`
Disk Space	10+ GB free	`df -h`

🚀 Quick Installation

Step 1: Clone the Repository

git clone https://github.com/duman-dp/Goethe.git
cd Goethe

Step 2: Install FFmpeg (if not installed)

sudo apt update
sudo apt install ffmpeg

Step 3: Setup Python Environment

# Create virtual environment
python3 -m venv venv

# Activate it
source venv/bin/activate

# Install dependencies
pip install -r requirements.txt

Step 4: Run Auto-Setup

python3 setup.py

This creates all necessary folders and configuration files automatically.

Step 5: Add Your Content

# Add videos to the videos folder
cp /path/to/your/german_video.mp4 videos/

# Add sentences to sentences.txt
nano sentences.txt

Step 6: Start Extracting

python3 main.py

📁 Project Structure

Goethe/
│
├── 📁 videos/              ← Put your videos here (.mp4, .mkv, .avi)
│   └── german_lesson.mp4
│
├── 📁 transcripts/         ← Speech-to-text results (JSON)
│   └── german_lesson.json
│
├── 📁 output/              ← All generated files
│   └── 📁 mp3/             ← Your MP3 library! 🎵
│       ├── 📁 Ich_habe_keine_Zeit/
│       │   └── 🎵 Ich_habe_keine_Zeit_001.mp3
│       ├── 📁 Wie_geht_es_dir/
│       │   └── 🎵 Wie_geht_es_dir_002.mp3
│       └── 📁 Das_ist_kein_Problem/
│           └── 🎵 Das_ist_kein_Problem_003.mp3
│
├── 📁 modules/             ← Python modules (don't edit unless you know)
│   ├── speech_recognition.py
│   ├── sentence_matcher.py
│   ├── audio_extractor.py
│   ├── beautiful_logger.py
│   └── setup.py
│
├── 📄 config.yaml          ← Configuration settings
├── 📄 sentences.txt        ← Sentences you want to find
├── 📄 main.py              ← Main program (run this)
├── 📄 setup.py             ← First-time setup
├── 📄 requirements.txt     ← Python dependencies
└── 📄 README.md            ← This file

⚙️ Configuration

Edit config.yaml to customize how Goethe works:

# ============================================
# Goethe Configuration
# ============================================

# Audio Settings
audio:
  padding_before: 0.03    # Add 0.03s before each sentence (natural sound)
  padding_after: 0.03     # Add 0.03s after each sentence
  bitrate: 192           # MP3 quality: 128, 192, 256

# Speech Recognition
speech:
  model_size: medium       # tiny | small | medium | large (best)
  language: German       # Language to transcribe
  device: cpu            # cpu or cuda (GPU)
  quiet: true            # Less output in terminal

# Sentence Matching
matching:
  fuzzy_threshold: 60    # 0-100 (lower = more matches, but less accurate)

# Paths
paths:
  videos: ./videos
  transcripts: ./transcripts
  output: ./output
  logs: ./logs

Model Size Guide

Model	Speed	Accuracy	Memory	Best For
tiny	🚀 Fastest	70%	~1 GB	Testing, short videos
small	⚡ Fast	85%	~2 GB	Daily use
medium	🐢 Medium	92%	~5 GB	Better accuracy (recommended)
large	🐌 Slow	95%	~10 GB	Best quality (if you have RAM)

📝 How to Use

Adding Videos

# Copy videos to the videos folder
cp /path/to/video.mp4 videos/

# Supported formats:
# .mp4, .mkv, .avi, .mov

Adding Sentences

Open sentences.txt:

nano sentences.txt

Add one sentence per line:

Ich habe keine Zeit.
Wie geht es dir?
Das ist kein Problem.
Guten Morgen!
Hallo, wie heißt du?
Ich liebe Deutsch.
Wo ist der Bahnhof?
Ich verstehe nicht.
Kannst du mir helfen?
Was ist das?

Tip: Use shorter sentences for better matches!

Running the Program

python3 main.py

Finding Your MP3s

# List all MP3 files
ls -R output/mp3/

# Play all MP3s
for f in output/mp3/*/*.mp3; do ffplay -autoexit "$f"; done

🎨 Example Output

  ╔════════════════════════════════════════════════════════════════════╗
  ║          📚 Goethe - German Sentence Audio Extractor              ║
  ╚════════════════════════════════════════════════════════════════════╝

  ┌────────────────────────────────────────────────────────────────────┐
  │  🎤 Step 1: Speech Recognition                                    │
  └────────────────────────────────────────────────────────────────────┘
  ℹ️ Found 1 video(s)
     📹 german_lesson.mp4
  ℹ️ ⏩ Skipping: german_lesson.mp4 (transcript exists)

  ┌────────────────────────────────────────────────────────────────────┐
  │  🔎 Step 2: Sentence Matching                                     │
  └────────────────────────────────────────────────────────────────────┘
  ℹ️ Searching for 9 sentences
  ✅ Found 6 matches!

  ┌──────────────────────────────────────────────────────────────────┐
  │  📝 Matches Found                                              │
  ├──────────────────────────────────────────────────────────────────┤
  │  🌟  1. Ich habe keine Zeit                                    │
  │     ⏱️  2:15 - 2:18  •  Score: 92%                            │
  │  🌟  2. Wie geht es dir                                        │
  │     ⏱️  5:30 - 5:33  •  Score: 88%                            │
  │  🌟  3. Das ist kein Problem                                   │
  │     ⏱️  8:45 - 8:48  •  Score: 85%                            │
  └──────────────────────────────────────────────────────────────────┘

  ┌────────────────────────────────────────────────────────────────────┐
  │  🎵 Step 3: Audio Extraction                                     │
  └────────────────────────────────────────────────────────────────────┘
  ✅ Created 6 MP3 files!
  📁 Saved to: output/mp3/

  ┌──────────────────────────────────────────────────────────────────┐
  │  📊 Summary                                                    │
  ├──────────────────────────────────────────────────────────────────┤
  │  🎬 Videos:          1                                        │
  │  📝 Transcripts:      1                                        │
  │  🔎 Matches:          6                                        │
  │  🎵 MP3 Files:        6                                        │
  │  ⏱️  Total time:      5.2 seconds                             │
  └──────────────────────────────────────────────────────────────────┘

❓ Frequently Asked Questions

How long does it take?

Video Length	tiny model	small model
1 minute	~30 seconds	~1 minute
5 minutes	~2 minutes	~5 minutes
10 minutes	~4 minutes	~10 minutes

First time: Longer (downloads AI model ~1GB)

Can I use this for other languages?

Yes! Change language: German to:

French, Spanish, Italian, Portuguese
English, Chinese, Japanese, Korean
Any language supported by Whisper

Where do I get German videos?

YouTube
Deutsche Welle
Easy German
Your own recordings
German movies/TV

My laptop is getting hot!

Normal! Speech recognition uses a lot of CPU. To reduce heat:

Use model_size: tiny (fastest, coolest)
Process shorter videos (2-3 minutes)
Take breaks between processing
Put laptop on a hard surface for airflow

No matches found?

Try these fixes:

Lower fuzzy_threshold to 50 in config.yaml
Use shorter sentences
Check if sentences actually appear in your video
Use simpler sentences first

Can I process multiple videos?

Yes! Put all videos in the videos/ folder. Goethe will process them all.

🛠️ Troubleshooting

FFmpeg not found

sudo apt update
sudo apt install ffmpeg

Python packages missing

source venv/bin/activate
pip install -r requirements.txt

Can't see videos folder

# Run setup to create folders
python3 setup.py

Hugging Face token warning

Warning: You are sending unauthenticated requests to the HF Hub

This is normal. To speed up downloads:

Create free account at: https://huggingface.co/join
Get token: https://huggingface.co/settings/tokens
Set token: export HF_TOKEN="hf_xxxxxxxxxx"
Or just wait - it works without token (slower)

Memory error

# Use tiny model
# Edit config.yaml and set:
# model_size: tiny

Error: No such file or directory

# Make sure you're in the right folder
cd ~/Goethe

# Check folder structure
ls -la

💡 Pro Tips

Tip	Benefit
Use short videos (2-5 min)	Faster processing
Use tiny model first	Quick testing
Clear audio = better results	Higher accuracy
Add simple sentences first	Higher match rate
Lower threshold to 50	More matches
Process one video at a time	Less memory usage
Close other apps	More CPU for processing

🎯 Complete Example

Here's a full example from start to finish:

# 1. Download a German video (example using yt-dlp)
yt-dlp "https://www.youtube.com/watch?v=example" -o videos/german_lesson.mp4

# 2. Create sentences.txt
cat > sentences.txt << EOF
Ich habe keine Zeit.
Wie geht es dir?
Das ist kein Problem.
Guten Morgen!
Hallo, wie heißt du?
Ich liebe Deutsch.
Wo ist der Bahnhof?
Ich verstehe nicht.
Kannst du mir helfen?
Was ist das?
EOF

# 3. Activate environment
source venv/bin/activate

# 4. Run Goethe
python3 main.py

# 5. Find your MP3s
ls output/mp3/

# 6. Listen to them!
ffplay output/mp3/*/*.mp3

🤝 Contributing

Want to help improve Goethe?

Fork the repository
Create a feature branch
Make your changes
Submit a pull request

Areas that need help:

Better documentation
More language support
GUI interface
Batch processing improvements
Performance optimizations

📄 License

MIT License — Free to use, modify, and share!

🙏 Acknowledgements

faster-whisper — Speech recognition
FFmpeg — Audio extraction
Hugging Face — Model hosting

📬 Contact

GitHub: @duman-dp
Project: Goethe
Issues: Report a bug

⭐ Support

If Goethe helped you learn German, please star ⭐ this repository!

Happy Learning! 🇩🇪🎧

Name		Name	Last commit message	Last commit date
Latest commit History 7 Commits
modules		modules
.gitignore		.gitignore
README.md		README.md
config.yaml		config.yaml
main.py		main.py
requirements.txt		requirements.txt
sentences.txt		sentences.txt
setup.py		setup.py

Folders and files

Latest commit

History

Repository files navigation