ClariFi - Speech Enhancer

ClariFi uses deep learning to enhance noisy speech, restoring clarity and quality in real-time.

Upload Audio

🔍 Key Features

✅ Trained DNN model using real noisy-clean speech pairs
✅ Real-time enhancement using spectral mapping
✅ Evaluated with PESQ & STOI metrics
✅ Simple and fast browser-based upload interface

📊 Model Performance Metrics

PESQ: 3.2

SDR: 8.7 dB

Latency: 50 ms