Voice Clone TTS

Overview

Implemented a real-time voice cloning system based on SV2TTS architecture.

Responsibilities

• Integrated open-source voice cloning model
• Processed speaker embeddings
• Generated natural-sounding speech

Approach

• Speaker encoder
• Synthesizer (Tacotron-like)
• Vocoder

Pipeline

flowchart LR
    A[Speaker Audio] --> B[Speaker Encoder]
    B --> C[Embedding]
    C --> D[Synthesizer]
    D --> E[Vocoder]
    E --> F[Generated Speech]

Tech

PyTorch · TTS · SV2TTS

Impact

• Enabled realistic voice cloning
• Demonstrated multi-stage deep learning pipeline
• Expanded into generative audio systems