High-Fidelity Simultaneous Speech-to-Speech Translation