Multi-Speaker Multi-Style Speech Synthesis with Timbre and Style Disentanglement