Codec-ASR: Training Performant Automatic Speech Recognition Systems with Discrete Speech Representations | Synapse