ECG-based authentication is an emerging biometric modality that provides strong resistance to spoofing attacks. This study evaluates the effectiveness of representation learning for generating discriminative ECG embeddings by comparing the TS-TCC and CLOCS architectures, and identifies optimal pipeline configurations for real-world ECG-based authentication. Using the LarField dataset, comprising over 20,000 h of single-lead ECG recordings from 48 subjects, we assessed the complete authentication pipeline. To ensure subject-level independence, no subjects used for training were included in the evaluation or experimental sets. Our results demonstrate that contrastive learning effectively mitigates physiological variability; specifically, CLOCS achieves the highest recall (0.99) and F1 Score (0.94), while TS-TCC matches it only in precision. Furthermore, an analysis of reference-set construction indicates that approximately 100 readings collected over a 10-day span are sufficient to achieve consistent performance for both models. These findings provide a comprehensive benchmark for state-of-the-art methods and offer practical guidelines for the design of physiological biometric authentication systems.