Analyzing impact of data augmentation on vocal effort classification with self-supervised models | Synapse