Text this: Comparative modeling of mixed cardiopulmonary sounds in a low-resource paired dataset