In this paper, a heart sound analyzer is presented for interpretation of heart sound signals and automated diagnosis of valvular heart disease. The heart sound analyzer includes data acquisition from multiple positions, signal analysis to extract auscultatory features and information, and a knowledge-based program to provide a likely diagnosis. Experiments using clinical data from real patients show good performance for the automatic analysis. Due to its simplicity and fast implementation, the method has potential for clinical deployment and eventually to be used in distributed diagnosis settings