The aim of the present study was to compare assessments of sperm concentration and sperm motility analysed by conventional semen analysis with those obtained by computer-assisted semen analysis (CASA) (Copenhagen Rigshospitalet Image House Sperm Motility Analysis System (CRISMAS) 4.6 software) using semen samples from 166 young Danish men. The CRISMAS software identifies sperm concentration and classifies spermatozoa into three motility categories. To enable comparison of the two methods, the four motility stages obtained by conventional semen analysis were, based on their velocity classifications, divided into three stages, comparable to the three CRISMAS motility categories: rapidly progressive (A), slowly progressive (B) and non-progressive (C+D). Differences between the two methods were large for all investigated parameters (P < 0.001). CRISMAS overestimated sperm concentration and the proportion of rapidly progressive spermatozoa and, consequently, underestimated the percentages of slowly progressive and non-progressive spermatozoa, compared to the conventional method. To investigate whether results drifted according to time of semen analysis, results were pooled into quarters according to date of semen analysis. CRISMAS motility results appeared more stable over time compared to the conventional analysis; however, neither method showed any trends. Apparently, CRISMAS CASA results and results from the conventional method were not comparable with respect to sperm concentration and motility analysis. This needs to be accounted for in clinics using this software and in studies of determinants of these semen characteristics.