OBJECTIVE Computed radiography of the musculoskeletal system has the potential to become a powerful tool in the practice of diagnostic radiology. It addresses many of the geographic and film-distribution concerns facing diagnostic imaging. We undertook this study to compare and document the quality of computed radiographs and conventional screen-film images before widespread implementation. METHODS We evaluated clinical images using direct comparison. Bilateral hand images from 50 patients were scored independently by six musculoskeletal radiologists. In each case one hand was imaged with a conventional screen-film technique and the other with computed radiography. Images were masked to eliminate as much bias as possible. The numeric scores assigned to the images by the observers were analyzed using Student's t test. RESULTS Computed radiographs were judged with statistical significance to be better than conventional screen-film images in all features judged by the observers, including bone cortex, bone trabeculae, corticomedullary junction, distal phalangeal tuft, soft tissues, fat planes, bone-soft-tissue interface, and overall contrast and density. CONCLUSIONS The statistically significant determination that the image quality of computed radiographs is at least as good as screen-film images allows confident use of computed radiography and enables radiologists to take advantage of its many other practical capabilities related to image distribution, storage, cost, and geographic coverage without sacrificing image quality.