This study used a modified Bramante technique and new digital subtraction software to compare root canals prepared by nickel-titanium (Ni-Ti) hand, Ni-Ti engine-driven, and stainless steel hand endodontic instruments. Sixty mesial canals of extracted human mandibular molars were randomly divided into five groups. The roots were embedded in clear resin and cross-sectioned in the apical and mid-root areas. In group A, canals were instrumented using a quarter turn/pull technique with K-Flex files. In group B, canals were prepared with Ni-Ti hand files (Mity files) using the same technique as in group A. Group C was prepared with NT Sensor engine-driven files. Group D canals were prepared with Ni-Ti Canal Master "U" hand instruments. Group E was prepared with engine-driven Ni-Ti Lightspeed instruments. Digitized images of the uninstrumented canals were compared with images of the instrumented canals. Engine-driven Ni-Ti instruments (Lightspeed and NT Sensor file) and hand instrumentation with the Canal Master "U" caused significantly less canal transportation (p < 0.05), remained more centered in the canal (p < 0.05), removed less dentin (p < 0.05), and produced rounder canal preparations than K-Flex and Mity files. Engine instrumentation with Lightspeed and NT Sensor file was significantly faster than hand instrumentation (p < 0.05).