Acoustic analysis techniques were used to investigate the short-term consequences of vocally violent behavior, and to compare voice production before and after training in hygienic laryngeal release (HLR) techniques. Twenty-seven actors ranging in age from 17 to 48 years were audiorecorded before and after multiple productions of 4 vocally violent behaviors: grunting, groaning, sobbing, and shouting. After training in HLR techniques, the experimental protocol was repeated. Audiorecordings of vowels (produced at 3 pitch levels: modal F0, minimum F0, maximum F0) before and after vocal violence, and before and after HLR training, were analyzed using the Multidimensional Voice Program (4305, Kay Elemetrics Corp, Lincoln Park, NJ). After vocal violence, no consistent acoustic changes were detected for voice generated at modal and minimum F0; however, significant increases in both fundamental frequency range and maximum F0 were observed. After training in HLR techniques, acoustic measures sensitive to pitch and amplitude perturbation, and non-harmonic noise, improved across pitch levels. The results also indicated that vocal training does defend the laryngeal system from undesirable changes related to vocally violent maneuvers that might surface at the extremes of an actor's pitch range. Because the HLR technique used in this investigation was multimodal, interesting questions are raised regarding which aspect of training is primarily responsible for the observed effects. Further study is required to identify such factors.