In the past years some studies had been directed towards the audibility thresholds of phase distortion and in general, electronically generated signals have been used. It has been proven that phase distortion is audible but little attention has been paid to its influence on sound quality. This thesis investigates the influence of phase distortion in the perception of sound quality. The work is focused on the perception of phase distortion in a normal listening environment with loudspeakers. Two subjective listening test were implemented. For the first one, an adaptive algorithm was designed in order to find phase distortion audibility thresholds for a set of sixteen stimuli, among which were speech, instrumental music, electronic generated sounds and pop music. The purpose was to select those signals which showed to be more sensitive to phase distortion. The selected signals were used in the second listening test. This test consisted of a paired comparison experiment and its purpose was to assess the listener preferences regarding the phase distortion. A training session was implemented in order to teach the differences to the listener. This session was ran before the preference parts. Results showed that phase distortion is not audible for most of the evaluated stimuli. Stimuli such as speech and electronic generated signals showed lower thresholds when played through headphones. Training showed to be a key factor on the phase distortion perception. From the preference test it was found that for speech some subtle distortion is rated similar to no distortion but when perceivable, distortion is not accepted at all. In the case of the electronic generated signal a monotonically decreasing scale was obtained, where the non-distorted sample was ranked highest and the most distorted lowest. An electronic music sample from a commercial record was evaluated as well. It was observed that distortion had a subtle effect through headphones, but there is no significant difference in the ranking. In general, phase distortion showed to be not audible through loudspeakers in a normal listening environment for any of the selected signals. No significant differences were found in the results and subjects showed difficulties trying to discriminate the distorted from the original.