Computer-Assisted Pronunciation Training:: The Four K's of Freedback

One of the biggest problem areas in relation to Computer Assisted Pronunciation Training (CAPT) applications, designed for Second Language Learning (SLL), has been that of feedback. As with human teachers, the ultimate goal of CAPT is to provide instantaneous, individualized feedback to the user on the overall quality of the pronunciation made. Attempts at accomplishing this goal have seen the light of day in many different ways. Some applications provide intonation curves and some provide spectrograms, color-coding or numerical scores. Unfortunately these feedback modes rarely inform the user of what is wrong instead relying on the users to self-correct through trial and error [2].