ABSTRACT
Digital technology makes many acoustic components of speech visible: displays of fundamental frequency (i.e. the acoustic correlate of pitch) facilitate the acquisition of lexical tone and the patterns of intonation phrases, waveforms demonstrate segmental duration, and spectrograms display formant characteristics of sounds. Recent advances have made this information available to language researchers, teachers and learners. This chapter outlines some of the issues in the field, contributions made by research on the topic, methodological considerations and recommendations for practice. At the forefront of the discussion are topics such as the interpretability of the information that programs provide to users, role of variability in experimental design, demonstration of improvement through training, and perceptions of the efficacy of training by participants.
Among the displays provided by signal analysis software, those for pitch are probably the most user-friendly and informative. They can emphasize the similarities and differences between a learner’s pre- and post-training utterance, or a learner’s and a native speaker’s utterance. Other types of displays, such as waveforms and spectrograms, can also be informative for users with some guidance. In addition, research findings (e.g. Okuno & Hardison, in press) indicate that these displays can facilitate perception and production improvement. Learners have reported positive impressions regarding the use of such tools and often recommend their inclusion in the language curriculum. Technology offers several advantages in second language speech learning, including the accessibility of public domain tools, convenience of web-based delivery formats, and applicability to a range of language features and target languages. For learners without target language input or guidance, corrective feedback from automatic speech recognition systems holds some promise for the future.
