<p dir="ltr">Suprasegmental features (i.e., features realized through pitch, volume, and/or duration changes) are particularly difficult for second language learners. This has led to the creation of many different methods of training, often employing the use of experimenter or instructor-created training materials. However, the use of authentic materials for training suprasegmental features remains underexplored. Authentic materials have recently been shown to be useful for some aspects of language learning, such as communicative competence and listening comprehension, while their usefulness for training suprasegmental features remains unclear.</p><p dir="ltr">This dissertation explores the potential usefulness of authentic materials for training two suprasegmental features, second language (L2) lexical tone in Mandarin and L2 lexical pitch accent in Japanese. For each language, participants were divided into an authentic group and an inauthentic group based on the type of materials they were exposed to during a shadowing task. Participants then used words they shadowed in a visual feedback task, which was chosen due to visual feedback’s previously proven usefulness for training these two features.</p><p dir="ltr">For each language, L2 productions at the pretest, posttest, and delayed posttest were compared to first language (L1) productions using a previously established analysis known as the <i>CID </i>measure analysis. The resulting time-series plots, <i>CID </i>measures, magnitudes (pitch height), and phases (timing), were analyzed to determine which group, authentic or inauthentic, performed in a more L1-like way at the posttest compared to the pretest. Results suggest that the authentic group outperformed the inauthentic group for three out of four Mandarin tones, and three out of seven Japanese pitch accent patterns. These results begin to show support for the usefulness of authentic materials, with experimental and pedagogical implications for researchers and learners of Mandarin and Japanese.</p>