Computational analysis of the prosodic characteristics of the poets’ voice

This thesis presents the development of a novel computational tool, Pause, designed to explore, and analyse the intricate prosody of poetry. Through the use of different state of the art techniques, the tool aims to unveil the rhythmic and melodic patterns inherent in poetic expression, allowing for the study of their role in conveying the emotion and aesthetics of poetry. The project begins with a comprehensive review of existing literature on prosody, computational linguistics, and their intersection in the realm of literary analysis. Emphasizing the significance of the voice in understanding the nuances of poetic form, this research sets out to bridge the gap between traditional literary analysis and cutting-edge computational methodologies. Pause utilizes tools that have been used extensively in previous research in the field to dissect the acoustic features of audio recordings, capturing the intensity, pitch, and speed that contribute to the unique prosodic fingerprint of each poem. They include algorithmic solutions as well as machine learning models that enable us to segment and analyse spoken language. These tools have proven their reliability and accuracy on the aforementioned research. The implementation of Pause is accompanied by a user-friendly interface, facilitating accessibility for scholars and enthusiasts interested in exploring the prosodic intricacies of poetry. The tool allows users to input poetic texts alongside recordings of those poems, visualize prosodic patterns, and generate insightful analyses. The results are different annotations of the provided text that can be used for a critical study of poetry. These annotations are accompanied by the relevant data to explain and contextualize them. Through a series of use case, this thesis also provides the user with an insightful guide on the use of the tool, as well as providing examples of its potential use in education and outreach. In conclusion, Pause represents a significant contribution to the field of computational linguistics applied to literature. This tool facilitates the work of researchers approaching a critical study of the prosodic components of poetry, allowing for a close reading of the material, and yielding measurable and replicable insights.