A Comparative Analysis of Poetry Reading Audio: Singing, Narrating, or Somewhere In Between?

Kahyun Choi,Minje Kim

DOI: https://doi.org/10.1109/ICASSP48485.2024.10447582

2024-04-01

Abstract:This paper provides a computational analysis of poetry reading audio signals at a large scale to unveil the musicality within professionally-read poems. Although the acoustic characteristics of other types of spoken language have been extensively studied, most of the literature is limited to narrative speech or singing voice, discussing how different they are from each other. In this work, we develop signal processing methods, which are tailored to capture the unique acoustic characteristics of poetry reading based on their silence patterns, temporal variations of local pitch, and beat stability. Our large-scale statistical analyses on three big corpora, each of which consists of narration (LibriSpeech), singing voice (Intonation), and poetry reading (from The Poetry Foundation), discover that poetry reading does share some musical characteristics with singing voice, although it may also resemble narrative speech.

Sound,Audio and Speech Processing

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to reveal the musical features in professional poetry - reading - aloud audio through large - scale computational analysis. Specifically, the research aims to fill the gap in the existing literature regarding the insufficient analysis of the acoustic properties of poetry recitation and explore the unique position of poetry recitation between Narration and Singing.

### Specific Description of the Problem

1. **Unique Acoustic Properties of Poetry Recitation**:
- Although there has been a great deal of research on the acoustic properties of other types of spoken language (such as narration and singing), most of these studies are limited to narrative voices and singing voices, discussing the differences between them.
- As a unique form of expression, the acoustic properties of poetry recitation have not been fully studied. Therefore, this paper hopes to capture the unique acoustic features in poetry recitation, including silence patterns, local pitch changes, and beat stability, by developing specific signal - processing methods.

2. **Musicality of Poetry Recitation**:
- Poetry recitation has a certain musicality, which is reflected in pitch changes and reading rhythms.
- The research attempts to quantify the musical elements in poetry recitation in order to understand its relationship with narration and singing.

3. **Cross - Domain Comparison**:
- Through statistical analysis of large - scale data sets (including narration, singing, and poetry recitation), the research attempts to find out whether poetry recitation shares some musical features of singing and may also be similar to narration.

### Main Objectives

- **Develop Signal - Processing Algorithms**: In order to analyze the acoustic properties of poetry recitation on a large scale, researchers have developed specialized signal - processing methods that can capture silence patterns, local pitch changes, and beat stability in poetry recitation.
- **Large - Scale Data Analysis**: Using three large - scale public data sets (LibriSpeech, Intonation, and The Poetry Foundation), quantitative analysis of different types of spoken language performance is carried out.
- **Fill the Academic Gap**: This is the first large - scale quantitative analysis of poetry recitation, filling the gap in the existing literature and providing a basis for more in - depth research in the future.

### Conclusion

Through the above methods, the research reveals the uniqueness of poetry recitation in acoustic properties, especially in the intermediate zone between musicality and narration and singing. This research not only enhances our understanding of poetry recitation but also provides new perspectives and tools for further exploring the art and technology of poetry recitation.