Lin, Xiaoling (2024) Identifying ASMR-Style Audio: Development of a Predictive Classification Model. Master thesis, Voice Technology (VT).
|
PDF
MSc-5476399-X-Lin.pdf Download (2MB) | Preview |
Abstract
Over the past few years, ASMR (Autonomous Sensory Meridian Response) videos have quickly become a popular genre with a significant emphasis on the auditory aspect in eliciting specific sensory reactions in the audience. However, "ASMR-style audio" lacks a clear definition when compared to common audio. This study aims to fill this gap by creating a model that can predictively differentiate ASMR-style audio from other audio. The CNN model designed in this study aims to accurately distinguish ASMR-style audio from common audio. The performance of the model is evaluated using accuracy in identifying ASMR-style audio. This study's findings suggest that synthesizing ASMR-style audio in the future could become possible, allowing individuals to select their preferred ASMR content. By automating the classification of ASMR-style audio, this research not only enhances content curation on streaming platforms but also contributes to the broader field of audio classification and voice technology.
| Item Type: | Thesis (Master) |
|---|---|
| Name supervisor: | Schauble, J.K. |
| Date Deposited: | 23 Aug 2024 07:16 |
| Last Modified: | 23 Aug 2024 07:16 |
| URI: | https://campus-fryslan.studenttheses.ub.rug.nl/id/eprint/554 |
Actions (login required)
 |
View Item |