| Author: | Lui, Kwai Hong |
| Title: | A study of the association between the parameters & sound quality of the designs of headphone core |
| Advisors: | Lee, K. M. Carman (ISE) |
| Degree: | Eng.D. |
| Year: | 2025 |
| Department: | Department of Industrial and Systems Engineering |
| Pages: | 299 pages : color illustrations |
| Language: | English |
| Abstract: | The study delved into the complex arena of earphone design, leveraging principal component analysis to develop a neural network model that characterises the design process and predicts the six key sound quality indicators: total harmonic distortion, output power, frequency response, signal-to-noise ratio, speaker impedance, and headroom. The design of earphones is recognised as a challenging task that traditionally relies heavily on extensive experience and technical expertise. The research aimed to construct a predictive model to streamline the design process, making it more accessible and efficient. Along with the development of previous research, a simplified model for earphone design was crafted. This model identifies a relationship between sound quality indicators and eight material-specific parameters, namely, the type of driver 1, the magnet of driver 1, the voice coils of driver 1, the diaphragm of driver 1, type of driver 2, the magnet of driver 2, the voice coils of driver 2, and the diaphragm of driver 2. The neural network approach was applied to seek the connection between the independent parameters and the sound quality, resulting in a more manageable and streamlined model. The neural network-based approach yielded a model with optimal predictive accuracy for the sound quality indicators in comparison with the linear counterparts, including linear regression mode and principal component analysis. Additionally, the model with 15 neurons in the hidden layer was found to be the best performer in predictive power. In the study's investigation into earphone design methodologies, various predictive methods were assessed, with their characteristics summarised as follows: analysis of variance offered insights into design trends, suggesting that higher quality materials tend to result in better sound quality. Regression Analysis provided a linear model linking design parameters directly to sound quality outcomes. Principal component analysis streamlined the design process by reducing the input dimensions from eight to four, thereby simplifying the evaluation process of the earphone design. Lastly, the Neural Network approach addressed the design challenge with a nonlinear model and recommended a configuration of 15 neurons in the hidden layer according to the test data analysis. The research developed a simplified predictive model, applied neural network modelling to a real problem, compared the performance of the developed predictive methods, and inspected the material quality and sound performance (0.5186 to 1.0622, the Score of 7). |
| Rights: | All rights reserved |
| Access: | restricted access |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| 8590.pdf | For All Users (off-campus access for PolyU Staff & Students only) | 6.62 MB | Adobe PDF | View/Open |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/14364