Author:
Maddali Anil Kumar,Khan Habibulla
Abstract
Purpose
Currently, the design, technological features of voices, and their analysis of various applications are being simulated with the requirement to communicate at a greater distance or more discreetly. The purpose of this study is to explore how voices and their analyses are used in modern literature to generate a variety of solutions, of which only a few successful models exist.
Design/methodology
The mel-frequency cepstral coefficient (MFCC), average magnitude difference function, cepstrum analysis and other voice characteristics are effectively modeled and implemented using mathematical modeling with variable weights parametric for each algorithm, which can be used with or without noises. Improvising the design characteristics and their weights with different supervised algorithms that regulate the design model simulation.
Findings
Different data models have been influenced by the parametric range and solution analysis in different space parameters, such as frequency or time model, with features such as without, with and after noise reduction. The frequency response of the current design can be analyzed through the Windowing techniques.
Original value
A new model and its implementation scenario with pervasive computational algorithms’ (PCA) (such as the hybrid PCA with AdaBoost (HPCA), PCA with bag of features and improved PCA with bag of features) relating the different features such as MFCC, power spectrum, pitch, Window techniques, etc. are calculated using the HPCA. The features are accumulated on the matrix formulations and govern the design feature comparison and its feature classification for improved performance parameters, as mentioned in the results.
Subject
General Computer Science,Theoretical Computer Science
Reference28 articles.
1. Gender recognition system using speech signal;International Journal of Computer Science, Engineering and Information Technology (IJCSEIT),2012
2. Automatic recognition of gender by voice” in proc,1988
3. Recurrent neural network language model adaptation for multi-genre broadcast speech recognition and alignment;IEEE/ACM Transactions on Audio, Speech, and Language Processing,2019
4. Kernel method for voice activity detection in the presence of transients;IEEE/ACM Transactions on Audio, Speech, and Language Processing,2016
5. Characterization of dysphonic voices utilizing a filter bank-based spectral analysis: sustained vowels and running speech;Journal of Voice,2013
Cited by
1 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献
1. Computer Vision and Speech Understanding;Proceedings of the 2nd International Conference on Cognitive and Intelligent Computing;2023