English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Sound is an effective and efficient magnitude for biometric characterization. However, the sound is a phenomenon that is a fusion of multidimensional and influenced many aspects, such as speaker characteristics (articulator configuration, emotions, health, age, sex, dialect), languages, and the environment (background and transmission media), so that the system has been developed until now has not been able to work well in real situations. This is the background of this research. In this research, we investigate higher order statistics (HOS) and Mel-Frequency Cepstrum Coefficients (MFCC) as a feature extraction, and integrated with a Hidden Markov Model (HMM) as a classifier to get a more robust speaker identification system, especially for Gaussian Noise. Research carried out more focused on feature extraction part of the speaker identification system. In classifier process stage, we use the HMM. This is a technique that has been widely used in voice processing provides good results. At the beginning, we empirically showed the failure of conventional MFCC using power spectrum in noisy environment. Then proceed with reviewing the matter, and proposed HOS-based extraction techniques to overcome these problems. Next is an experiments to demonstrate the effectiveness of the proposed method. Data used in this study came from 10 people who say the phrase 'PUDHESA' as much as 80 times with different ways of utterance. In this research, we use signals that are spoken with different variations of pressure, duration, emotional, loud and weak. Figure 1 presents the forms of signals for different utterances of a speaker. In accordance with the focus of this research is to build models that are more robust to noise, then we add a Gaussian noise signal to each original signal with a signal-tonoise ratio (SNR) of 20 dB, 10 dB, 5 dB and 0 dB.

intech ebooks

Mel-Frequency Cepstrum Coeffficients as Higher Order Statistics Representation to Characterize Speech Signal for Speaker Identification System in Noisy Environment Using Hidden Markov Model