International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 04 Issue: 07 | July -2017
p-ISSN: 2395-0072
www.irjet.net
Investigations of Formant Extraction of Male and Female Speech Signal Via Cepstral De-convolution Method Sakshi Bedi1, Randhir Singh2, Saleem Khan3 1, 2 Dept.
of Electronics and Communication Engineering, SSCET, Badhani, Punjab, India 3Dept. of Electronics, GGM Science College, J&K, India ---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - The capability of speaking the language is one of the most amazing skills human possess. It serves as a very effective way of communication, sharing experiences, thoughts, ideas among the people. In this research work, cepstral analysis of sentence level speech signals of male and female speakers is carried out. Spectrogram analysis of segmented speech is carried to determine various speech parameters. The Cepstral analysis shows the unique trait between male and female speakers in terms of low time liftered speech processing. Key Words: The speech signal, sentence analysis, Hamming, Cepstral, Spectrogram.
1.INTRODUCTION Speech is the vocalized form of Human communication. The creation of each spoken word is in accordance with the phonetic combination of a limited set of speech sound units such as vowel and consonants. The process of generation of sound is based on three main factors for effective voice generation:
1.1 Breathing Our purpose of producing voice is signaled to the parts of the body by taking the involvement of impulses from the brain [2]. The body shows its first response by exhalation so that to power voice there is enough air in the lungs. The exhalation from the mouth and nose passes down the wind pipe and is inhaled into the lungs. The air which we breathe in through the lungs the ribcage needs to be expanded and the diaphragm which forms the base of the chest needs to be smooth downwards. At the time we breathe in we feel most of the expansion in the lower ribs area. Once the air we inhaled down into the lungs reaches its capacity, the elastic tissue of the lungs recoils and the air we inhaled is breathed out [3].
1.2 Phonation The production of electric supply for voice is done by breathing air out of the lungs. This process of exhalation of airflow from the lungs makes the vocal chords in the voice box vibrate to make the basic sound of the voice. This is referred as Phonation. The basic tone of sound can be Š 2017, IRJET
|
Impact Factor value: 5.181
|
differentiated in many ways, depending on the way in which we use the other parts of the voice mechanism and the vocal folds. The basic facets of voice that can be varied are: Pitch: Pitch refers to as the highness and lowness of the voiced sounds. It is determined by the length of the folds, the thickness of the edge of the folds and the speed of vibration of the vocal folds. The highness level of the pitch depends on the elongation and thinning the edges of the vocal folds. The less the elongation and thinning of the edges of the vocal folds the vocal folds vibrate at the slower rate. Pitch describes the main indication of gender. The average pitch of the female voice is about 200bHz where as the average pitch of the male voice is about 110 Hz. Hence pitch is the main indicator of gender. Emotion signals such as excitement, stresses that are present in voice also describes as a part carried by pitch variations. Shouting as a means of stressing on a particular point or an expression of anger describes more with rising of the pitch than loudness. More often pitch variation is often correlated with loudness variation. Happiness, distress and extreme fear in voice are signals led by the fluctuation of the pitch in voice. The accent is the part conveyed by changes in the pitch and rhythm of the speech signal. For example in more accents, such as Northern Ireland accent, at end of the sentence, the pitch signal is raised instead of being lowered [4-5]. Loudness: Loudness referred to as the how loud or soft a voice. The amount of air pressure from the lungs and muscle tension of the vocal folds decides the loudness of the voice. In the nutshell, the loudness of the sound depends on the air pressure and tension of the vocal folds. The greater the air pressure, the tenser the vocal folds so the louder the voice will be and lesser the air pressure, the slacker the vocal folds so softer the voice will be. Quality: Quality referred to as the clarity of the voice sounds. The determination of voice quality depends on many complex factors including the relaxation of muscles of the larynx and how moist the cover of vocal folds is and how smoothly the vocal folds vibrate. The dryness of the cover of vocal folds depends on the muscles of the larynx. If the muscles of the larynx are excessively tense, the cover is dry and the folds cannot close together or move in an irregular
ISO 9001:2008 Certified Journal
|
Page 2951