I have been grateful to Ms. Shahini Haque, Assistant Professor at the Department of ETE Daffodil International University, Dhaka, and I want to express my sincere thanks to her. I would like to express our sincere gratitude to Ms. Shahina Haque, Assistant Professor, Department of ETE, for her kind help in completing my thesis and also to other faculty members and staff of the Department of ETE, Daffodil International University.
CHAPTER-01 INTRODUCTION
- WHY CHOOSE THIS TOPICS ?
- MOTIVATION
- OBJECTIVE
- General Objective: The objective of this thesis is to analyzing disordered speech of hypernasalityin cleft palate or lip children
- Specific Objective: The specific objectives of the thesis work are-
- ORGANIZATION OF THE THESIS
In this work I use the wave keyboard to get the exact sound of children's cleft palate vowels from their voice. In this thesis I will analyze the abnormal hypernasal speech voice of children with cleft palate.
CHAPTER- 02 LITERATURE REVIEW
- INTRODUCTION
- PERCEPTUAL EVALUATION
- OBJECTIVE EVALUATION
- ACOUSTIC MEASURE OF SPEECH
- NASALANCE MEASURE OF SPEECH
- ONE THIRD OCTAVE SPECTRAL ANALYSIS AND VLHR
- PERTURBATION MEASURES(JITTER AND SHIMMER)
- PREVIOUS WORK ON SPEECH AND HYPERNASALITY ON 90’S
- PREVIOUS WORK ON SPEECH AND HYPERNASALITY FROM(2000-2018)
The perceptual assessment was considered the gold standard[7][8] for the analysis of speech among people with CLP. In the initial stages, the perceptual evaluation was limited to the examination of articulation skills, which mainly focused on the description of articulation errors, the frequency of these errors and the type of errors. The comparisons of these errors were usually performed with non-cleft children using traditional SODA (substitution, omission, distortion, addition) analysis[9]. Later during the 1950-60s, researchers identified many other speech parameters that were affected in the individual with CLP. The parameters such as the presence of compensatory articulation, scoring of articulatory errors using different types of Rting scales, etc. were explored. It was during this decade that the focus was shifted to parameters such as resonance, nasal grimaces, etc. [10]. This gave rise to the development of different perceptual assessment protocols. Some of the protocols have been reviewed and used in various research studies[11]reat Omond street.
Henningsson (2007) studied several parameters such as hypernasality, hyponasality, nasal air emission and consonant production errors. These parameters were used to report speech outcomes in individuals with cleft palate to achieve greater consistency in the reporting of speech outcomes worldwide, regardless of language or language spoken. A working group of six individuals with experience in speech and cleft palate was formed to develop a system of universal parameters for reporting speech outcomes in individuals born with cleft palate. The system was adopted in conjunction with a workshop in Washington, D.C. dedicated to developing the universal system. The system, which was further refined after the workshop, includes a three-part plan consisting of 1) evaluation, 2) mapping and 3) reporting. This report focuses primarily on the third phase, reporting speech. However, they felt that perceptual evaluation remains the gold standard for evaluating speech, as well as the most commonly used method. The perceptual evaluation is influenced by several factors, such as the type of stimuli, phonetic context, voice quality, articulation pattern, previous experiences and listener expectations[14]. The perception of hypernasality varies as a function of other aspects of speech. The perceptual assessments can complement the objective assessments which, as a conjoint assessment procedure, can provide detailed insight into the speech skills of an individual with CLP.
Review of various studies that have used perceptual assessment protocols has revealed that these measures are influenced by multiple factors such as stimuli, judges, rating scales, listening conditions, recording quality, and articulatory characteristics of speech in children with CLP. The reviewed studies have shown that acoustic measures of speech have been widely studied by different researchers and many studies have used acoustic measures to assess the speech of people with CLP. Nose mass is widely used to assess nasality. Among the indirect assessment procedures, the concept of nose measures was extensively researched during the 1970s. The concept was largely based on earlier work developed by Fletcher.
The spectrographic analysis has been used to investigate the spectral and temporal parameters of the speech of individuals with CLP [48]. Specifically, it was noted that the opening phase of nasal vowels was shorter due to altered vocal cord vibration, as reported by Hamlet (1973). The authors attributed the results to the increased force during vocal fold adduction without vocal effort in the presence of nasalization. 1985, two.
CHAPTER-3
BASIC ACOUSTICS OF SPEECH SIGNAL
- THE SPEECH SIGNAL
- SPEECH PRODUCTION
- IMPORTANT OF HUMAN ORGANS OF SPEECH
- BRIEF DESCRIPTION OF HYPERNASALITY
- DIFFERENT TYPES OF CLEFT PALATE
- Submucosal cleft:Asubmucosal cleft could be a cleft that's underneath the membrane that lines the roof of the mouth — thence the term “sub.” as a result of a submucosal cleft
- Soft palate cleft:A cleft of the tongue runs from the tip of the flap and stops before or at the junction of the soft and surface. Not solely is it additionalobvious in its look than a
- Soft and hard palate cleft: A cleft that involves each the arduous and taste bud can embody the complete taste bud and any a part of the surface up to the
- SPEECH IMPAIRMENT OF CLEFT PALATE CHILDREN
- ACOUSTIC FEATURES OF SPEECH
- ACOUSTIC PHONETICS OF VOWEL FORMANTS
Phonemes will be composed in a light weight of the properties of either the time waveform or repetitive properties and ordered into different sounds produced by the human vocal tract. The cross-sectional zone of the vocal tract is limited by the tongue, lips, jaw and velum and varies from 0-20 cm2. In less complicated terms, it involves only the secondary roof of the mouth, because it does not extend all the way forward to include the appendix.
Not only is it more obvious in its appearance as the junction of soft and surface. A superficial|congenital defect|congenital anomaly|congenital defect|congenital disorder|congenital abnormality} in children develops in very vertebrates when the two halves of the sky fail to close and fuse in the middle. These children face social, communication and language problems. A speaking voice is perceived because the F0 (tone) and its harmonics are multiples of the surface unit F0.
F1 has a lot to do with the height of the tongue, and F2 relates to the anterior-posterior position of the tongue. The frequencies of the formants are affected by three factors: the degree of constriction created by the tip of the tongue; the opening of the cartilage structure where this narrowing takes place and therefore the amount of lip rounding, lip protrusion or lip spreading. The resonance frequencies of the vocal tract are called the formants. The frequencies of the three primary formants of the vowel within the words listen, hid, head, had, hod, hawed, hood and who would be shown in Figure 3.4. (b) Comparison with Figure 3.4(a) shows that there is no relationship between actual tongue position and formant frequencies. Nasal vowels, in which the mouth is lowered so that some of the airflow exits through the nose, are found in several languages.
CHAPTER- 04
ANALYSIS TECHNIQUE AND MATHEMATICAL TOOLS USED 4.1 INTRODUCTION
- DISCRETE TIME SIGNAL
- WINDOWING
- WINDOW FUNCTION
- THE FOURIER TRANSFORM
- THE DISCRETE FOURIER TRANSFORM(DFT)
- FAST FOURIER TRANSFORM(FFT)
- SPEECH MATERIAL
- LINEAR PREDICTION CODING
- LPC ANALYSIS TECHNIQUE
- LPC MODELING AND USING LPC ANALYSIS TECHNIQUE
One cluster includes , 1≤B≤1.8 and also the alternative cluster includes B≥1.98 . The Gauss and Emperor windows include families of area units that include each team, although there are only 1 or 2 examples of each area unit shown. Many of the window capabilities don't expect to be unequivocally zero beyond the associate degree interim, as long as the window times the competition results are square. Fourier reprocessing is the main analysis tool for the many problems related to signal and LTI system as a result of Fourier reprocessing of a wave {shape | waveform | wave | waviness}.
Speech Analysis Techniques provides a brief scientific summary of the speech signal analysis techniques involved in a selected specialization in Variable Resolution Spectral Analysis, i.e. Emphasis, Variable Resolution Spectral Analysis, Filter Bank Analysis (Filter Bank Speech Analysis), Linear Prophetic Analysis (Linear Prediction Speech Analysis), Cepstral Analysis, Deltas and social control (Delta, acceleration and normalization). Our goal in processing the speech signal is to get an extra convenient or extra useful illustration of the data carried by the speech signal. Since the parameter used in most speech process applications, the area unit, is derived from the frequency domain illustration, the biggest task is to reason about the sound spectrum.
The experimental half consists of recording every word beginning with a vowel from the sentence. Each of these words ar attached to the English vowel a/e/i/o/u. Registration. Direct expectation models of the human vocal tract as an infinite motion feedback (IIR) frame that makes the speech signal for vowel sounds and different speech localities, which have a robust structure and high level of to the similarity of overtime movements that measure the square of the goods of their quantity height, this demonstration produces a good illustration of the sound. Murmurs and squeaks are created by the activity of the tongue, lips and throat between sibilants and plosives.
CHAPTER-05
ANALYSIS AND RESULT
- FEATURE EXTRACTION OF NORMAL AND DISORDERED SPEECH
- DATA COLLECTION METHOD
- LPC ORDER ANALYSIS
- USING LPC ANALYSIS TECHNIQUE
- BLOCK DIAGRAM OF ANALYSIS OF DISORDERED SPEECH
- WORKING PROCEDURE
- SPECTRAL ANALYSIS OF HYPERNASALITY
- RESULT
- Discussion and Overview
F3 of the vowel /e/ /i/ during this case is low due to F3 tuned to front versus back constriction constriction within the back produces a high F3 constriction within the front produces a coffee therefore these vowels are that the part of front vowels. The vowel /a/ /o/ /u/ for F1 hyper nasal is incredibly slightly if the frequency traditional vowels because of these vowels is that the part back vocal pronunciation and build tubular cavity house for Somali therefore say F1 the realm of the rear or tubular cavity. Linear prophetic commitment to writing analysis tries to see the properties of the vocal tract filter. in particular, it tries to see the formant frequencies, or peaks within the filter.
An LPC filter is expressed as an operation with a group of coefficients. the amount of coefficients is called the order of the filter. each combination of coefficients defines a resonance of the filter. In this experiment I have compared the speech variation of the 7 children according to their depth of the hypernasal problem. The speech samples were analyzed by a speech spectrograph, a machine that created an achievable visual examination of the phone's 3acoustic properties.
These properties included: 1) frequency, which refers to the rate at which the vocal cords vibrate throughout the composition of the sound. telephone square measure, said to be primarily high-frequency, low-frequency, or a mixture of the two; 2) time, which refers to the amount of milliseconds. The spectrograph was able to display the acoustic properties of speech patterns on a graph called a sound picture or icon. We took the three vowels /a/,/i/,/u/ from Figure 5.2, then searched for the normalized signal of a normal child and spectrographed from the normalized signal. Since the child has a perfectly normal voice, there is no hypernasal problem, so the range of voice quality was normal.
F1 has more to do with the shape of the tongue, and F2 is identified with the anterior-posterior position of the tongue. In Figure 5.7, the child needs to speak more because of his severe hypernasality. All vowels were difficult to identify with these children. spectral graph differs from other spectrographs shown in the figure above. Fig.5.14 Vowel space for CP children: F2 versus F1. For Figure 5.13, we used two frequency formants F1 and F2 to create the vowel house graph of traditional youth among youth with a congenital condition. From the figure we have shown that the vowel space of standard young people is larger due to their smooth production of the speech.
CHAPTER-06
CONCLUSION AND FUTURE WORK
CONCLUSION
FUTURE WORK
