Development of the Arabic Voice Pathology Database and Its Evaluation by Using Speech Features and Machine Learning Algorithms
Table 9
Comparison of AVPD with two publicly available voice disorder databases.
Sr. number
Characteristics
MEEI
AVPD
SVD
Language
English
Arabic
German
Recording location
Massachusetts Eye & Ear Infirmary (MEEI) voice and speech laboratory, USA
Communication and Swallowing Disordered Unit, King Abdulaziz University Hospital, Saudi Arabia
Saarland University, Germany
Sampling frequency
Samples are recorded at different sampling frequencies (i) 10 kHz (ii) 25 kHz (iii) 50 kHz
All samples are recorded at same frequency (i) 48 kHz
All samples are recorded at same frequency (i) 50 kHz
Extension of recorded samples
Recorded samples are stored in .NSP format only
Recorded samples are stored in .wav and .nsp format
Recorded samples are stored in .wav and .nsp format
Recorded text
(i) Vowel /a/ (ii) Rainbow passage
(i) Vowel /a/ (ii) Vowel /i/ (iii) Vowel /u/ (iv) Al-Fateha (running speech) (v) Arabic digits (vi) Common words (All vowels are recorded with a repetition)
(i) Vowel /a/ (ii) Vowel /i/ (iii) Vowel /u/ (iv) A sentence
Recording of vowels
Only stable part of the phonation
Complete phonation including onset and offset parts
Only stable part of the phonation
Length of recorded samples
Normal (i) Vowel: 3 sec (ii) Rainbow: 12 sec Patient (i) Vowel: 1 sec (ii) Rainbow: 9 sec
(i) Vowel: 5 sec (ii) Al-Fateha: 18 sec (iii) Digits: 10 sec (iv) Words: 3 sec (The length of a complete recorded sample is 60 sec approx.)
Vowels: 1~3 sec Sentence: 2 sec
Ratio of normal and pathological subjects
Normal: 7% Pathological: 93%
Normal: 51% Pathological: 49%
Normal: 33% Pathological: 67%
Perceptual severity
✗
✓ Perceptual severity is rated on a scale of 1 (low) to 3 (high)