Perceptual Distribution Of Monosyllables And Hearing Assessment

Language is one of the most important communication way of human.People produce sound or speak based on some given natural language model.The sound signals produced purely or when people are talking are called speech.Production,perception,and interaction of speech signals constitute speech chain.Speech perception is the most important part of the speech chain.In area of phonetics and clinical audiology,there are plenty of research focused on speech perceptual computing and hearing assessment.In this paper,we set the beginning of our research at the stage of auditory awareness,and auditory distinction.In this work,we build the perceptual vector in the level of monosyllable,compute the perceptual distribution of monosyllables,figure out the computational relationship between the acoustic physical property and the perceptual property of speech signals,and propose a method to compute vocabularies in clinical audiometry based on this computational relationship.Contributions of this paper can be concluded as follows:· We propose an algorithm of pure tone audiometry based on multiple judgement.In this work,we propose a model to transform amplitude to SPL of speech signals with acoustic parameters based on some auditory characteristics.We design an adjustment rule of test sound signals during the pure tone test based on human auditory perception.And we propose an algorithm based on multiple judgment to compute the hearing threshold of pure tone test.This work increase the efficiency and the validity of pure tone assessment,ensue the astringency and the reliability of the test results.· We propose a systematic approach to compute the perceptual distribution of monosyllables based on perceptual vectors of syllable.In this work,we generate a perceptual vector in the level of monosyllables based on acoustic physical property,construct perceptual distribution of monosyllables based on the perceptual vectors.This work bring the speech perceptual computing from the level of consonants vowel and tones to the level of syllables,increasing the accuracy of the describing of perceptual properties based on acoustic physical properties.This work produce more possibility for speech perception assessment.· We propose a method to design vocabularies in speech assessment based on perceptual distribution in the level of monosyllables.This work generate the confusion items based on the result of perceptual distribution of monosyllables,and generate the vocabularies based on the confusion items.This work combine the result of speech perception computing and speech assessment,increasing the validity and reliability of speech audiometry.· We propose a method to give speech audiometry based on the vocabularies above.This method set different weight to calculate the result of speech assessment considering human auditory characteristics.This work increase the validity and reliability of speech assessment.