Speaker-independent ASR for Modern Standard Arabic: effect of regional accents

Droua-Hamdani, Ghania; Selouani, Sid‐Ahmed; Boudraa, Malika

doi:10.1007/s10772-012-9146-4

Cited by 14 publications

(5 citation statements)

References 2 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The significant effects of speech quality for syntactic and semantic interpretation accuracy are consistent with previous research findings, demonstrating that speech recognition technology is dependent on audible clarity in order to consistently provide accurate responses (Rohlfing et al, 2020;Luger & Sellen, 2016). The significant effect of speaker accent for semantic interpretation accuracy is consistent with previous research analyzing differences in software response accuracy relating to participant accents (Wilson et al, 2017;Palancia et al, 2019), but is contrary to accent studies examining ASR performances of speech software (Dichristofano et al, 2022;Droua-Hamdani et al, 2012). The final main effect of speech rate was not significant for either WER or confidence scores, contradicting previous findings suggesting that speaking at moderate speech rates compared to fast and slow rates results in higher user satisfaction due to greater software response accuracy (Choi et al, 2020, Christenson, 2020.…”

Section: Discussionsupporting

confidence: 88%

Talking With Siri: Analyzing and Detecting Error Patterns in Speech Recognition Technology

Steger

View full text Add to dashboard Cite

Voice assistants are a subset of speech recognition technology that continue to exponentially improve in terms of their accessibility and functionality in response to user voice commands. The current research examined the effects of speaker accent, speech quality, and speech rate on syntactic and semantic interpretation accuracy for the voice assistant Siri. 60 adult participants read standardized command phrases to Siri to determine syntactic and semantic interpretation accuracy differences between Canadian and non-Canadian accents, masked and unmasked speaking, and slow, moderate, and fast rates of speech. Analyses demonstrated that poorer speech quality inputs resulted in significantly lower syntactic and semantic accuracy interpretation scores. Furthermore, non-Canadian accents generated significantly lower semantic accuracy interpretation scores. Accuracy measurements for speech rate and speaker accent (syntactic accuracy) did not support previous research findings. These results signify the importance of additional algorithmic training improvements and continuing to investigate influential factors on voice assistant response accuracy.

show abstract

Section: Discussionsupporting

confidence: 88%

Talking With Siri: Analyzing and Detecting Error Patterns in Speech Recognition Technology

Steger

View full text Add to dashboard Cite

show abstract

“…At present, the most popular and successful speech recognition systems use Hidden Markov Models (HMM) [23][24][25][26] in the acoustic modeling. HMM is used to train the acoustic models of sixty one phonemes along with a model of silence (sil).…”

Section: Experimental Setup and Discussionmentioning

confidence: 99%

“…This greatly enhances the performance speech recognizers. The derivatives of the features are calculated through the use of regression formula [23].…”

Section: Frequency Partitioning and Wavelet Based Feature Extractionmentioning

confidence: 99%

Articulation based admissible wavelet packet feature based on human cochlear frequency response for TIMIT speech recognition

Biswas

Sahu

Bhowmick

et al. 2014

Ain Shams Engineering Journal

View full text Add to dashboard Cite

To deal with non-stationary and quasi-stationary signals, wavelet transform has been used as an effective tool for the time-frequency analysis. In the recent years, wavelet transform has been used extensively for feature extraction in noisy speech recognition. These filters have the benefit of having frequency bands spacing similar to the auditory Equivalent Rectangular Bandwidth (ERB) scale. Central frequencies of ERB are equally distributed with the frequency response of the human cochlea.This paper deals with the speaker-independent Automatic Speech Recognition (ASR) system for continuous speech. This Hidden Markov Model (HMM) based ASR system was developed for English using recordings of four regions taken from TIMIT database. A new set of features were derived using wavelet packet transform's multi-resolution capabilities and having an advantage of ERB filter based on the human cochlea. New set of wavelet features have shown significant improvements in the noisy environment, especially at low SNR values.Ó 2014 Production and hosting by Elsevier B.V. on behalf of Ain Shams University.

show abstract

“…Berdasarkan pemaparan tersebut, sintaksis sebagai cabang ilmu linguistik yang membahas penataan dan pengaturan kata-kata, serta merupakan subsistem gramatika yang berfokus pada struktur juga turut mengalami perkembangan seiring perkembangan linguistik. Selain membawa pengaruh terhadap sintaksis, perkembangan linguistik normalnya berpengaruh juga terhadap perkembangan bahasa yang menjadi objeknya, termasuk bahasa Arab sebagai salah satu bahasa yang cukup khas dengan konsep tata bahasanya (Hamdani, 2012).…”

Section: Introductionunclassified

“…Bahasa Arab sendiri memiliki klasifikasi berdasarkan penggunaannya. Ghania Droua dalam Hamdani (2012) menyebutkan bahwa bahasa Arab terdiri dari Modern Standard Arabic (MSA) dan Colloquial Arabic (CA). Menurut penjelasannya, MSA adalah bahasa Arab yang digunakan di ruang-ruang resmi, seperti institusi pendidikan, institusi media, dan komunikasi formal pada umumnya, sedangkan CA digunakan dalam komunikasi sehari-hari.…”

Section: Introductionunclassified

Fungsi, Kategori, dan Peran Sintaksis Bahasa Arab: Perspektif Linguistik Modern

Fitriani

2023

ICEIL

View full text Add to dashboard Cite

Penggunaan bahasa Arab semakin meluas dengan latar belakang pengguna yang juga semakin beragam. Oleh karena itu, dibutuhkan konsep kaidah yang bersifat universal agar bisa diterjemahkan dan diselaraskan dengan perspektif yang berlaku secara umum, sebagaimana prinsip universal grammar dari Chomsky. Konsep tersebut ditemukan dalam perspektif linguistik modern. Oleh karena itu, peneliti merumuskan penelitian yang mencoba untuk membuka paradigma sintaksis bahasa Arab menurut perspektif linguistik modern. Tujuan dilakukannya penelitian ini adalah untuk menganalisis konsep fungsi, kategori, dan peran sintaksis bahasa Arab berdasarkan perspektif linguistik modern. Metode penelitian yang digunakan adalah metode deskriptif berdasarkan jenisnya sebagai penelitian kualitatif. Adapun data dan sumber data yang digunakan adalah buku yang menyediakan satuan sintaksis bahasa Arab. Data-data yang diperoleh dianalisis dengan metode simak dengan teknik analisis yang terdiri dari tiga tahap, yakni penyediaan data, analisis data, dan penyajian hasil. Adapun hasil penelitian ini menunjukkan tiga poin: pertama, bahwa dalam bahasa Arab dikenal dua pola fungsi dasar sintaksis, yakni subjek-predikat-objek/komplemen-keterangan dan predikat-subjek-objek/komplemen-keterangan, 2) kategori sintaksis bahasa Arab tidak hanya fi’l, ism, dan harf, melainkan diperluas menjadi verba, nomina, adjektiva, adverbia, preposisi, pronomina, numeralia, dan konjungsi, serta 3) Analisis peran sintaksis pelaku, tindakan, sasaran, dan pelengkap juga dapat diterapkan dalam kalimat bahasa Arab.

show abstract

Speaker-independent ASR for Modern Standard Arabic: effect of regional accents

Cited by 14 publications

References 2 publications

Talking With Siri: Analyzing and Detecting Error Patterns in Speech Recognition Technology

Talking With Siri: Analyzing and Detecting Error Patterns in Speech Recognition Technology

Articulation based admissible wavelet packet feature based on human cochlear frequency response for TIMIT speech recognition

Fungsi, Kategori, dan Peran Sintaksis Bahasa Arab: Perspektif Linguistik Modern

Contact Info

Product

Resources

About