تصميم اختبارات اللغة العربية في ضوء الذكاءات الاصطناعية: VOICEMAKER أنموذجا

Arabic Test design Based on Artificial Intelligence: Voice Maker as a Model

Abstract

This simple research aims to reveal some of the superior features of the voicemaker.in site and some of the techniques used in reproducing artificial human voices for listening skills texts on standardized Arabic tests, while also offering solutions to some of the obstacles in using the site. This research uses a qualitative approach with a case study model and triangulation technique in analyzing the findings obtained through observation and interviews. The results of this study are: (1) this site pampers its users with several voice options based on gender and nationality, of which 14 Arabic native voices are currently available; (2) there are features that make it easy for site users to get the desired recordings through speed, volume settings. , formats, effects, tones, sizes, and sample rides; (3) the site can be accessed for free with a maximum limit of 400 words per recording and 30 languages ​​with 70 different natives, as well as cloud storage is available; 4) the steps for using the site start with logging in go to the url address, login with email, type the text to be recorded on the white board provided (you can use the copy-paste method), select the driving engine, language and native (male or female), give a touch according to your taste through the settings feature provided. available, press the record button then download the results (if the results are not suitable, you can repeat by improving the available setting features); 5) The obstacles that are often found are the pronunciation of harakat, the pronunciation of nakirah and makrifat, and the pronunciation of two words into one word. all of this can be overcome by placing the harakat according to the i'rab rules and above the alif lam qamariyah (makrifah) and placing the second word on a new line below it.