Mobile-Based Video Assessment for Speaking Skills: Improving Pronunciation, Fluency, and Confidence in ESL Learners

The speaking component is the focal area that many ESL learners face challenges, this includes pronunciation, fluency, and confidence. Most of the conventional resources for language learning address and teach the constituents disjointedly while ignoring their interaction with the technicalities and emotive content of speech. Meeting these challenges demands an out-of-date, multimodal perspective that involves the context of speech and feeling. According to this research, the MT-SEC model is suggested to be a multitask transformer model for learning pronunciation, fluency, and confidence with attention to speech features and emotional content of ESL learners. The necessity of touching on technical as well as emotional dimensions associated with speaking is a valuable feature of the model. A pre-test and post-test design was employed to assess the effectiveness of the MT-SEC intervention, showing significant improvements across all speaking dimensions: phoneme errors of 50.6%, an increase in speech rate by 50.0%, fluency coherence increased by 29.2%, and an increase in emotional confidence by 38.7%. With regards to the application based on the mobile the learners affirmed that they found user satisfaction with it to be better by 30.6%. It is thus found that the MT-SEC model has the capability to help target individualized emotion sensitive feedback, thus the proposed model is indeed a one-stop solution for ESL learners, addressing technical and emotive aspect of their written and spoken language.