Reading Sounds

Simulation of Subjective Closed Captioning Quality Assessment Using Prediction Models

“…Recently, regulators have attempted to include quality provisions in the CC regulations but have only adopted quantitative measures of accuracy as their de¯nition of quality. While errors are important to assess, there are also other factors that a®ect quality [1,4,6]. In addition, the D/HOH communities are advocating for their subjective perspective to be included so that they can have input into CC quality assessment as a procedure and as speci¯c ratings for various shows [1,7].…”

“…The core quality factors of CC have been identi¯ed in the literature; they include: synchronization delay (between caption and the actual audio), presentation speed, number of missing words between the captions and the audio transcript, number of 46 S. Nam & D. Fels spelling and grammar mistakes, coloring and positioning, display methods, verbatim state, speaker identi¯cation, and the inclusion of captions for non-speech sounds [4,7,8]. In this paper, we selected the¯ve factors to be used including delay, speed, spelling and grammar, missing words, and verbatim state as these have been iden-ti¯ed by D/HOH viewers as the¯ve most important factors a®ecting quality [1,6,9,10]. The verbatim state (or verbatim accuracy) factor of CC represents the quanti¯ed accuracy of the translation from the spoken words in audio format to CC word by word [4,5].…”

“…Closed Captioning (CC) is intended to share the television experiences that hearing audiences enjoy with Deaf or Hard of Hearing (D/HOH) viewers by displaying the text equivalent of audio [1]. However, because of the intensity of the translation process, the speed of speaking and human errors, CC is a challenging task, particularly for live or real-time content.…”

“…The current live captioning method involves human captioners who type or respeak the words that are spoken in a show. Typed captions use the stenograph keyboard and comprehensive dictionaries similar to court reporters [1]. Respeaking method require captioners to listen to what is said in the live program, then repeat the words to a trained voice recognition system to generate captions [2].…”

The End of the Old and the Birth of a New Order, 1918‑1919

“…From the questionnaire distributed by the Russka Narodna Rada, represented by the priest, Emilian Nevyts'kyi, reproduced in PeškaZdeněk;Příspěvek v ústavním dějinám Podkarpatské Rusi". Bratislava 5: 526.…”