Research for the Developing of Assistive System of the English Learning for Japanese Blind Students and their Cognitive Process of English Learning

阅读不可避免地依赖于读者的知觉或视觉效率。本研究调查了以英语为外语学习的视障大学生的单词再认速度，以开发英语学习辅助软件来弥补他们的视觉问题。受试者的视力有困难，视野也有困难。与好读者相比，我们发现受试者中有两种阅读速度慢的人。视力正常和一种低视力组(LV1)的好读者可以在几乎相同的时间段内阅读单词，而不考虑单词的长度。LP(每秒字母数)的速率随着字长的增加而线性增加。然而，另一种类型的低视力(LV2)的LP是恒定的。这意味着他们对单个单词的阅读时间会随着单词的长度而增加。阅读是一项压力很大的工作，特别是对于后一类低视力学生。然后我们开发了一款名为‘ReadKon’的阅读支持软件，该软件配备了…更多的语音合成设备。ReadKON针对这两组低视力阅读器进行了测试。两组每个单词的解码时间都缩短了，他们准确阅读的单词数量增加了。结果表明，ReadKON适用于任何类型的阅读速度较慢的视障读者的英语学习。其次，我们进行了一项试验性研究，为缺乏盲文阅读技能的视障人士开发了一种英语语音学习测试。语音测试有三种版本：自动从文本转换的合成语音、被试者自己用计算机操作的旁白数字录音和DAISY。受试者可以自己回答三种类型的考试。观察听力策略，如单词拼写、重复听力、章节间跳跃、多语言语音合成听力和速度听力。受试者对声音质量的评价，对人声的评价更高。操作性能和对测试的用户友好性的评价取决于测试者的技能和经验。最后，我们对盲文使用者进行了两种类型的英语听力测试。其中一项是日本领先的英语水平语言评估，另一项是国家大学入学考试中心为大学申请者举办的英语听力测试。我们观察到盲文使用者在英语听力和阅读的平行加工上比视觉使用者弱，这是认知资源分配困难的原因。较少

Reading inevitably depends on perception or visual efficiencies of readers. We investigated the word recognition speed of visually impaired college students who studied English as a foreign language to develop assistive software of English learning compensating for their visual problems. The subjects have difficulties in visual acuity and also in their visual field. Compared with good readers we found two types of slow readers in our subjects. Good readers with normal vision and one type of low vision group (LV1) can read words in almost the same time period independent of their length. The rates of lps (letters per second) increased linearly with word length. The lps of the other type of low vision (LV2), however, were constant. This means that their reading time of individual words increases according to the length of words. Reading is stressful work especially for the latter type of low vision students. Then we have developed reading support software called ‘readKON' equipped with a … More speech synthesis device. The readKON was tested for the two groups of low vision readers. The time of decoding each word shortened for both groups, and the number of words they read accurately increased. The results show that the readKON is applicable to English learning of any type of slow readers with visual impairments. Secondly, we conducted a trial study to develop a scholastic test by voice for English, for the visually impaired who lack the skill of Braille reading. Voice tests were given in three versions ; synthesized voice which was automatically transformed from text, narrated digital recording operated by the testees themselves with computer, and DAISY. Subjects could answer tall three types of the test by themselves. Listening strategies, e.g., spelling-out of words, repeat listening, jumping between chapters, multi-language voice synthesis listening, and speed listening were observed. The evaluations by the experimentees, regarding the sound quality, was higher for human voice. Operational performance and the evaluation for the user-friendliness of the tests depended on the skill and experience of the testees. Finally we conducted two types of English listening comprehension test to Braille user. One was EIKEN Japan's Leading Language Assessment for English proficiency, the other was a trial test for English listening comprehension for college applicants by the National Center for University Entrance Examinations. We observed Braille users were weak in parallel processing of listening and reading of English than sighted cause of allocation difficulties of cognitive resource. Less