Oral Dictionary

Oral Dictionary application is developed with aim to enhance the existing dictionary,
therefore promote flexibility in English dictionary usage through implementation of
speech recognition technology. Speech recognition addresses an effective and faster way
for word definition searching as word input is retrieved through human's voice,
regardless the knowledge on the word spelling where current dictionary relies to.
Capability of providing pronunciation playback based on phonetically British English
standard overcome the major limitation of current English dictionary. The application is
using the text-to-speech (TTS) technology which is different from other commercial
dictionary software who use to record each sample of word. In dealing with homophone
words (words having similar pronunciation but different spelling), the system
successfully provide an approach to reduce the misrecognition. The approach is by
providing possible matches, relevant to the phonetic matching of the word. Oral
dictionary also offers hassle-free solution to users who only want to hear the
pronunciation of the word without having to search throughout the dictionary. Currently,
the scope of Oral Dictionary only cover 100 words, within the noun context and using
British English (Br.E) standard, as it is English language correctly spoken or written. The
application is developed using the combination of vocabulary collection process and
incremental delivery model. The major processes involve in vocabulary collection
process are words collection, syllabification and ASCII-phonetic transcription. Process
activities in Oral Dictionary application will then follow the incremental model which
allows system to be reworked in response to change request. Since the output speed is
critical, binary search tree algorithm is chosen as best data structure to be implemented in
this application to improve the searching time performance. Procedures to obtain the
result of the system effectiveness are discussed in this paper, where product evaluation I
testing is conducted to two groups of users and the result on searching time comparison,
misrecognition rate, verification on British-English standard, system functionality and
overall system interface design are discussed in detail throughout this paper.