21st Century Eloquence, the Speech Recognition Specialists

Vocabulary Builder Tips

Preparing Documents

Vocabulary Builder can use only file formats as: Text files (.TXT), Rich text format files (.RTF), Microsoft Word files (.DOC), version 6.0 or later, Corel WordPerfect files (.WPD), version 8.0 or later, HTML files (.HTM and .HTML).
Running a spell checker on the documents you will process will save you time by preventing Vocabulary Builder from identifying misspelled words as new, unknown words. It is much easier to do this beforehand and it helps create a more accurate language model.
Text files process faster than the others so, if you have a choice, use text files.
Microsoft Word (.DOC) files are supported only if you have Microsoft Word installed on your computer. To process Microsoft Word 97 documents, you must either have Microsoft Word 97 installed or first convert the documents to another format.
Corel WordPerfect (.WPD) files are supported only if you have Corel WordPerfect installed on your computer. If it is not installed, you must convert the documents to another format.

Using the Vocabulary Builder

On the Tools menu, click Vocabulary Builder.
Click Add Document to select documents for processing.
Click Begin to start processing.
The New Words dialog box displays, listing the unknown words found in the document.
Click to place a check mark next to the word(s) you want to add to the active vocabulary.
Add words that you use frequently and that are important for your particular subject area. Do not add words you rarely use, odd word forms that you think would not be useful, or capitalized words unless the capitalized form is significant.
Optionally, you can proceed without selecting any words. Either way, you will modify the language model from this dialog box.
If necessary, individually edit words before adding them to correct misspellings, typographical errors, or unusual capitalization. Double-click a word to edit it.
Click the Train and Build (or Build) button.
If you have selected words, the Train Words dialog box opens for you to train the new words. When you are finished training, the new words are added to your active vocabulary and the language model is modified.
If you have not selected any words, Vocabulary Builder modifies the language model without adding any new words. Vocabulary Builder analyzes how often known words in the processed documents are used alone and in combination with others. This information helps improve recognition accuracy, even without adding or training unknown words.