All text materials must be normalized before any processing by IDOL Speech Server. The text normalizer prepares the data in a consistent form so that words with the same meaning (for example, “two” and “2”) are treated as being the same.

The text normalizer focuses on the following parts of a set of text:

You can also format the text as metadata to mark text that you want IDOL Speech Server to exclude from the normalizing process (for example, notes about the transcript). Metadata sections pass through the text normalizer unmodified, and IDOL Speech Server flags them with metadata tags in the normalized transcript. Metadata tags do not affect further processing, such as transcript alignment. For further information on how to use metadata tags, see Metadata Tag Syntax.