TTS Text Normalization & Context Expansion Workbench

Raw Unnormalized Input

Raw Text with Ambiguous Tokens

Context Scenarios & Presets

Preprocessing Model Engine

20 Total Tokens

6 Ambiguous Resolved

98.4% Model Confidence

Side-by-Side Context Normalization Pipeline

Interactive Token Scrubbing & Category Map

Dr.

Inferred Entity Category TITLE

Regex Rule Expansion Doctor

Context Model Expansion Doctor

Confidence & Context Scope 0.99 (Pre-name boundary)

DETERMINISTIC REGEX RULES Naive Engine

CONTEXT-AWARE MODEL ENGINE Active Target

Canonical Proof Target:

Normalized Spoken Text: Doctor Smith paid fifty dollars and fifty cents on August fourth for two hundred fifty milliseconds of audio at one hundred twenty-three Main Street. Contact Miss Jones at five five five, zero one nine nine or ten thirty A M.