Multilingual Script Reading Project

Multilingual Conversational TTS Data Recording Project (2024~2025)

In this project, ARAMedia managed large-scale voice data collection sessions across multiple languages including Japanese, Turkish, Polish, English, Arabic, Dutch, French, and others.
Each participant read scripts for up to three hours in controlled acoustic environments, ensuring high-quality, natural conversational recordings suitable for text-to-speech (TTS) and AI voice synthesis.

Our team handled the end-to-end process — from multilingual script preparation and participant coordination to quality assurance and file delivery — maintaining linguistic accuracy, audio consistency, and dataset compliance for AI model training.