Designing a Synthesized Content Feed System for Community Radio.

The use of text-to-speech to generate radio content is largely unexplored, despite the importance of radio in remote parts of the world, where TTS offers a robust means of transforming data into media for low-literate audiences and those without regular internet access. How suitable are TTS voices for meeting the expectations of radio listeners and what type of content are these voices best suited to deliver? We present an application for generating automated daily synthesized weather forecasts for selected locations and language varieties, based on the provision of a regularly updated weather data service. We present results from a pilot listener study aimed at exploring people’s reactions to this and other synthesized audio content, as we begin to explore best practices around the design of a synthesized content feed system for community radio.