Challenge on Sound Scene Synthesis: Evaluating Text-to-Audio Generation