Scenario description uses a picture or video frame as a speaking prompt.
Start with nouns, add actions and locations, then connect them into complete Chinese sentences. The visible scene makes output more concrete and reduces the pressure of finding a topic.