Convert the whole input image into the requested visual medium and style; the conversion must cover the entire frame, not a local filter or only some objects. Keep: composition, aspect ratio, viewpoint, subject placement and pose, spatial relations between objects, and the position and legible content of existing signage and text; add no slogans, captions or watermarks. Convert: repaint every object in the requested style with consistent brushwork, colour layering, material rendering and light-colour relationships; light direction matches the original. Output one single converted image.
