I didn't use lo-res models as the tiny trees in the background, so this image had over 22 trillion polygons. (69.6 trillion when rendering the dynamically generated parts.) Defiantly NOT a model of efficiency.
The memory of a beautiful scene tends to exaggerate the key points, trees are taller, valleys are deeper, vistas are wider, and the moon is bigger. I used a telescopic photo of the moon to get more detail.