In the Canon world, RAW format images can be saved as "Large" or "Medium" or "Small". I always choose "Large" to get the most detail possible. But, with the camera mounted on a tripod and taking 3 identical shots, one at each setting, would there be an easily noticeable difference when viewed as an entire picture? Probably not.
But when enlarging those images, that's where the difference would show up...'edges' of items. Think of a diagonal object, such as a yard rake leaning against a wall. Now visualize that as a staircase when viewed from the side. With "large" RAW (or JPG files, for that matter), the 'steps' on the staircase would be quite small, and the differences between light areas (the wall behind the rake) and the rake handle would be quite discrete. In the "small" format, what was 4 adjacent 'boxes' (pixels) of one light color value each in the 'large' is 'averaged' by the computer in the camera to be one single value in a single box the size of one of the 4 original boxes. In short, some detail is lost. In all likelihood, unless one blows up the picture to 'poster size', the differences won't be terribly discernible.
Consider this, my first digital camera was 3mp Canon G3 point-and-shoot. The pictures, even when projected on an 8'x8' screen were quite acceptable. Are the pictures from my current 22+mp 5Diii clearly more sharp on the same screen? Not to the average viewer. In my estimation, the resolving 'power' of the sensor, lens, and even the noise level have more of an affect on image quality than original image size. Even reduced quality pictures posted on this forum are more than acceptably sharp.
In reference to the tall building referred to by the OP, unless the shot has a large DOF -AND- shot from roughly half-way up the building, DOF issues and non-parallel-sensor-to-the-building issues will result in more loss of IQ than how many megabytes the image is compressed to.
