You can create most of the details using procedural modeling, automatically. We also know enough about the actual detailing of the buildings to create much more

You can create most of the details using procedural modeling, automatically. We also know enough about the actual detailing of the buildings to create much more detailed representation of Rome than we see in this video. See e.g. https://www.youtube.com/watch?v=zx_8RbFNjes https://www.youtube.com/watch?v=zx_8RbFNjes https://www.youtube.com/watch?v=necqGAYknIY https://www.youtube.com/watch?v=necqGAYknIY https://www.youtube.com/watch?v=5dxQbqVB7iU https://www.youtube.com/watch?v=5dxQbqVB7iU