I have trouble understanding why people here say "the modeling is bad" (I think what is really meant is the level of detail).
If this is meant to be an "as realistic as possible" representation of ancient Rome, then the level of detail is linked to our knowledge of what specific buildings may have looked like back then. That's why eg some of the statues have a very high LOD because they were preserved until today.
Now, comparing this against a multi-hundred-million budget for-profit video game seems a bit pointless, because in a video game (apart from time and money invested) there's no need for historical accuracy. But if we assume that standards of scientific accuracy apply to this visualization, then it makes a lot of sense that they didn't just invent a bunch of details. Imagine them publishing a paper and just filling some historical gaps with "this is how we think it might've looked"..
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.
If this is meant to be an "as realistic as possible" representation of ancient Rome, then the level of detail is linked to our knowledge of what specific buildings may have looked like back then. That's why eg some of the statues have a very high LOD because they were preserved until today.
Now, comparing this against a multi-hundred-million budget for-profit video game seems a bit pointless, because in a video game (apart from time and money invested) there's no need for historical accuracy. But if we assume that standards of scientific accuracy apply to this visualization, then it makes a lot of sense that they didn't just invent a bunch of details. Imagine them publishing a paper and just filling some historical gaps with "this is how we think it might've looked"..