It strongly depends on the book but in general a book will focus on just giving you enough of a description that matters for the story. Some will describe a wall by just saying, "there was a wall". Others will describe the features of the wall that may be relevant to the story, "it was made of brick that you could tell was repaired often due to the changes in color". Some books describe a wall with a whole history of where the bricks came from, how they were made with the ground up bones of local pets, and the fact that its curvy playful design was meant to invoke joy in order to hide the evil origin. In a movie, such a wall would only look a certain way based on how the designer wanted it to look, but you don't get the additional context unless they have the actors specifically say something about it (which usually comes off unnatural). In a book, only the things the author describes actually matter, and the rest can be up to you. What is a curvy playful wall? One that wiggles back and forth? One that has circular holes in it? Is it colorful? In full honesty, in this example none of that matters because as long as you imagine something "curvy" and "playful" then any wall will work.

When talking about historical information or documentation, you are absolutely right. Lots of words are needed to describe what one photo will give, and lots of photos are needed to show what one video will give. I argue we are at the point where VR models should be considered for documentation since a video can capture everything so long as you look at it at every angle, but what about with different lighting? Why stop there? What about X-ray videos as well? In the end it goes back to how much is needed to share the important information. Is it a wall, or 3 terabytes of digital information with full spectral 3D imaging?