There's two kinds of documentation you can reasonably achieve, forensic, and narrative.
Placing a camera in the room and pointing it at the screen is forensic. You can go much farther than a camera in the room, and record an in-experience motion trace of objects in the virtual world, as well as HMD and controller pose.
Narrative documentation is crafted recordings, meant to communicate specific points, and the creation of the narrative documentation is no different than making a movie.
The fundamental problem with forensic documentation is the same as in court room. Maybe the camera was pointed in the wrong direction when the murder occurred, or you neglected to record a crucial signal such as in-experience audio cues.
The fundamental problem with narrative documentation is that you can't omnisciently provide the answers to future questions.