I am reading and doing exercises using the book Rendering with Radiance.
In P55, when I use the rvu command to view the stool, I am confused about the functions of -vu -vd -vp and how they work together to establish the vantage point. What I understand is that -vp is used to determine the position(coordinates) of the viewpoint (like the camera); “-vd” the vector to determine the view direction of viewpoint (the vector from 0 0 0 to the vector gave by -vd defines the direction in which the camera faces). Am I right for the function of -vp and -vd.
In this case, what is the function of -vu and how it works?
Yes, your understanding is correct. One easy way to compute the -vd vector is to find the point in your scene that you want centered in the image. Subtract the view point (-vp) from the center of view, and that can serve as your view direction (-vd) vector. This vector does not need to be normalized, and some rendering options even use its length to get focus distance to compute depth-of-field effects.
The view up vector (-vu) is used to orient the camera “roll” to 0 relative to the specified vertical direction. It keeps the camera “upright,” and only really needs to be changed if the vertical of your scene is non-standard, or you want the camera not parallel to the horizon.