Not exactly nothing ...
For me, part of understanding what we mean when we talk about the concept 'natural sound' involves thinking about how it ties to live acoustic music, the reference from which the notion derives. Natural sound is a goal for stereo listening.
Imaging is a function of listening to a stereo and one of its effects on listeners is our minds supplying or translating what we hear into something visual or quasi-visual. I say 'something' as I don't know how that varies across people. Maybe it is a picture in our mind of performers on a stage. Maybe it is a sense of the location of sonic origins. Maybe it is a thought picture of instruments in our heads -- hear a timpani, imagine a timpani. Whatever you wish to call the experience.
I find little discussion of imaging tied to listening to a live acoustic orchestra. I believe such talk is largely derivative from talk about stereo listening and largely the province of audiophiles, that is applying what we think we understand about imaging in stereo listening to a live event. So, in a way, live acoustic listening imaging is the reverse of natural sound listening. I never hear concert-goers talk about it.
There is a psycho-acoustic aspect to concert hall listening. The ambience of the hall when musicians play and when they are not playing -- how sound exists in the hall and how the hall impacts that sound. There I believe we have a sense of energy in a space. For me the extent to which stereo listening represents that psycho-acoustic is definitely one aspect of natural sound.
I don't believe we exert or intend control of what images occur in our heads while we listen. As a non-audiophile listener in the concert hall I don't think about imaging there. If I have a different imaging experience when listening to reproduction than I do when listening in the concert hall, then no, imaging at best is neither necessary nor sufficient to gauge a stereo as sounding natural.
A question that came to me as I wrote this: is directionality an image?