Well, I have had this experience consistently and so I feel able to report it and share it with others. Some here agree (including the more 'scientific' participants), some don't. I guess that is the point of a discussion to get various views on the subject?
With regard to your point about the image forming in your eyes and brain, absolutely. My day job for the last 10 years has been as a senior research fellow in computational neuroscience at a UK university and I have studied the path from the retinal cells themselves, through the various ultra-specialised neurons at the back of the eye which help to provide its remarkable dynamic range of up to 18 bits, down the optic nerve and through the layers of the visual cortex until it finally gets to neocortex. All the brain ever 'sees' is action potentials i.e. events where the illumination falling on a retinal cell has changed up or down slightly within a given time window. So the signal travelling down the optic nerve is something like the snow pattern of a detuned analogue TV except that the white dots will have spatial orientation along edges. One way that the system preserves information flow is saccades - tiny eye movements - to refresh the edges in the scene if not much is happening dynamically. From this highly abstract and digitised signal the magic starts happening in V1 and V2 to knit together the signals, correct for the overlap and the fovea, create and maintain 3D spaces with edges, shapes, colour and an allocentric orientation of the observer within that space and portray a perceived stability and solidity which is a long way from the retinal signals themselves. It's quite a remarkable magic trick and is generating a lot of research work in order to understand the underlying computations because event-based vision and computation is at the heart of the latest developments in robotics and AI for their speed and energy benefits.