I can perhaps answer some of that from myself, which may not be a universal experience of it. When I try to, I always see whole words. If I need to spell it out loud, I bring up an image of the word and then look at each letter and say it.
When I see stuff like that, if anything it is like a chalkboard in my head in its own visual universe about where a "third eye" might be. Not coherent or overlayed into the real world. Amusingly, my eyes look up and scan like reading when I am putting effort into recalling something visually.
I don't think my pictures are fully encoded as much as they represent really rich patterns that are recconstructed as needed.