I don’t know exactly how Ben does it, but you might want to check out this video and blog post:
From the way I interpreted it, I literally visualize images placed from left-to-right or top-to-bottom and the relationships between different images sort of grow out of practice.
Examples:
- 51 for me is a bamboo grove. If it comes first in a group, the second image is emerging from the right side of the bamboo grove. If the bamboo comes second, then the first image is in the top of the bamboo. (The bamboo in my image isn’t as tall as in that video though.)
- 01 for me is a glass of iced tea. If it comes first in a group, it usually has fallen over to the right and the second image is coming out of the glass, or the second image is submerged under the tea. If the iced tea comes second, then the first image is floating on the tea.
What the images are doing sometimes depends on what the previous or following images are doing.
I experimented with three images per location but went back to two images per location because it seemed to work better for me.