mgrego2:
Yes, I think this goes to my point about needing to examine the entire vision system (eyes and brain together) as a working unit. And the need for a trained subject, not just anyone.
Rolf:
The realtime processing is part of the system. Why exclude it? For what it's worth, I feel that the emotional state of the subject may also be a factor. It might have an impact on the ability of the brain to process those peripheral images.
In addition, you'd need to establish some norms for lighting, colors, size and quantity of text, duration of exposure, and so on. It's a difficult and complex question.
Finally, you're in a difficult logical situation: you've essentially set out to prove a negative (photereading can't work as advertised). Most investigators would probably start with an observed instance and try and design a trial that tests why and how the process does seem to work.
James Randi would probably try to duplicate the performance by some covert means, thereby exposing the "trick." That could be one approach. At some point, Occam's Razor is in play, though. The simplest answer is perhaps the best: maybe the system just works.
--George