4 ms·

What he is trying to explain is that he is NOT making a permanent video record of everything he sees. Any computer vision system is going to have an image buff

• betawolf33 14y ago
  Consider how long ago these events are in his narrative. He gives a still of 'Possible Witness 1' who he met and talked with as he entered the store. Presume that that image is at the end of their conversation (as it will have been pretty long). He then describes going to the counter and ordering food. That is to say, his daughter ordered food in French. That's a minute or two of activity, at the least. Then he goes and sits down to eat his food. Assuming really prompt service and that he was attacked almost as soon as possible in the narrative, that's at least another half-minute to take a seat and unwrap your sandwich. There are then a series of images and a description of him trying to reason with several people who 'deliberated on [his documentation] for some time' before pushing him out. That, to me, is at least five minutes of confrontation. I can't see why his image buffer would be over seven minutes long.