raph: As I noted in response to a similar post by telsa last fall, crowd sizes are notoriously overestimated. I think your four square feet is extremely conservative. But it's at least reasonable to put an upper bound on crowd size.
Years ago, when I was a young journalistic whippersnapper in Pasadena, Calif., I tackled the Pasadena Police Department's traditional estimate of 1 million people along the Rose Parade route. I asked the cops how they came up with it year after year and they said they looked out the helicopter window and said it seemed pretty much like last year, and last year it was a million so it must be a million this year, too. Any idiot with a calculator can demonstrate how absurd that number is: 5-1/2 mile parade route, both sides of the street, that's 58,080 feet of parade route. One row of people standing 18 inches apart, up one side and down the other, is 38,720 people.
The 18 inch assumption is too conservative - people don't pack that tightly together. But even with that assumption (and subtracting 100,000 people in grandstands) you'd have to have a crowd 23-plus rows deep to get to a million. If you look at the parade route, the crowd is rarely more than 5 rows deep. Twenty-five rows wouldn't fit on the sidewalk. So we went out and actually did our own crowd counts and made our own estimates, and concluded it was around 300,000, and likely less.
Every year, I eagerly read the Rose Parade story out of Pasadena. Every year, the cops estimate the crowd size at a million. My work, apparently, was for naught - except I love retelling the story at parties. And I learned a lesson. Now when I'm covering anything like that, I try mightily to actually do a rigorous count/estimate, rather than asking the organizers or the cops.