Theater View Optimization
Palace Theater View Optimization
Shulong Ren, Xueyin Lu, Yunlong Fan, Yukun Tian, Juno Lee | Generative Design |Fall 2021 | GSAPP
Introduction
To ensure a good viewing experience is the focus of theater design. We found that the viewing experience depends on the distance between the audience and the stage, the Angle from which the seat deviates from the central axis of the auditorium, the Angle from which the audience needs to turn their heads, and the sight occlusion area. We established an evaluation system for the viewing experience.
The best viewing distance and Angle can turn through the plan of the auditorium and adjust the seat Angle to find the best solution. The deviation from the stage center axis is inevitable. In the general case, sight occlusion caused by the front audience’s head. In the traditional theatre design, C value is the parameter that evaluates sight occlusion. By adjusting the stage slope, seat spacing and stage height, the architect ensured sufficient C value to avoid obstacles.
However, in some existing theatres, the sight occlusion has always existed due to space limitation and unreasonable seat arrangement. We tried to minimize sight occlusion area by optimizing seat position through Genetic Algorithm in the existing theatres without changing the original building structure.
Methodology
We chose Palace Theater in Saint Paul as our optimization target. The century-old Palace Theatre was originally operated as a vaudeville theatre in 1916. The first floor of this theater is completely horizontal. For flexibility, there are no fixed seats in the audience area. Due to the limited space on the upper floor, the slope of the auditorium does not meet the C value, which will block the view.
Proposal 1
For the first floor, we put forward a more radical design strategy. Seats are randomly generated in this area. We then calculate the average sight occlusion area of each choice, and optimize the whole area by Genetic Algorithm. Flexible arrangement can fit different types of performance.
Proposal 2
For the upper floor, we kept the traditional arrangement of the auditorium — arranged in rows — and the number of seats. We used the original structure of the auditorium, and only adjusted the left and right spacing of each seat to reduce the shielding area of the audience.
Algorithm
The algorithmic parts of both proposals are essentially the same:
First, we draw a cone with the eyes of a random viewer as vertex and the head of all other audiences as base. After extending the length of the generatrices, we selected the cones that have intersected area with the screening. We found that those who are far from this audience have little influence. So in order to simplify the calculation, we only calculate cones in a small range. We also used collision detection to prevent any two seats from being too close or out of bounds.
For the first floor, we input the boundary curve of the audience area and the XY coordinates of all seats to calculate the occlusion area of each audience by the above method, and output the average value.
For the upper floor, we input the position curve of each row and the number of seats in each row, calculate the occlusion area of each audience by the above method, and output the average value.
Results
Proposal1
The trend of the optimal result is that the seats are concentrated in the front, and the front follows the principle of staggered arrangement. The back row is scattered as far as possible, and the distance between the front and rear rows is enlarged, which will reduce the shielding of the front row from the back row.
Proposal2
The trend of the optimal results is that the farther away from the stage, the less the seat in the rear row is staggered than the seat in the front row. Conversely, the more directly facing the stage, the greater the staggered distance of the rear row (up to 50% of the width of the seat). This tendency gives everyone the best view.
Conclusion
The result of proposal 2 is closer to expectation, which proved the reasonability of the algorithm. Proposal 1 is a brand new attempt to arrange the auditorium. Although the limitation is that the seating arrangement is too chaotic and the space utilization efficiency is low, it has the potential for further development. For example, more complex collision algorithms can be used to make seats less left to right and more forward to ensure audience accessibility.
