A content provision system provides a content reproduction apparatus with content including main information which is at least either of video information and audio information generated by a content generation apparatus. The content generation apparatus generates a bodily sensation element by applying numeric conversion to characteristics of a bodily sensation given to a user, generates instruction information to reproduce the bodily sensation element in synchronism with reproduction of the main information, generates bodily sensation information by adding the instruction information to the bodily sensation element, and generates content including the bodily sensation information by adding the bodily sensation information to the main information. The content reproduction apparatus separates the main information and the bodily sensation information from the content, reproduces the main information, separates the bodily sensation element and the instruction information from the bodily sensation, and reproduces bodily sensation element based on the instruction information.