A method and a device for preparing a media stream containing more than one component stream for picture-in-picture applications are described along with the corresponding method and device for rendering the media stream in a picture-in-picture mode. The invention allows for live and dynamic picture-in-picture rendering of the component streams contained in the media stream. Extensions to the media formats that shall support various multimedia applications, such as the MPEG-2 Transport Stream and ISO media format, are proposed to enable such a dynamic picture-in-picture functionality.