A video recording process is commenced by a camera of a device. The video recording process produces a sequence of frames. While the video recording process is in progress, a gesture directed at the device is detected. A new first frame from the sequence of frames is displayed in a camera view user interface (UI) of the device, based on a characteristic of the gesture. A video file is generated using frames captured during the video recording process and the new first frame is set as a starting point of the video file.