Live video streaming of outdoor and built environment could be a source of scenarios generations for training games, especially when scenarios are on real-time dynamics of the environment. Such so-called live virtual reality (LVR) is, among other things, one of the most recent gaming technologies. There is an increasing need for effective solutions that can be utilized to process videos in attempts to gain desirable result, as the requirement of video quality is becoming higher and higher. The issue at the core of real-time transcoding is to control the right combination of transmission bitrate and resolution. This work aims to develop a prototypical workflow for dewarping videos from fisheye camera to the cloud and re-stream it to a client. This prototype is used for testing combinations of bitrate and resolution in various living scenarios of indoor, outdoor and street traffic environments. The study of performance is enabled by KTH students who supply feedbacks about quality of experience. Comments of observers prove the promising use of LVR in training and education games. The results reveal that the combination of bitrate 3.5–4.5 Mbps and resolution 720p is best suited for transmission in order to avoid noticeable lagging.