From my experience this only ever looks good in a recorded video. Because we are used to assume that the real object captured within the video is perceivable in a stereoscopic form. But it‘s not. The real usage of head and viewtracking is as a controller, not to improve immersion.