There are a variety of approaches. This was recorded on macOS with Screenflick¹ and composited with Adobe After Effects. Screenflick records very high quality (both fps/resolution). Recording on an Apple Studio Display gives ~5K pixels to play around with in After Effects, which is useful for displaying in 3D and zooming in close without losing quality. In After Effects, create a new composition and add a camera². Drag in the footage and toggle on 3d for the footage layer. You can use any combination of keyframing the 3d position and rotation of the footage layer, or the position and point of interest of the camera. The camera also has other properties that can be tweaked and animated over time (via the stopwatch icons), such as focus distance and aperture. This is useful for simulating rack focus adjustments. The rest is trial and error! @videocopilot and @mtmograph are very helpful resources for After Effects. _ 1: 2: 40–50mm will feel close to human perspective, wider fov will make angled perspectives look more fisheye, and narrower more telephoto.

@wesbos There's a good tool in Edge for it: It won't be a problem if your page is small but the performance will scale poorly with the total number of DOM nodes on the page.