Due to the increased complexity and number of variables in 360 video, there is a need for metadata standards to accommodate this. Google has a suite of standards and metadata injection tools on GitHub under the title of "spatial media" here.
In order to upload 360 video to YouTube and have it recognised and played back properly, it is required that spatial metadata be injected using the Google tool. This supports MP4 and WebM (video only) containers. You can find specification information on GitHub for the support elements for video and audio.
It supports:
Projection type
Stereoscopic mode
Projection pose (yaw, pitch, roll) - presumably for initiation position?
[Add audio]
Spatial Media Metadata Injector is a open source software that can be used to "convert standard" videos (equirectangular for example) to inject 360° / stereoscopic 3D (top/bottom) and spatial audio (ambix / sn3df) datas
360° video can be either monoscopic or stereoscopic. Monoscopic video supports what is perceived as a 2D representation of the scene i.e. there is no perception of depth. Stereoscopic video supports a 3D representation of the scene with a perception of depth.
Monoscopic 360° video contains video captured from a single viewpoint within the scene.
Stereoscopic 360° video contains video captured from two viewpoints within the scene. These viewpoints can be different point of view for each eye. These are packed into a single video file where they can be arranged side-by-side or top-bottom.
Codecs: h.264/AVC, h.265/HEVC, VP9
Wrappers: MP4, Mastroska, WebM
Significant variables in choice of video file format:
Achievable bitrate / compression
Metadata container options
...?
Projection format refers to the way in which data representing a 360° or spherical field of view is mapped to a flat image when it encoded. It is similar to the way in which a map of Earth is a flat representation of the spherical surface of the planet.
Some common projection formats include:
Equirectangular
Cubemap
Equi-angular cubemap
Pyramid
Significant variables in choice of projection format:
Pixel density
Tool support (encoding, decoding)
Requirements of video streaming platforms (e.g. YouTube)
...?
360-degree has specialised transcoding requirements if properties such as projection format and stereoscopy to be properly managed. At this point the tools listed here have not been tested by us and inclusion should not be taken as a recommendation.
List of tools:
Headjack VRencoder, https://headjack.io/vrencoder/
360° video is a video format in which every direction of view is available to the viewer. Though the direction of view is free, in most circumstances the viewing position in space is either fixed or on a predetermined “on rails” path.
360° video can be created in a number of different ways:
Captured by a camera or array of camera lenses;
Generated as an export from 3D rendering software (e.g. Blender);
Generated from a real-time 3D game engine. The 360° video exported from a game engine can be the artistic end product (s. for instance "Passage Park #7: Relocate" by Studer / van den Berg) or the documentation of a real-time 3D artwork.