If we take an example of photograph or a video recording. I think for each we would have an RDFSource describing it as creative work. It may include information about who took photo / video, geolocaiton, what it depicts, who appears in the creative work (aka 'tagged in' social feature). Also if someone would like or comment, the would reference that 'crative work' which I would consider RDFSource.

At the same time, each one of those creative works would have multiple encodings. For photos, it could have image/jpeg version or image/png. Also having Responsive images feature, we can expect even more encodings of the same image. Similar for videos we can expect multiple encodings of the same creative work with different codecs and qualities and frame rates (720p60 etc.)

I'd like to clearly define difference between the conceptual creative work and multiple encodings of it each denoted by distinct URI. I think we also should notice that in some cases we want to have meta data for each of the encodings with information like codec used, encoding software which produced it etc.

EDIT: See also:

• https://developer.mozilla.org/en-US/docs/Web/HTML/Element/source