status |
---|
accepted |
The timerange of the media samples represented by the Flow Segment can either be represented as a timerange (currently range
, but will be renamed to timerange
) or a combination of first_ts
, last_ts
and exclusive_last_ts
.
The original design used a first_ts
and last_ts
because the timeranges were required to be inclusive to support Flows with or without a media rate using the same representation.
The internal implementation also used the first_ts
as a sort key.
The design was later changed to provide a choice between representations and to allow Flows with a media rate to use an exclusive timerange end. A timerange with an exclusive end makes it easier to check for containment of a sample within the timerange given that audio and video samples for example have a duration.
- Option 1: Allow either representation to be used
- Option 2: Only support the timerange representation
Chosen option: Option 2: Only support the timerange representation, because it simplifies the API by removing duplication.
Implemented by #27
- Bad, because 2 representations for the timerange needs to be supported in the Flow Segment
- Good, because only 1 representation for the timerange needs to be supported in the Flow Segment