Open HansBusch opened 3 weeks ago
Cautionary Suggestion
In any case, since this requirement is related to Profile S/T features, devices (if interpreted/implemented incorrectly) would need time to adapt so that Profile S/T conformance is not affected?
I think option 2 is more appropriate, option 1 would create confusion/mismatch, since for video source we already mentioned below note. "Note that in case of e.g. a 90 degree rotation the width parameter corresponds to the height of the Video and vice versa."
I would also prefer option 2, since its been our assumption that VEC follows the same requirement as the VSC. If its not part of DTT testing, I would suggest to have it added once we have clarified the expected behavior.
Option 2 would mean that we need to change the definition of the following two fields:
Resolution [VideoResolution2] Configured video resolution
Width [int] Number of the columns of the Video image.
Height [int] Number of the lines of the Video image.
The Media2 specification clearly describes how width and height parameters must be interpreted for the
VideoSourceConfiguration
when a rotation of 90 or 270 degree is configured. Such a clarification seems to be missing for theVideoEncoderConfiguration
.Two possible options for the
Resolution
field of theVideoSourceConfiguration
:Width and height are swapped so that they correspond to the streamed video.
Width and height correspond to the
Bounds
field of theVideoSourceConfiguration
. As a consequenceWidth
describes the height of the encoded video and vice versa.