Re: OpenTimelineIO meeting agenda 2019-09-05
|
Joshua Minor
We're looking for AAF expertise in these areas: 1. Our understanding of how to find the timecode associated with a MasterMob has some holes. We have encountered problems like this: https://github.com/PixarAnimationStudios/OpenTimelineIO/issues/523 and slowly improved based on test cases, but we are not confident in our understanding of how this is supposed to work. Specifically, we use pyaaf2's walk() method and this logic: https://github.com/PixarAnimationStudios/OpenTimelineIO/blob/master/contrib/opentimelineio_contrib/adapters/advanced_authoring_format.py#L149 to find which segment of media is used in a clip. Most of our confusion revolves around the relationship between a MasterMob, FileSourceMob, PhysicalSourceMob and the various Timecode objects connected to them.2. Similarly, we have some misunderstandings about how nested clips work, which leads to problems like this: https://github.com/PixarAnimationStudios/OpenTimelineIO/issues/481 We adopted a two-stage approach to parsing an AAF. First we transcribe the AAF structure verbatim into OTIO, then as a second pass, we simplify the structure into something that more closely resembles what the user sees in the UI of their NLE. This has been reasonably successful, but produces unexpected and undesirable results sometimes. We would rather have a clear understanding of the mapping between the AAF structure and the user's data model, so we can go to our simpler OTIO model in one step. 3. Non-linear timing effects. We have read the AAF Object Specification documentation on OperationDefinitions, and the AAF Edit Protocol documentation on Video Speed Control effects, but neither of these documents have details on the actual Motion Control effects that we find in AAFs exported from Media Composer. Specifically, we are trying to deeply understand the PARAM_SPEED_OFFSET_MAP_U, PARAM_SPEED_MAP_U, and PARAM_OFFSET_MAP_U parameters which seem to hold the critical information about non-linear time warps. Any advice or expertise on any of these would be tremendously helpful.
|
|
|