Closed sandersaares closed 4 years ago
Comments from F2F discussion:
I think we just need to tidy up the text to be tolerant of edit lists, without giving them any special treatement (they are just extra low level math to arrive at the same numbers we already work with). Refer to CMAF for details.
Addressed in timing model, closing.
There was some discussion of edit lists in https://github.com/Dash-Industry-Forum/DASH-IF-IOP/issues/168
While the referenced thread implies to me that edit lists are legal in certain situations and that this got clarified, I still find this confusing statement in IOP v4.3 that, in my mind, conflicts with this:
Even beyond that, I see BMFF mention edit lists in context of the segment index box, whereas the new 3.10 section in IOP 4.3 does not, so I am again confused. How exactly does an edit list interact with DASH timing?
This statement also confuses me:
What does that even mean. Edit lists are not to be used for synchronization, only for other purposes? If so, what exactly? We need to present a more holistic view of this topic, these individual statements make it very hard to understand any sort of big picture here.
So I raise the topic again - what is the correct use of edit lists? Are they allowed in DASH-IF IOP context? And how do they affect calculations (e.g. relationship with PTO that Kilroy pointed out in the abovementioned thread).
This topic seems insufficiently described to ensure creation of interoperable DASH implementations right now. I do not understand edit lists to any sufficient degree myself, so I request that a more qualified expert make this topic clear to all readers of our IOP document. Alternatively, if we cannot come up with clear guidelines, I propose a ban on edit lists on the grounds of "you can't use what you can't understand".