Re: [Cellar] QuickTime timecode tracks in Matroska, was Ancillary data in Matroska

[Jerome Martinez <[e-mail]>]

Le 14/03/2017 à 20:14, Dave Rice a écrit :
> [...]
>
> After looking at this some more, I decided to create an alternate 
> proposal for a T_QUICKTIME codec mapping, rather than to store 
> QuickTime-style timecode in the more vague concept of an ancillary 
> N_QUICKTIME.

Trying to find some counter-arguments, I have to agree that this may be 
better as we don't have any real example for usage of a generic 
QuickTime ancillary track.

>
> [...]
>
> The attachment contains a SimpleBlock using the T_QUICKTIME with a 
> example value of 0x0001fc8c, which according to the QuickTime Timecode 
> Sample Data is 01:12:23;28 and subsequent timecode values for the 
> track would be calculated according to the incrementation defined in 
> the Private Data.

Content location is another issue.
Time code may be stripped (only the first time code is stored and other 
time code must be computed from it + header metadata), and there is no 
way to know it in Matroska : QuickTime has the list of blocks in the 
track header so the player knows if the the time code is stripped (1 
block of 4 bytes) or not (more than 1 block), this is not the  case with 
Matroska.
How could we indicate to the player that the time code is stripped 
(=only one value stored, then player should compute value of other time 
code frames)?

Jérôme