-
Notifications
You must be signed in to change notification settings - Fork 46
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
allow TrackTimestampScale in Matroska v4 #437
Conversation
@mbunkus any opinion on that ? It seems libmatroska is not ready for that move but it's possible to work around it without breaking everything. That has a direct impact on support for this feature in VLC as well. I have patches for libmatroska, VLC and libavformat (read and write). So it's at least possible to use it (with sample "accurate" timestamps, see #425). It should not be used by default because most existing readers, if not all, will not handle the value properly. But it can gradually be added. After all it's something that should have been supported since the beginning. And it is there in the specs up to version 3. |
For the record, the feature is now available in libavformat for reading. For writing it depends on codec packing size(s) so we need a list of known packing sizes (in my test to generate files I used 8 by default which should give long enough Clusters in most cases). |
8f1c72c
to
a7054ea
Compare
This is necessary but not sufficient to do sample accurate timestamps. For that we also need proper #439 documentation. |
a7054ea
to
a12ce71
Compare
Following the investigation in #422 it seems like a good tool to achieve sample accurate timestamp in many (more) cases.
Either it should not be used and we already have the default formula Or it needs to be explained in #437.
a12ce71
to
9b4e66c
Compare
Turned into a draft until #521 is merged. This now contains the formula to use |
9b4e66c
to
7327ac1
Compare
Either it should not be used and we already have the default formula Or it needs to be explained in #437.
The consensus on #422 is that reusing the element with some extra trick will likely not bring real world support. So it should be done with new elements instead. Therefore I'm closing this, but parts could be reused to describe properly these elements. |
Either it should not be used and we already have the default formula Or it needs to be explained in #437.
Either it should not be used and we already have the default formula Or it needs to be explained in #437.
Following the investigation in #422 it seems like a good tool to achieve sample accurate timestamp in many (more) cases.
In posts starting from #422 (comment)