2024-02-13 Sequencer meeting notes

Meetings
1

Attendees

  • Aras Pranckevičius
  • Francesco Siddi
  • Richard Antalik
  • Sebastian Parborg
  • Sergey Sharybin
  • Thomas Dinges

Notes

  • Filtering changes are completed #116980

    • There is still an open topic about the setting for the preview. The meeting decided not to pre-emotively add such an option: the performance in projects used by benchmarking is as good as gets with the rest of the changes, and
      adding an option always adds complexity and confusion to the setups for directors.

  • The scaling re-usage context patch #118130 is ready to land.

    Further improvements are possible (maybe re-use FFmpeg context for movies), but the change is good and isolated.

  • The FFmpeg contexts needs to be covered by a cache limiting mechanism, allowing to have multiple of them open at the same time.

    No actual development time has been allocated to it yet, unfortunately.

  • There is an on-going investigation of improving early outputs for cases when blending mode does not allow early output, but the underlying media is fully opaque.

    There is WIP code, but we need to extend the regression suit first. Aras is looking into it.

  • Richard is looking into pitch-preserving retiming. Sergey is available to help with the math side of it.

  • Delayed audio initialization is awaiting review from Audaspace.

  • Changes to the available codes:

    • The meeting agreed to expand the list of officially supported output codes, for example h265. Support from platform maintainers might be needed to ensure FFmpeg is built with them. Thomas is helping with this.
    • The meeting talked about removing option to output legacy codecs (DV, Xvid, …):
      • Agreement is that we should not have presets for the legacy codecs and videos.
      • Jeep legacy codecs unless they cause maintenance burden.
      • The UI can be improved to make things more clear, which will also allow exposing more parameters for people with FFmpeg proficiency.

  • The meeting talked about AVI Jpeg output:

    • It does have maintenance burden.
    • Does not have much added functionality compared to FFmpeg output.
    • The proposal is to possibly make FFmpeg output similar video, remove AVI JPEG, and rename “FFmpeg Video” output to “Video”.
    • No development time is allocated to it yet.

Practical Info

This is a forthlightly video chat meeting for planning and discussion of VFX & Video in Blender. Any contributor (developer, UI/UX designer, writer, …) working on these areas in Blender is welcome to join and add proposed items to the agenda.

For users and other interested parties, we ask to read the meeting notes instead so that the meeting can remain focused.

11 Likes
2

yay been waiting for this

3

Please include DNxHR / HQ options. There are currently no modern industry-standard professional video output options available. Thank you

1 Like
4

I believe we’ve discussed DNxHR outside of the meeting in the Sequencer module chat room. So it is definitely on the table!