Skip to content

Latest commit

 

History

History
49 lines (34 loc) · 1.49 KB

ffv1-v4-goals.md

File metadata and controls

49 lines (34 loc) · 1.49 KB
Raw

FFV1 v4 goals and promises

  • Cleanup of a few special cases that resulted out of bugs
  • Better Compression and or Speed
  • BAYER support to efficiently store RAW color CCD images
  • Better error resilience
  • Arbitrary color spaces
  • Limit worst case size after compression

Anything that finds no volunteers will be pushed to v5

Compression and Speed

Proposals are welcome, preferably with a patch to the FFmpeg implementation. Proposals should come with a speed and compression comparison against FFV1 v3

Ideas

  • Use the previous frames with motion estimation / compensation
  • Neural network based compression

Better error resilience

Proposals are welcome, preferably with a patch to the FFmpeg implementation. Proposals should come with a PSNR per overhead and damage comparison of damaged streams

Ideas

  • Include concealment motion vectors for each slice. This way if a slice is damaged the corresponding area from the previous frame can be used even if there is movement, this is very low overhead and quite simple
  • Include a lossy compressed representation of the frame

Arbitrary color spaces

Allow storage of any 2D plane of samples not limited to red, green, blue, alpha or transforms of these.

Examples:

  • Infrared
  • Radar
  • surface vectors
  • height
  • age
  • temperature
  • charge
  • elasticity
  • velocity
  • acceleration

This is mostly a question of metadata, the existing compression should be largely reusable. Of course compression could be fine tuned if any of these uses have non niche uses.