Skip to content

Latest commit

 

History

History
50 lines (33 loc) · 1.42 KB

README.md

File metadata and controls

50 lines (33 loc) · 1.42 KB

A rudimentary video codec

https://video-codec-v0.onrender.com/

Constraints

  1. The video size should end up smaller after the encoding
  2. The encoding is reversible
  3. The compressed video should contain enough information to be recongisable

technique

Consecutive frames have spatial redundancy, pixels in consecutive frames are likely to have similar information. This allows us to compress like how png does.

Each frame contains pixels -which contain three attribuutes: r, g, and b. This gives a total size of (width * height * 3) per frame.

We convert each frame to YUV420 format. Then for each 4 adjacent pixels, we average the chominance this reducing information stored by each pixels to 0.25

on the webpage

https://video-codec-v0.onrender.com/

  1. Use the video.rgb24 file available in this repository
  2. upload the video, add its width 384 and height 216
  3. the video should be compressed and downloaded to your computer
  4. upload the compressed video, width 384, height 216 and number of frames 217
  5. the video should be decompressed and downloaded o your computer

play the videos using ffmpeg

type the command onto the terminal: ffplay -f rawvideo -pixel_format rgb24 -video_size 384x216 -framerate 25 <filename>.rgb24

references

  1. https://github.com/kevmo314/codec-from-scratch
  2. ffmpeg blogs