Tools for various operations on raw YCbCr video files.http://en.wikipedia.org/wiki/YCbCr

I would like to compare a large number of videos to a good reference video in order to find videos with poor quality. I want to do this in a batch mode by calling a command line utility, I'm already using ffmpeg in this manner to grab video frames.

ycbcr.py - is the main class that supports the following formats:

• IYUV
• UYVY
• YV12
• NV12
• YVYU
• YUY2
• 422

Supported operations:

• basic info about a file
• convert between any of the formats above (including correct sub- re-sampling of chroma-data)
• split a file into individual frames
• creates a diff between two files
• PSNR calculations, one value per color-plane including weighted and average
• get luma-data per frame
• SSIM calculation on luma
• convert between 8bpp and 10bpp
• flip left/right, upside/down
• draw frame number in luma-data
• crop
• visualization of PSNR/SSIM using matplotlib
• reduce framerate by throwing away frames

Xp c2008 thermal receipt printer drivers for mac. Also includes a simple GUI using wxpython that wrapsthe psnr/ssim operations above.

Usage

Files

• ycbcr.py - main class
• plot_diff.py - matplotlib wrapper around PSNR/SSIM-calculation. Generate nice plots using luma-data.
• verify.py - unittest
• visual.py - matplotlib wrapper around PSNR/SSIM-calculation. Generate nice plots.

Screenshots

Here's one of the output from visual.py Jersey number font cricut.