This module was intended for a Python-equivalent of the Pixylator (ImageJ plugin). In fact, now that the module becomes a more general toolkit, you can in principle perform any types of frame-by-frame conversion of videos.

The current version is 1.0.

Note that I do not provide any test code here! This is just out of my laziness... Please poke me in case of any weird behavior of the code!

Rough documentation is here.

• Can take number of files as input (processed/output separately).
• Frame-by-frame conversion is applied (see below for commands) to each of the video files.

• projection: equivalent to ImageJ's "Z-projection" command.
• profile: equivalent to ImageJ's "Z-profile" command.
• pixylation: equivalent to the "Pixylator" plugin.

videobatch requires numpy, scipy, matplotlib (weirdly enough) and scikit-video.

They will be automatically installed with videobatch if you use the pip command:

pip install git+https://github.com/gwappa/python-videobatch.git

The following scikit-video library uses ffmpeg-related commands to read/write videos. What you need to do before running videobatch is:

1. Install ffmpeg; make sure that you install sub-commands, particularly ffprobe.
   • For Macs and Linux computers, it may be easier that you install through a package manager such as apt or Homebrew.
2. Make sure that your terminal emulator can find ffprobe by modifying the environment variables, if necessary.

Actually, ffprobe is the command that scikit-video uses to find out the actual ffmpeg binary to use.

It looks like my use of FFmpegWriter is inappropriate for some video players (incl. QuickTime Player on the Mac). If you encounter this problem, try using VLC. Note that I don't plan to fix this issue for the time being!

The MIT License