masdude / moviepy

Script-based movie editing with python

MoviePy

MoviePy is a Python module for script-based movie editing, which enables basic operations (cuts, concatenations, title insertions) to be done in a few lines. It can also be used for advanced compositing.

See the full documentation and demo videos here, or have a look at this page of animated GIFs made with MoviePy.

A typical MoviePy script looks like that:

from moviepy.editor import *

# Load myHolidays.mp4 and select the subclip 00:00:50 - 00:00:60
clip = VideoFileClip("myHolidays.mp4").subclip(50,60)

# Generate a text clip (many options available ! )
txt_clip = TextClip("My Holidays 2013",fontsize=70,color='white')
txt_clip = txt_clip.set_pos('center').set_duration(10)

# Overlay the text clip above the first clip
final_clip = CompositeVideoClip([clip, txt_clip])

# write the result to a file in any format
final_clip.to_videofile("myHolidays_edited.avi",fps=25, codec='mpeg4')

Download and Installation

Dependencies

MoviePy cannot run without these dependencies:

• The software ffmpeg is needed for writing, reading, converting the sound and the video.
• Numpy is needed for image and sound manipulation.
• The Decorator module is used in the MoviePy code for better code readability.

Make sure to use a recent version of ffmpeg. You can either install it or save the binary files in any folder and specify the path to these folders before installing MoviePy (for the latter, see Manual installation below).

Normally Numpy and Decorator will be automatically installed when you install MoviePy. In case of doubt/problem they can also be installed manually. Numpy can be installed with most software managers on Linux distributions. Both Numpy and Decorator can be installed as follows with pip:

(sudo) pip install decorator
(sudo) pip install numpy

Installation with PIP

On Linux, if you have pip installed, just type this in a terminal

sudo pip install moviepy

Maybe it works with easy_install too, but it hasn't been fully tested yet.

Manual installation

You can install moviepy manually by downloading the sources, either on PYPI or (if you want the development version) on Github .

Then just unzip everything in one folder, open a terminal and type

sudo python setup.py install

Before doing that, you should make sure that MoviePy can locate ffmpeg on your computer. To do that, run the script moviepy/conf.py. It it fails, then you must enter the path in the first line of this file

FFMPEG_BINARY = path/to/your/ffmpeg

(Not so) Optional dependencies

You are not obliged to install these but for many uses MoviePy will scream at you and say that the package or the software is missing. All these dependencies can be installed any time after the installation of MoviePy.

• PyGame is needed for video and sound previews (really essential for advanced editing).
• imageMagick is needed for all text generation, GIF support, and much more in the future.

There are many packages for image manipulation/processing in python. Most effects are coded such that none of these packages are needed, or such that having at least one of these packages is sufficient. For instance, the feature clip.resize will be available if you have either Scikit Image or the PIL or OpenCV installed (PIL or OpenCV are to be prefered).

• Scipy is needed for many advanced functionalities (tracking, segmenting, etc.)
• Scikit Image may be needed for some advanced image manipulation.
• The Python Imaging Library can be used for resizing videos.
• OpenCV 2.4.6 (which provides the python package cv2) or more recent may be needed for some advanced image manipulation. See below for the installation of OpenCV.

If you are on linux, these will surely be in your repos.

Installing OpenCV 2.4.6

OpenCV is very optional, its installation is not always simple and I found it to be unstable, be warned. The installation seems easy for Windows. On linux, here is what I found on the Internet:

• Remove any other version of OpenCV if you installed it through a package manager.

• Unzip the source code of OpenCV 2.4.6 in some folder. open a terminal in this folder.

• Make a new directory and go into this directory:

  mkdir release
  cd release
  

• Run cmake. Here is the line I used:

  cmake -D WITH_TBB=ON -D BUILD_NEW_PYTHON_SUPPORT=ON -D WITH_V4L=OFF -D INSTALL_C_EXAMPLES=ON -D INSTALL_PYTHON_EXAMPLES=ON -D BUILD_EXAMPLES=ON ..
  

• Run make. This may take a few minutes (15 minutes on my computer).

  make
  

• Finally, install.

  sudo make install
  

And voilà !

You can check if it worked by opeing a Python console and typing

import cv2
print cv2.__version__

Advice: do not throw your release folder away. If later you have strange bugs with OpenCV involving .so files, just redo the sudo make install step.