Install

This page covers the installation of the MotionClouds library.

This library depends on a number of dependencies, among which numpy is the only that is mandatory.

There is a dedicated page describing instructions oriented toward psychophysicists

MotionClouds: installing and running the python scripts¶

MotionClouds are parameterized stimuli with controlled motion content and no spatial coherence. They are precisely tuned in the frequency space.

• A full description can be found in Paula S. Leon, Ivo Vanzetta, Guillaume S. Masson, Laurent U. Perrinet. Motion Clouds: Model-based stimulus synthesis of natural-like random textures for the study of motion perception, URL . Journal of Neurophysiology, 107(11):3217--3226, 2012, see http://www.ncbi.nlm.nih.gov/pubmed/22423003 or http://jn.physiology.org/content/107/11/3217 .

download¶

• to get the latest version, download the current state from the repository:

  git clone https://github.com/NeuralEnsemble/MotionClouds.git

installing and running the python scripts¶

This package consists of:

• MotionClouds.py : the API with all the math and display routines,
• test_*.py : different types of stimuli are tested,
• experiment_*py : different experiments.
• fig_*py : different scripts used to generate wiki pages.
• figures : resulting figures of tests and experiments (as set in the variable MotionClouds.figpath.
• Makefile : a GNUmake file to edit files make edit, generate figures make figures or compile the documentation make doc.

Installation¶

• MotionClouds consist of a one-file python script: Installation consists simply in downloading the MotionClouds.py file which contains all routines to compute and use the MotionClouds textures. This script uses python (http://python.org) which comes (pre-installed / easy to download and install) on many operating systems.

• MotionClouds may easily be installed using

    pip install MotionClouds

• Other solutions consist to simply grab this file at: https://raw.githubusercontent.com/NeuralEnsemble/MotionClouds/master/MotionClouds.py and to use it along with the examples.

• Installation has some dependencies:

  1. mandatory: numpy is the core library used to compute textures,
  2. optional: vispy is used to visualize envelopes,
  3. optional: ffmpeg is used to generate movies.
  4. optional: matplotlib, scipy (with PIL support) and imagemagick are used to generate figures in the documentation.
  5. optional: progressbar for displaying progress of encoding.
  6. optional: texlive-latex-recommended latexmk latexdiff to compile the documentation
  7. optional: ipython, jupyter-notebook, pymovie to test MotionClouds within an ipython notebook (recommended)

• These dependencies may be easily met using the dedicated scripts:

  • For MacOsX, first install command line tools by typing the git command in the terminal (a pop-up message should happen if these tools are not installed). Then I recommend using a python toolchain such as that described in https://github.com/laurentperrinet/config-scripts/blob/master/osx_brew_python.sh

  • For Debian-like systems, use the package manager to install missing dependencies (most likely, everything is already installed)

Again, these line of code may be copy-and-pasted in your terminal.

testing¶

• To test the basic functionality of Motion Clouds, try in the terminal:

  python -c'import MotionClouds as mc; fx, fy, ft = mc.get_grids(mc.N_X, mc.N_Y, mc.N_frame); z = mc.envelope_gabor(fx, fy, ft)'

• other sources exist to get a stable release, such as ModelDb (http://senselab.med.yale.edu/modeldb/ShowModel.asp?model=146953).

• This was tested on several MacOsX/Mavericks 64bit machines and on ubuntu 14.04 LTS and Debian squeezy 64bit machines.

• If that fails, we will be happy to help out, so send an e-mail to laurent.perrinet@univ-amu.fr.