getting-started(7)						  Numm Tutorials						getting-started(7)

NAME

       getting started - numm installation and usage

SYNOPSIS

       numm-run FILE
       import numm

DESCRIPTION

       numm is a python library -- the bastard child of Processing.org and MATLAB, if you will -- that fuses python, numpy, and gstreamer together
       to create a numerical-computing environment for multimedia generation, analysis, and interaction.

   Installation
       A Debian package is provided for ease of installation on Debian and derivative operating systems, such as Ubuntu; we recommend you  abandon
       alternative systems and install Debian, but in the interim direct your attention to  virtualization software such as VirtualBox.

       The following command downloads and installs numm:
       % wget http://numm.org/numm/releases/python-numm_0.1-1_all.deb
       % sudo dpkg -i python-numm_0.1-1_all.deb

   Usage
       At its simplest, numm provides three pairs of invertable functions that connect common media formats with numpy:

       Images represented as (height, width, color) numpy.uint8 arrays.

       o   image2np(path) -> np

       o   np2image(np, path)

       Sounds as (frames, channels) numpy.int16s

       o   sound2np(path) -> np

       o   np2sound(np, path)	      # XXX: must be .wav

       Videos as (frames, height, width, color) numpy.uint8s

       o   video2np(path) -> np

       o   np2video(np, path)	      # XXX: must be .mkv

       For  rapid-prototyping  of  interactive	numpy-based  audio-visual  experiments,  the  numm package has a real-time mode.  The numm package
       installs a program named numm-run, which launches (and reloads on modification) python scripts that may implement any subset of the follow-
       ing functions, which are asynchronously called as needed:

       o   video_in(a)			    # webcam

       o   video_out(a) 		    # mutate a in-place to set video

       o   audio_in(a)			    # mic

       o   audio_out(a) 		    # mutate a for audio output

       o   mouse_in(type, px, py, button)

       o   keyboard_in(type, key)

SEE ALSO

       numm-run(1), numm.one-bit-instrument(7), numm.spectral-analysis(7)

numm								   February 2012						getting-started(7)