Mining the science content of modern, complex astronomy data sets requires the development of new approaches to visualization, and accordingly many groups have begun to investigate new visualization tool and technologies. Brian Kent of the National Radio Astronomy Observatory (NRAO) has shown how the free application Blender is a powerful 3-D visualization tool for astronomy data. The paper has been published in the Publications of the Astronomical Society of the Pacific (PASP), but you can download a copy from astro-ph or you read Brian’s summary post on Astrobetter.
Blender is a free desktop application, maintained by the Blender Foundation under an Open Source license, that realizes high-resolution production quality 3D graphics, modeling, and animation. It runs on Mac OS X, Linux and Windows (32 and 64 bit). The application GUI has the same look and feel and operation on all platforms, and does not have dependencies on OS specific development toolkits such as Cocoa. The core code itself is written in C and C++. Blender has a large user base already (3.4 million downloads per year) and it has found applicability in many fields such as gaming and video editing, and – in the sciences – in medicine, fluid dynamics and other fields.
The application has a number of features that are valuable to astronomy. The graphical user interface (GUI) has a Python application program interface (API) for scripting, which enables data loading through the standard PYFITS and CFITSIO libraries. Moreover, there is support for JPEG, PNG, TIFF and GIF files, and for mesh models, as well as support for a variety of output files, which may be high resolutions stills or video. Table 1 from the paper summarizes the data formats supported:
Credit: Brian Kent, Astronomical Society of the Pacific.
Scalability is an important issue in the development of Graphics applications , and Blender supports study of images on parallel processors or a Graphical Processing Unit (GPU). The paper describes performance and benchmarking numbers. Finally, Blender offers a wider range of graphics features than are generally offered even in commercial packages, as summarized in Table 3 of the paper:
Credit: Brian Kent, Astronomical Society of the Pacific.
The GUI is designed to follow typical workflow steps in creating and examining 3D images. I won’t go into them here – Brian provides a detailed description and you can see it in action in this companion video:
(Credit: Dr. Brian R. Kent, National Radio Astronomy Observatory – Visualizing Astronomical Data with Blender)
The paper very useful examples from several areas of astronomy, and Brian’s webpage provides and tutorials for reproducing them; see also the images at the end of the post:
- Mapping and animating a fly-through of the Extragalactic Distance Database (EDD) of nearby galaxies.
- An N-body simulation of a spiral galaxy (10,000 disk particles, 20,000 halo particles)
- 3D asteroid models bases on the Database for Asteroid Models from Inversion Techniques.
Three-dimensional view of a nearby galaxy catalog (cz⊙ < 3000 km s−1) from the Extragalactic Distance Database (EDD). (a, b, c) Front, right, and top orthographic projections, respectively. (d) Single render frame from the animation in a perspective projection. Credit: Brian Kent, Astronomical Society of the Pacific.
Three-dimensional view of a simulation with colliding galaxies. (a, b, c) Front, right, and top orthographic projections, respectively. (d) Single render frame from the animation in a perspective projection. Credit: Brian Kent, Astronomical Society of the Pacific.
Asteroid models in this figure show how OBJ files can be rendered in Blender. The models are not shown to relative scale and have been increased in size for detail. Credit: Brian Kent, Astronomical Society of the Pacific.
Astronomy Computing Today 