This section contains technical information for preparation of the movie submission. As the target display for all submissions are dome theatres and planetarium domes, video have to be prepared in a 180 degrees angular fisheye projection (http://paulbourke.net/dome/fisheye/).  The following two images show examples of this rendering technique.

Submissions to this contest have to be submitted using this projection techniques with a resolution of either 4096×4096 or 8192×8192.  The individual frames should be combined into an MP4 file using either H.265/HEVC or H.264/AVC encoding.  Optional audio tracks shall be submitted as separate files and also be included in this MP4 file.  Submissions are free to use either 2.1, 5.1, or 7.1 audio, which means the submission consists of an MP4 video files and either 3, 6, or 8 individual audio files, one per channel.

The remainder of this section is split into three parts, which follow the production pipeline for creating the prerendered visual components.  Additional information can be found on the accompanying GitLab page (https://gitlab.liu.se/wisdome-contest-2020) and a tutorial page for an in-house developed fisheye stitching tools (http://sgct.itn.liu.se/tutorial/offline).

If the software you are using the create the content for the contest has the ability to generate fisheye images directly, feel free to skip this part of the instructions.

The high-level overview of rendering a fisheye method consists of rendering at 4 separate orthogonal viewpoints.  These viewpoints then get merged into a cubemap, which in a second step gets reprojected into the fisheye representation.  These two steps are already implemented in the Stitcher application, which is provided below.  The following images shows these 4 orthogonal viewpoints as “left”, “right”, “top”, “bottom”.  Please note that in the case of this submission (180° fisheye), only half of the “top” and “bottom” are used in the final result.

Whichever software you are using for creating the submission content has to be able to render to a square field-of-view and be able to output those renderings to individual files, such that in the end you have four image sequences of the form:
basefilename_top_00000.png  to  basefilename_top_36000.png
basefilename_bottom_00000.png  to  basefilename_bottom_36000.png
basefilename_left_00000.png  to  basefilename_left_36000.png
basefilename_right_00000.png  to  basefilename_right_36000.png
With 36000 frames in this case for 10 minute video @ 60 fps (10 [min] * 60 [sec/min] * 60 [frames/sec]).

The stitcher application then coverts these into 180° fisheye images.

These files are of the form:
basefilename_00000.png to basefilename_36000.png

The Stitcher application to transform cubemap images into fisheye projection is available from (http://sgct.itn.liu.se/uploads/stitcher.zip).  The application is only provided as a pre-build binary for Windows, but is available as source code on GitHub (https://github.com/opensgct/sgct).  The commandline version of the application should be compilable for both MacOS and Linux, the UI is build in C# and probably requires Mono, but is very much untested on non-Windows platforms.

The easiest way to convert an image sequence of the form basefilename_00000.png to basefilename_36000.png into a video is to use the open-source solution FFmpeg (https://www.ffmpeg.org), which is a commandline tool that can be used to convert image sequences into a video format and add multiple audio tracks at the same time.  The available options in FFmpeg are too numerous to elaborate, so this guide is not meant as a comprehensive explanation (see the FFmpeg documentation (https://ffmpeg.org/ffmpeg.html) for complete information.

Generating a video from an image sequence

ffmpeg -framerate 60 -i basefilename_%05d.png -i audio.mp3 -c:v libx265 -c:a copy  basefilename.mp4

Explanation:
ffmpeg:  Name of the commandline tool installed from ffmpeg.org
-framerate 60:   Uses 60 of the provided images per second of video runtime
-i basefilename_%05d.png:  Search for files basefilename_00000.png to basefilename_99999.png as frame input for the visual part
-i audio.mp3:  Use the file audio.mp3 as the input for the sound part of the video
-c:v libx265:  use the H.265 video encoding format (alternatively: use libx264 if the size of images is 4096×4096)
-c:a copy: Do not reencode the audio stream but copy instead
basefilename.mp4: The output file that is to be submitted for the contest

• Alexander Bock, Linköping University & University of Utah
• Erik Sundén, Linköping University

• Ingrid Hotz, Linköping University
• Patric Ljung, Linköping University
• Niklas Rönnberg, Linköping University

• Thomas Clifford, Projectleader, Visualizationcenter C
• Konstantin Economou, Lecturer in Media & Communication Studies, Linköping University
• Gabriel Eilertsen, Postdoctorial Researcher, Linköping University
• Carter Emmart, Director for Astrovisualization, American Museum of Natural History
• Jackie Faherty, Senior Scientist, American Museum of Natural History
• Mats Falck, External Relations Coordinator, Umeå University
• Chuck Hansen, Professor of Scientific Visualization, University of Utah
• Ralph Heinsohn, Designer, Curator & Producer of Immersive Media, Hamburg
• Tim Florian Horn Director of the Zeiss-Großplanetarium and Archenhold-Sternwarte
• Ralph Kaehler, Manager of the Visualization Facilities at the Kavli Institute for Particle Astrophysics and Cosmology
• Malcolm Lindsay, Composer, malcolmlindsay.com
• Konrad Polthier, Professor of Mathematical Geometry Processing, Freie Universität Berlin
• Ben Shedd, Director, Producer, Writer, and Professor at the School of Art, Design and Media, Nanyang Technological University
• Mark SubbaRao, President of the International Planetarium Society, Director of Visualization, Adler Planetarium
• Ryan Wyatt, Senior Director of Science Visualization, California Academy of Sciences