HIPASS galaxy catalogue (HICAT) animations |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
A series of animations constructed from the HIPASS galaxy catalogue (HICAT) and its northern extension to declination +25° (NHICAT) depicts the three-dimensional distribution of galaxies in the nearby universe (z < 0.03) as seen by a hypothetical observer who travels geometrically through them. |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
One frame of the 720 in nhicat-DVD-111a.flc shown at 85% actual size. True 3-D structure only becomes evident upon animation although clustering is clearly discernible in this 2-D view. |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The celestial sphere is shown in a zenithal perspective projection at unit radius corresponding to a recession velocity of 3000 km/s (z = 0.01, 40 Mpc, 140 Mly). A little over half the catalogue lies within this sphere, thereafter the density falls rapidly to about 12000 km/s (z = 0.04, 170 Mpc, 540 Mly). Galaxies are represented by coloured spheres with volume proportional to M(HI), the neutral hydrogen mass. To enhance the perspective effect, the marker size is corrected for its distance from the observer with appropriate object occlusion. Marker hue represents recession velocity starting at blue for 0 km/s and cycling through cyan, green, yellow, and orange to red at the surface of the celestial sphere (3000 km/s) and beyond:
The Milky Way at the centre of the celestial sphere is located by a fiducial marker with arms that point to the celestial poles and ra = 0, 6, 12, and 18h on the celestial equator; that which points to 0h is coloured yellow. The celestial sphere is inscribed with a 15° equatorial graticule in grey, with IAU constellation boundaries in magenta labelled with their standard three-letter abbreviation. These are useful for orientation and identification of particular structures, particularly when aligned with the fiducial marker. Constellations associated with particular clusters or superclusters are marked in maroon. These are Virgo (Vir), Hydra (Hya), Centaurus (Cen), Fornax (For), Eridanus (Eri), Pavo (Pav), and Indus (Ind). The supergalactic equator is marked on the celestial sphere in yellow with the galactic equator in green. The observer is in the equatorial plane of one of these systems when its equator is projected as a straight line intersecting the central marker. HICAT looks straight through the galactic plane which is not an obstacle for HI observations of distant galaxies. The observer traverses a path that describes one or more figures-of-eight when projected onto the celestial sphere (marked in dark blue). In 3-D this motion is composed of three independent cycles: |
In the "a", "i" and "o" animations the observer's trajectory is circular, whereas in the "s" animations the observer spirals in very close to the centre then out again. The "i" (inner) animations show the region close to the Milky Way, while the "o" (outer) animations skim past the surface of the celestial sphere (at 3000 km/s, z = 0.01) mainly showing its contents. In these animations the observer moves gradually allowing plenty of time to appreciate the 3-D structure. On the other hand, the "a" and "s" (spiral) animations cover a wide range of distances and angles in order to depict both the near- and far-field galaxy distribution. Consequently, the motion in these animations is much more dynamic, particularly in the latter, and they may seem like a roller-coaster ride in parts! Stereoscopy is an obvious extension for representing 3-D structure. This is provided here in two ways; by stereo pairs, which require stereo projection equipment; and a red/cyan anaglyph which requires anaglyphic glasses (red on the left). In each the depth has been set so that the Milky Way at the centre of the celestial sphere is in the focal plane whence galaxies in the foreground appear to fly out of the screen. In earlier animations (produced several years ago) the stereo calculation was based on a simple rotation, the so-called "toe-in" method, which introduces vertical parallax and is not strictly correct. This has now been fixed by using a proper "off-axis" calculation. The 3-D representation is generated by applying a parallax transformation to the galaxy distribution as seen from Earth to simulate what the observer would see in the night sky. That is, the distance and direction to each galaxy as seen from Earth is transformed to the distance and direction as seen by the observer. The distribution on the observer's celestial sphere is then projected onto the map plane using a gnomonic projection; it can be shown that this sequence of operations is equivalent to computing a zenithal perspective projection of the Earth-bound galaxy distribution with the observer at the point of projection. However, the question naturally arises of using something other than the gnomonic projection to map the observer's sky. In particular, the choice of a cylindrical projection would allow the whole sky to be mapped, including the galaxies behind the observer. This is done, mainly out of mathematical interest, in nhicat-XGA-111C using Gall's stereographic projection - which is unrelated to stereoscopy! |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Animations currently available:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Notes on each animation |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
A number of structures are common to all of the animations:
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Instructions |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The animations have been produced in FLC format with reverse play capability when using the xanim player on unix/X11 systems. This player also allows on-the-fly speed adjustment and frame-by-frame stepping in each direction which is very useful for investigating 3-D structure. Suitable players are said to be available for Macs and Windows. Some animations have both left and right stereo pairs, but if you do not have a stereo viewer there is little point in getting both views. Because of their size they have been bzip2'd and then split into 10 MiB files for download. Firstly download the split files. You might find wget useful for this but recursive fetches are disallowed |
by our robots.txt so follow the instructions in the links given in the table. When playing the animations, particularly the larger ones, using xanim on my Dell Latitude D600 laptop (512 MiB, Debian sarge) I find that it is much better not to buffer them to memory; the i/o rate from harddisk (i.e. not network mounted) is more than adequate. (The i/o rate for the 60 GiB Fujitsu MHT2060AT internal disk, as measured by hdparm -tT /dev/hda, was 24 MiB/s with DMA enabled, though it was too slow without DMA.) Thus use xanim +f nhicat-DVD-111a.flcThis should start playing in under 5 seconds, while nhicat-XGA-475a takes about 30 seconds to begin. |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Converting to other animation formats |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The venerable FLC format was found to be the most suitable for these computer animations. While MPEG provides better compression it does so at the expense of image quality, particularly in losing the fine grid lines. Although MPEG image quality may be improved by increasing the bit-rate, the MPEG file size then becomes comparable to the bzip'd FLC files without achieving the same quality. Since FLC animations use lossless compression they may be decomposed into individual frames for conversion to other formats. The unflick utility undoes the operation of ppm2fli which was used to create them. (However, to support more than 4000 frames for the larger animations you will need to increase FLI_MAXFRAMES in upro.h and apro.h before compiling.) Version 2.1 of unflick supports output filtering so, for example, to extract the frames as GIFs rather than the default |
PPM, you could use unflick -f"ppmtogif -sort" -v \ animation.flc animation- gifwhich uses the netpbm filter, ppmtogif. Once you have the separate frames you can recompose the animation in another format. The 720 x 576 image size was chosen to match that of PAL DVD which is based on MPEG2. For example, a unix shell script was used to produce the 8 Mib/s MPEG-1 video nhicat-DVD-111a.mpg using the netpbm utility, ppmtompeg. However, although much more compact, image quality is notably inferior to FLC. MPEG-2 and MPEG-4 video can be produced with mplayer/mencoder, unix shell scripts flctompeg2 and flctompeg4 may be used for this. However, the resulting MPEG-2 and MPEG-4 videos at the same bit rate (8 Mib/s) as the MPEG-1 video are virtually indistinguishable from it. |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Dr. Mark R. Calabretta ( mcalabre@atnf.csiro.au) Last modified: 2006/06/14 |