Renderman Challenge

These color and elevation maps are designed for use in 3D rendering software. They are created from data assembled by the Lunar Reconnaissance Orbiter camera and laser altimeter instrument teams.

Color

The color map was adapted from the Hapke Normalized WAC Mosaic, a composite built by the camera team from over 100,000 WAC (Wide Angle Camera) images. The readme file provides a technical summary of how the mosaic was constructed along with references to more detailed publications. A less formal introduction to the process is in this LROC blog post and in this one. For the color map here, the visualizer modified and combined three of the seven wavelength bands of the LROC color data to more closely match what the human eye sees. The red channel contains the 643 nm band, while green and blue were created from different linear combinations of the 566 and 415 nm bands to more nearly center them on 532 nm (green) and 472 nm (blue). A gamma of 2.8 was applied (the LROC data is linear), and the channels were multiplied by (0.935, 1.005, 1.04) to balance the color. The intensity range (0.16, 0.4) was mapped into the full (0, 255) 8-bit range per channel. Small data dropouts near the top and bottom were inpainted using Photoshop's content-aware spot healing brush. The source data covers the lunar globe from 70°N to 70°S. Because the Moon's axial and orbital tilts are both small, many areas outside these latitudes remain shrouded in shadow, even after thousands of passes by LRO's camera, so they are left out of the LROC mosaic. For this color map, the missing latitudes were filled in with a combination of monochromatic LROC data and an albedo map (LDAM) from LRO's laser altimeter. When rendered with realistic shadows, these parts of the map aren't particularly visible, and while they comprise more than 20% of the map's pixels, they represent only 6% of the Moon's surface. This image is optimized for aesthetics, not science. Scientific applications should use the source data.

Displacement

The displacement map (also known as a height map or elevation map) was taken directly from the latest (as of spring 2019) gridded data products of the Lunar Orbiter Laser Altimeter instrument team. LOLA data is archived on the Geosciences Node of the Planetary Data System. A small subset of the LOLA data stored there, the global cylindrical projections at 4, 16, and 64 pixels per degree, has been reformatted here as uncompressed TIFF files, in vertical units of either floating-point kilometers or 16-bit unsigned integer half-meters. The reference surface for all LRO data is a sphere of radius 1737.4 km. LOLA's gridded elevation data is published as signed 16-bit integers in units of half-meters relative to this radius. For the floating-point TIFFs, the source data was divided by 2000. For the unsigned 16-bit TIFFs, the source data was offset by +20,000 (10 km) so that all of the values are positive. This latter format is provided for software that doesn't work well with either floating-point or signed integer files.

How These Maps Are Used

Within 3D animation software, an object like the Moon begins as a simple geometric shape, in this case a sphere. Texture maps like the ones on this page are used to add detail to the model. The color map tells the software how to paint the surface, and the displacement map tells it how to add the shape details that define the lunar terrain. Without them, the Moon model is just a smooth, monochrome ball. Although these maps are flat rectangles, the software understands them as maps of a spherical surface and knows how to warp them onto spherical geometry. Each pixel in these texture maps corresponds to a point on the lunar surface defined by a longitude-latitude pair. Pixels in the color map contain the base color of the surface, before applying the effects of varying light and camera angles (called incidence angle i and emission angle e in the technical description). Pixels in the displacement map contain the height of the surface at the corresponding locations.

More Data

These maps are a tiny subset of the data publicly archived by the LRO instrument teams. All of the data is in standard and fairly simple PDS file formats (files with .IMG filename extensions). Although not typically supported by general-purpose computer graphics software, these formats are directly readable by some GIS programs. The data formats are documented in PDS labels, which are either separate text files with a .LBL filename extension or embedded as fixed-length headers. Other than the possible header, most .IMG files are uncompressed 2D binary arrays of numbers. Some data is also available in TIFF or JPEG2000 formats. As of this writing, the best global-scale elevation data for equatorial and mid-latitude regions of the Moon, at resolutions up to 512 pixels per degree, is SLDEM2015. For regions near the poles, where the density of laser shots is much higher, the GDR data in polar stereo projection reaches resolutions of five meters per pixel. The LROC narrow-angle camera (NAC) has imaged most of the Moon's surface at a resolution of one meter per pixel or better. LROC imagery can be explored graphically using its Quickmap interface and searched on its RDR Product Select page. Of particular interest are the Topographic Products (menu option in the upper left of the page), which are matched with high-resolution DEMs derived from stereo image pairs. This dataset includes all of the Apollo landing sites and dozens of other locations of interest.

This set of star maps was created by plotting the position, brightness, and color of 1.7 billion stars from the Hipparcos-2, Tycho-2, and Gaia Data Release 2 star catalogs, with help from the Yale Bright Star Catalog, UCAC3, and the XHIP Hipparcos cross-reference. The constellation boundaries are those established by the International Astronomical Union in 1930. The constellation figures also come from the IAU, although they're not official.

---

Skip the explanation

---

The maps are presented in plate carrée projections using either celestial (ICRF/J2000 geocentric right ascension and declination) or galactic coordinates. They are designed for spherical mapping in animation software. The oval shapes near the top and bottom of the star maps are not galaxies. The distortion of the stars in those parts of the map is just an effect of the projection. The celestial coordinate mapping will be the more useful one for 3D animation, since camera rotations in the software will correspond in a straightforward way to the right ascension and declination in astronomy references. The galactic coordinate mapping is probably better for 2D animation and compositing. It also works as a standalone image showing the edge-on view of our home galaxy, from the inside. Update: The galactic images were replaced on January 4, 2021. The original images used ICRF/J2000 coordinates in a galactic coordinate transformation meant for B1950. The new images use the transformation described in the Hipparcos and Gaia documentation. The boundary, figure, and grid images are conventional grayscale TIFF files. The star maps are in OpenEXR's half-float format, which provides higher dynamic range in a linear colorspace while easily accommodating very large files. Most 3D animation and HDR image processing software can read OpenEXR.

Catalog Completeness

Hipparcos-2 provided the data for stars brighter than magnitude 8.0. To check HIP2's completeness, it was compared with the Yale Bright Star Catalog by matching positions, by using the XHIP cross-reference catalog, and by consulting the SIMBAD database. Of the 9096 stars in Yale, about one hundred have no matching Henry Draper ID in HIP2. Most of these are the second members of double or multiple stars that are represented in HIP2 as a single entry, and their omission has no visible effect on the star map. Another few are highly variable stars listed in Yale at their brightest magnitude, even though this magnitude is atypical. An example is T Coronae Borealis, the Blaze Star, which reached the magnitude listed in Yale only in 1866 and 1946; it normally hovers around magnitude 10. The remaining 18 stars can be considered genuinely missing:

HR	Mag	Comment
4210	4.30	Eta Car
4375	4.41	Xi UMa A
4374	4.87	Xi UMa B
5978	4.77	Xi Sco A
5977	5.07	Xi Sco B
4729	4.86	256 Cru, 90" from Acrux (Alp Cru)
2322	5.98
5343	5.98	CN Boo, 17' from Arcturus (Alp Boo)
2950	6.02	12' from Procyon (Alp CMi)
1982	6.15	AK Lep, 97" from Gam Lep
5034	6.18	61" from J Cen
2366	6.20	HIP 31067, omitted from HIP2
4619	6.37	3.6' from Del Cen
1704	6.37	15' from Rigel (Bet Ori)
6660	6.38	part of M7 in Sco
6263	6.45	part of NGC 6231 in Sco
2341	6.51	10' from Canopus (Alp Car)
6848	6.84	part of M24 in Sgr

These were added to the star maps. Some are relatively near very bright stars that may have made the measurement of their dimmer neighbor by Hipparcos problematic. The magnitude of Eta Carina is an estimate based on the AAVSO Light Curve Generator as of July 2020. All other data was taken from Yale. Missing stars at fainter magnitudes are in general much less visually apparent. The completeness of Tycho-2, used for stars with visual magnitudes between 8.0 and 11.5, was assumed. And since data in the Hipparcos and Tycho catalogs share the same provenance, there's less concern about drawing stars twice or losing them in the cracks at the crossover magnitude. Gaia DR2, however, has two obvious data holes centered at r.a. 97.9°, dec. 57.5° and 34.2°, 22.1°. These were filled in using stars from UCAC3.

Star Colors

The colors of the stars from both Hipparcos-2 and Tycho-2 were based on the B-V color index. This requires a mapping from B-V to effective temperature, T_eff. In previous versions of this product, the mapping used a high-degree polynomial fit, but it was found that this fit was calculated using a relatively narrow range of B-V. Outside this range, the fit behaved poorly, producing a number of unrealistically red stars. For this version, the mapping was a slightly modified version of a function ascribed to F. Ballesteros. The spectrum of light emitted by a blackbody with a temperature T_eff can in turn be mapped to an RGB triple, yielding the star color. See Mitchell Charity's What color is a blackbody?, which was used for the present work. In common with Charity, the modest goal here was not-completely-bogus colors. The colors of Gaia DR2 stars were taken from the fields called G, G_RP, G_BP, scaled by color balance factors estimated by eye. See this article from the Gaia team for more information about the passbands for these three measurements. Roughly a quarter of the DR2 stars lack G_RP, G_BP values, and those are set to white.

Tour

The animation demonstrates the use of the maps in a tour of the sky. The tour starts at W-shaped Cassiopeia, then heads south through Perseus to the winter constellation of Orion the Hunter and the Hyades and Pleiades star clusters in Taurus. It moves southeast past Orion's canine companion and its star, Sirius, brightest in the sky, eventually pausing at the rich southern hemisphere portion of the Milky Way in Carina and Crux, the Southern Cross. East of the Cross, in Centaurus, is the binary star Alpha Centauri, at 4.4 light-years the naked-eye star system nearest to the Sun. Also visible as a fuzzy spot near the top of the frame is the globular cluster Omega Centauri. The number of stars used to draw the star maps is large enough to reveal many globular and open star clusters as well as the Large and Small Magellanic Clouds. After passing near the celestial south pole, the tour moves north along the Milky Way to the center of our galaxy near the teapot in Sagittarius. The tour veers northwest from there, finally stopping at the familiar Big Dipper or Plough asterism in Ursa Major. This is an update to entry 3895.

The visualizations here are based on the visualization Monsoons: Wet, Dry, Repeat (#4397). Each data set is presented in three resolutions: 8192x4096, 4096x2048, and 2048x1024. Each the 8192x4096 and 4096x2048 layers have been rendered with alpha transparency channels to allow you to create your own combinations of layered data. NOTE: the preview movies are composited over a black background, but the individual 8192x4096 and 4096x2048 frames have transparency channels. The 2048x1024 frames do not have transparency channels. The layers have frame numbers from 01000 through 13000. Frame 01000 corresponds to 02 Jun 2014 at 00:00 GMT. Each successive frame is 15 minutes later. Frame 13000 corresponds to 05 Oct 2014 at 00:00 GMT. To determine the time for a specific layer's frame number, you can look at the date sequence's corresponding frame. Even though all layers are provided at 15 minute intervals, most data sets do not have such a high cadence. In these cases, the frames simply show the same data. See the annotation with each data set for the cadences.