Text and most images © Don Davis

              On Painting and animation techniques then and now

  1. Traditional painting

  The creation of convincing astronomical images, especially those with hardware or other geometry requires knowledge and use of perspective and lighting principals as well as a working knowledge of the subject matter. Next comes mastering the nuances of your medium of choice for your given pictorial goals.
  One reason I gave up oil painting was the hassle of achieving even dark photographable blacks in that medium. Bonestell excelled in this, thanks to the use of oils mixed with a painting medium he preferred ( 2/3 turpentine, 1/3 sun thickened linseed oil, a dash of Cobalt Drier) which if added to paint in sufficient amounts can cause the brushstrokes to melt away before drying. If time is not that important a few layers of such 'black' applied in nearly liquid state can end up as a wonderful enameled like surface which nowhere deviates from the horizontal by more than a few degrees. This can be lit properly to show off the rich blacks against the sunlit surfaces below. Smooth board panels or smoothly sanded gessoed panels are the preferred painting surface with this method. I have only done one oil on canvas painting with the sky done this way, as many sanded gesso coats were required before the sky could be even started.
  Acrylics freed me of the black sky hassles ( I still consider oils the ultimate traditional media!) but only with the computer can I be sure the skies will indeed look black in the final printed page.
Many variations of what constitutes 'black' are done, from straight use of 'Mars black' pigments to mixing up lots of deep thalo blue and umber colors, or otherwise trying to tap in on blue's 'receding' property. Geoffrey Chandler prided himself on never using black. I am inclined to agree that black pigment should be no more than a third of the volume of the color mixture I use for my skies (if I ever paint a traditional media picture again!)
  One place in an artist's work that almost allows you to look inside an artist's mind is the way they portray randomness.
  Pattern dominates efforts to mentally process textures into paintable forms, but the forces which scatter rocks, stars, and clouds about don't often behave predictably. The discovery (or invention) of fractal geometry is a landmark in perceptions of nature whose properties can be applied to such random things, and has provided new textural possibilities for our work as it has for our way of seeing things.Looking at my work of 20 years ago, my painted attempts at rocky surfaces now look weird to me, perhaps the computer makes me want to steer away from stylization, an impossible thing and undesirable in many cases.
  I am a worshiper of the look of the hyperreal, the 'take off my glasses and explore closeup' the micro textures nearby while the distant vistas peek from behind the nearby stuff. A simple Lunar landscape is among the most challenging textures to paint, from the scattering of the rocks to the soil details between the rocks. In space the distant scenery is distinguished by concentration of textures into other 'meta textures' in the distance, and by superimposition.
  The photos from the Apollo missions are our best record of being on another world, the later mission's surviving video records have much to offer on behavior of dust in a vacuum, gross lighting characteristics and variety of scenery on the Moon. As such exploration continues to recede into the mists of history, these sources will only grow in value.
  Getting an impression across through ones work of an unfamiliar place is obviously one of the things a skilled artist can do with a vast variety of approaches, the skill is in tapping into the visual cortex in such a way as to create a visual language most of the audience can 'read'.

   Much of the work of such paintings has traditionally been the preliminary sketches in which the composition and lighting are determined. Perspective drawing and shadow plotting techniques have provided the means of trying out the effect desired within the chosen composition. For the quarter century from the start of my professional painting career in late 1968 until the beginnings of my use of the computer with regularity, labor intensive perspective drawings of things like spacecraft and planetary ring perspective views accumulated in folders with each new painting. Now most such preliminary work, as with the primary act of painting, has largely moved to the computer.

   The creation of a desired element in a picture justifies any approach taken to achieve it.  In other words, if it works, do it.
   When Chesley Bonestell wanted to paint the rings of Saturn as a line or an extremely thin ellipse, a technical pen like a rapidograph would be used to outline the critical region as delicately as needed, then he would paint (with oils highly thinned with a painting medium) up to the line.

  For a space colony oil miniature, I opted to do it all in paint, including many cables nearly parallel but showing noticeable perspective. After the background painting was very dry, I hunched over the piece (about 12X16, oil on gessoed Masonite) with a steel straight-edge suspended solidly on a knee just above the painting. I had pecked tiny dots where each of the cables began and ended, and anchoring my little finger on the rulers edge I could glide my hand along just the right path to let the fine brush paint a narrow line. This line had to be thinner on one end to carry the desired impression of perspective.
   Each cable had to be painted after shifting the straight-edge and trying to insure just the right amount of paint (of just the right consistency) was on the brush. During the critical moments of painting the line the world disappears, my being focused on becoming a precision machine. A mistake could be wiped out with a 'Q-tip' dampened in turpentine, providing the background is thoroughly dry.
  Once layer 1 of the cables were dry, layer 2 could then be safely painted! The results were....OK.   Variations of width were inevitable.
  This is like a memory of a past life now, the turning point for painting such things was my Galileo cover for Sky and Telescope. The parachute lines were rendered in 3D, and repainted in Photoshop. I wouldn't want to think of doing them in traditional media, yet heroic effort used to be part of such works. In the end I like the look of the computer paintings for such hardware work.
  A painter can create an image with an effective resolution of, say, 2 pixels per millimeter (let's call that an average, anyway) The standards for an acceptable print are a good deal higher, roughly up to 10 pixels per millimeter.
  When painting digitally, as I now do, it is important to keep in mind the vast size such paintings must be to indure good display prints and reproductions in print media. More new work in shows will be digital prints, and there are advanced 'printers' which even spray an image onto canvas and art board surfaces using a digital file. I can see using such a device as an 'apprentice' and finishing up the result by hand.  

  Over the years I became familiar with the mathematical tools for calculating horizon distance, but I never used them, and they lie buried among my copious notes. I hated my math classes in public schools, and to this day I cannot do anything but add, subtract, multiply and divide. Despite my stunted math skills, I recognize the power and usefulness of mathematics to our world and to such work as mine.
  Now that I can 'look' at a virtual Earth with great detail and create an instant 'perspective drawing' at a whim. I can afford to try variations of views before committing myself. Only the first preliminary drawings are still exclusively done by hand, with my personal contribution to the painting concentrated on the things computers do poorly or slowly such as various amorphous details, and random things like rocks and other textures.  Following is an attempt to summarize some of my thoughts regarding the dawn of the digital age in art.

2. Film and animation

  When I became part of the Cosmos team of artists in 1979, visual effects for films were still based on traditional 'in camera' and model methods. Layering of images beyond simple multiple exposures required the use of tricky and very expensive optical printer work. At that time computer graphics had just reached the stage of creating images of solid plastic looking objects, but texture mapping was just being implemented and ray traced cast shadows were still some time in the future.

  All the shots of the surfaces of other planets required the production of 'tabletop' sized models, some built after design drawings I prepared. Although I was in charge of the planetary surface landscape models, all of the artists had some overlap of skills and it was common to join in on each other's projects to help get something done on schedule.

Left: three Cosmos artists pose for a Polaroid as the first planetary landscape model, that of Titan, begins to take shape. Left to right: Don Davis, Jon Lomberg, Rick Sternbach.

Right: I am painting 'lava' flow details on postulated 'ice volcanoes', the older surfaces were imagined as a reddish brown color due to continuous deposits of dark 'rain' of particles.







Left: The background panel being painted, fortunately Adolf Schaller and I handle clouds well enough we were able to create a work of consistent quality. Most of the left side was painted by Adolf, the right by me.

 Below: The final painted backdrop behind the finished model, ready for the cameras, rigged with tubing for the bubbly vapor from the 'ice volcano'.






  Many globes as well as planetary surface models were made during the production. One shot was of Pluto and it's then recently discovered moon. Charon. Two globes were set up in front of a large sky backdrop made of perforated opaque black sheets attached to a large sheet of frosted plastic, strongly backlit with banks of lights. The shot depended, of course, on the smoothness of the camera move as it was wheeled and moved on it's crane.



Below left: the models, right: How the shot appeared in the series 'Cosmos'.












3. The computer emerges.

  My first peek at a good computer paint system was in 1981, when I was asked to paint a series of texture maps of Saturn, the outer planets and some of their moons for the use by the Jet Propulsion Laboratory in upcoming Voyager encounter animations.  

  At that time Jim Blinn, the originator of many of the texture mapping methods now widely used, was working at JPL and had written much of a then sophisticated painting program which used an electronic stylus (with a cord attached) to apply painting, or rather scratching motions on a flat 'tablet' which caused blobs of a chosen color to instantaneously appear on a monitor facing me, a miracle to behold then. Both Jim and Pat Cole, another CGI pioneer, coached me in using the equipment.
  Rick Sternbach had used the same system to create texture maps for the earlier Jupiter encounter.
  Unfortunately such systems were one of a kind major setups then, a few years later the Quantel Paintbox was the best electronic paint tool. Used for retouching of video elements, time on such a system was fantastically expensive. A comet animation sequence for the 1986 reissue of Cosmos was one of the few occasions I actually got to paint for extended times on one. I had painted with acrylics on acetate cels several comet landscapes which were then shot on video and captured as high quality frames in the Paintbox. I then painted with the electronic stylus many layers of gas jets expanding and fading on frame after frame, which were later cross dissolved in a video editing suite to create an animation of dynamic erupting jets.
The first home computer I saw which allowed one to paint pictures with a mouse was an Amiga in 1984, and I was not encouraged by what I saw at that time.
Later glimpses of what Joel Hagen was able to create on an Amiga nearly lured me to that platform, but fate intervened.

  My 3 years career at Hansen Planetarium in Salt Lake city started in 1987. During the process of my learning the various aspects of show production I began my first continuous relationship with computers. The Digistar computer graphic projection system allows users to create detailed 3 dimensional models made of lines and dots using a 'puck' (fancy mouse with cross hairs) to trace outlines and contours while adjusting the 'working plane' with a dial with the other hand to define where in the third dimension the next line or dot will be placed. The show production dept. purchased a number of Macintosh computers, and I was encouraged to learn to operate them. After my first rocky encounter with the UNIX operating system used by the VAX computer which powered the Digistar, I found the Mac's intuitive approach a joy to use.
The first machine I began to play with was a Mac II with 8 bit color, with I believe 8 megabytes of RAM, and an early paint program called Pixelpaint. I learned to work around it's limitations and create some images worth using. One Pixelpaint drawing tool in particular, meant to simulate charcoal rubbing, was capable of surprisingly subtle textures. During a visit by David Egge, it was a delight to see how he became entranced with this same tool.  He sat until the early morning hours one weekend, building up in rubbed textures a Mars landscape. One observation he made then was that trying to paint with a Mouse was "like trying to draw with a bar of soap", which I heartily agree with.
 Later the Planetarium bought a crude scanner and an early tablet, which allowed me to establish the rudiments of computer assists for certain aspects of my art.

   A series of planet images were needed for a painting for JPL celebrating the Voyager encounters, and for this painting I first used the computer to generate a series of perspective views of the ringed planets using a rudimentary modeling program called Mac3D, these were photocopied to the desired size then taped together. I then laid a sheet of tracing paper over this comp, and while producing my 'master drawing' I would by hand smooth out the curves made of many straight line segments making up a ring model, for example.
  I could then cut out the separate friskits of the planets and rings from copies of that drawing and went at it with the air and paintbrushes. This acrylic on illustration board is the result.














 At the time I was also working on an animation of the Galileo probe entering the atmosphere of Jupiter, and simple models of the spacecraft were made in the computer to allow accurate views to be rendered and then traced on acetate cells from various angles.  For the probe entry sequence (left), I had to paint the fiery ablation from the heat shield on hundreds of acetete cels. The sequence of the probe deploying the parachute (right) also involved many drawings, the parachute cords were all drawn on acetete with a 'rapidograph' pen using white ink. All sequences tried to follow accurately the appearence and timing of the events in the entry sequence.  

 One sequence in the animated short (below) used over a hundred 'Mac 3D' renders of a slowly shifting perspective as the templates for the hand painted cels of the probe drifting down near us from overhead. These frames were patched together from a move which panned as it tilted, thus the black areas outside the shot. This animation was the last time I used 35 mm movie film as a primary medium. I painted, designed, and shot the film myself, at times using 'bi-pack' mattes to combine foreground action onto backgrounds.







 During the Voyager Neptune encounter I was in transition between Salt Lake city and the Los Angeles area. I was starting to work for WQED science effects, an effects house used for science related Public Broadcasting shows such as 'Infinite Voyage', and 'Space Age'. At left is a painting of the Voyager Neptune encounter, one of a series commissioned by Charles Kolhase of the Jet Propulsion Laboratory of each of the first Voyager encounters of the outer planets. All were based on printouts of 'line drawing' representations of actual views the spacecraft would see.











  During my career at WQED the transition began in earnest in visual effects creation strategies from the use of physical models and camera tricks to the growing digital options. John Allison, a Cosmos veteran, headed the visual effects facility in a converted warehouse in Glendale. We worked and played hard there.

 This is how we used to 'texture map' a planet map onto a sphere before the days of computers! We prepared an animated map of Jupiter's clouds, modified a 'wall' 35mm camera to project the developed film onto a carefully aligned spheroid, and filmed the result with our motion control mounted Mitchell camera. In the case of Jupiter the limb darkening was even simulated by a weak blue gel filter with a hole cut in it so when carefully positioned in front of the light only the edges would be affected from the camera point of view.









 In other cases realistic painting or modeling would be simply applied over a specially prepared globe, such as this 12 inch earth globe I painted for the PBS 'Planet Earth' series in the mid 1980's. Many of these shots were simple moves near the models against a backlit perforated black wall such as had been employed during 'Cosmos'. Some shots, however, required the use of 'mattes' to selectively expose parts of the film at a time. These were prepared and literally sandwiched onto the color film in the camera, using the precise, relatively cheap, and very labor intensive 'bipack' method. A special 'double decker' film magazine was required for this process, with one set of reels holding the color film which was exposed twice in the long process, the other set of reels holding the black and white film. Often parts of the surroundings had to be painted out directly on the film using a brush and opaqueing fluid, one frame at a time!


 We soon realised that computer graphics had not only evolved geatly in the last few years, they were becoming available in affordable equipment and software. As soon as we could we tried to 'automate' some of the tasks we were spending days at a time on in some cases. We bought a Mac II FX, with the then maximum RAM of any Mac at that time of 32 megabytes as well as 24 bit color on a large monitor.

 While at WQED I learned Electric Image, then and still the best Mac rendering and animation software. I also learned Adobe Photoshop, still my favorite digital paint software. By then the painting and animation capabilities available to users of such affordable home computers had surpassed that of the original system I had painted the Voyager texture maps with at the beginning of that decade.   The cheaply available electronic stylus now was not only cordless, but pressure sensitive as well, allowing great subtlety in the 'brushwork'. I produced my first computer animation on the Macintosh 2 FX, of a proposed satellite called EOS.



 In order to get these early sequences on film quickly we would actually set up the motion controlled camera to look at the screen, and after trying out various contrast adjustments and filter combinations we were able to obtain acceptable video res 'film transfers' using this decidedly transitionary method! Here John Allison makes the final adjustments before starting the shot.  During the cycle of exposing a frame in the camera the computer in the camera system would send a pulse the Mac was made to read as a 'mouse click', causing the next frame to appear in the sequence. We would start a shot and leave the room until the it was done.


 4. The computer as a painting tool

 Despite the traditional minority following of Apple, the Mac is the artist's computer as shown by the migration of large numbers of graphics people to that platform. The fact is, however, that clever people can do great work on any of the major platform choices. I now make virtually all my living with my computer, my drawing board is used more for retouching damaged old works and the preparation of that 'extra special' large painting I keep promising myself I'll do.
  Certain trends and changes in the methods of creating images come to mind that digital media has imposed on the process.
The act of painting on the computer is at first totally alien to the traditional tactile nature of creating art. The mark produced is in a totally different place than where the 'pen' touches the 'drawing surface'. In a way, when I learned the use of the Passche AB airbrush, I had to conquer a similar disturbance at seeing a mark appear 'off axis' from the tool I was holding!
  The Photoshop airbrush is quite good, and if used at a very low percentage value a good varied buildup can be achieved.
   Part of the effectiveness of computer airbrush effects lies in the number of brightness levels available in the picture. The current standard for home computer graphics is 24 bit, meaning 16,777,216 colors can be shown at any one time on the screen.
   I am told the human eye is capable of distinguishing some 50 million colors or so, crowded in sensitivity along the yellow to green transition of the spectrum. Thus 24 bit color (and more so the larger color numbers used in high end digital color work) approaches the ability of the eyes to perceive colors.
There is an important caveat to this, the limitations in the monitors to accurately convey all the colors in the spectrum.
   The color phosphors hit some parts of the spectrum nicely and miss others badly, a particular example being the green to blue portion. In order for the full color potential to be realized, a high quality print or transparency must be prepared from the picture file.   Most monitors are tuned way too bright, ideally areas of black in a picture should emit no light at all. For visual relief the computer should be in a room where at least one incandescent light falls across the walls, but not on the screen. This supplies continuous light, that is a non flickering illumination which will cover much of the visual field and reduce the screen's fatiguing effects on many people due to it's own subtle flickering.
   With the advent of multi-layering in paint programs new techniques of painting suggest themselves. a set of highlights on a rocky hill can be painted atop another layer, then that layer separately dimmed to see how a softer lighting would look.The entire project can be saved at critical points to allow fearless exploration that the traditional media of no return can restrict.
   It seems that indeed the process of creating an image is mutating beyond all bounds.
The use of the computer to do the perspective and part of the lighting setup for certain hardware paintings has advantages in that much of the perspective plotting 'donkey work' is now done by the machine, allowing me to concentrate on painting the 'personal' touches that distinguish the image from a purely CGI look.


   For a digital painting of the Galileo probe separating from the orbiter commissioned by NASA, I built in a 3D modeling program a complete Galileo orbiter and probe, taking care to show the thermal coverings over the instruments, the partly open antenna, etc. spending the better part of a week on it. I then assembled the model within the Electric Image software with appropriate color values assigned to the various surfaces. No texture maps were applied, the idea was to compose the picture and determine the lighting scheme.
  Once I liked an angle, a simple wire frame render was done of the spacecraft, black lines on white paper. This was then faxed to NASA for approval. As things turned out, they didn't like that or the second compositions, but because I could quickly generate other angle views the third try faxed in that day did the job.
  Already the work of building the 3D model had been justified.
Next a large (4096 pixels wide) image of the nicely lit model was rendered in Electric Image, and the next 3 days were spent painting on top of that render all the extra details specific to that view using photos of the spacecraft as reference.
  I thus let the computer do the work up to a point, then took over where the human touch mattered the most, what is normally the last ten percent of the work.
I think that even without significant painted embellishment, most hardware art could now be done by machine, saving the inspired painted images of that genre to people like Paul Hudson who continue to display the inherent superiority of talented artists over imagery created by people working in the realm of technology more than technique.
  Despite my absolute embracing of digital painting, there are still pangs of regret in the demise of physical painted surfaces as my primary technique, and I hope that traditional media won't soon be forgotten. So much of a fine painting is not reproducible on a phosphor monitor or even a high quality print.
  The inner light of the fiery glazed painted surface transcends the cold digital palette. A projected slide of a fine painting shot in polarized light will surpass in richness and subtlety all but the most expensive digital efforts to show the same painting.

  Ideally the computer is to be regarded as only another tool to convey our inner visions to the world.
  It is still necessary to learn perspective drawing and how to draw by hand the things we see. Perception is sharpened as we grapple with understanding and processing on a visual level, then define our own understandable visual language to communicate ideas with markings on a surface.
  And yet the writing is on the wall. Ours is the last generation to regard doing artwork in terms of purely physical media.

         Below: an animated 3D model of an alien insect like animal which evolved on a low gravity earthlike world, done for the Learning Channel series, 'Cosmic Safari', produced by Invincible films in London.