My work environment includes: a computer; black-and-white and color monitors housing the essential color and memory boards; a mouse and keyboard for inputting commands; a camera to act as a digitizing frame grabber; a separate camera to record the processed image; and programs to grab and store images, manipulate brushes, position grids, and analyze color.
The black and white monitor is used for typing commands to store or retrieve an image from or to the color monitor and tape. Errors and reminders that the storage
space is running out also appear in a window on this monitor.
The mouse and keyboard can be used interchangeably to activate a command.
The digitizing frame grabber accepts Red Green Blue analog signals from a videocamera and converts the signals to RGB digital data. A frame grab is done in one frame at a time (1/30 of a second).
To digitize, or grab a frame, I position a picture on the scanner. The image is focused, and the colors and lights balanced. The scanned picture will appear on the color screen. Once I am satisfied that the scanned image compares favorably to the photograph of the art I then use the mouse to click on Save to store the picture for further work.
Once I have selected a palette I restore a crack or fissure by using brushes of different sizes. I refer to brushes in the same way I would refer to brushes in the traditional sense but brushes in the computer have different attributes. One difference is that I do not hold my brushes but use a device, the mouse, to make the brushes and to paint with it. These computer brushes exist only in the machine.
It is not possible with traditional means to make a brush with a stroke that is actually an image. For example, once I identified the repeatable pattern in the background of \fISt. Julien\fR I made a brush, actually the shape of the pattern, and filled the damaged area by simply repeating the pattern or overlaying it. Unlike traditional brushes, I can also unpaint an area if I am not satisfied with the results.
Another advantage with the computer brush is to be able to save it. In electronic
restoration this is a more dependable way of working with color rather than relying on subjective memory of color.
Whenever a brush or an image is saved in the computer it must be named in order to retrieve it from memory.
Grids & Templates
I can use a variety of programs to process the stored picture. For example, I can place a grid over the entire image to aid in the study of specific areas and to check the proportions of the image. The grid size can be easily changed with different distances between the vertical and horizontal spacing lines of the grid and removed at different steps of the analysis. Some pictures need smaller grids to examine details. It is possible to work with the grid on since it is
actually sitting on the screen on a different level from the scanned picture. It exists as an overlay and doesn’t affect the picture. Upon completion of the analysis I simply remove the grid, or as in the \fIResurrection\fR, where I compare the sizes of the various Christ heads, I left the grid in place as a measurement device.
Another program permits me to take any area of a picture and test whether the shape would match or map into a geometric template such as a cube, cylinder, or triangle to abstract a composition. See of a soldier’s head in the \fIResurrection\fR.
Another program permits me to select colors from an image, a picture element (pixel) at a time, and store them for future use. This feature is useful in electronic restoration to identify surrounding areas of a crack.
Before I begin restoration I select a number of colors and record them in small boxes or color pages. I can click on the mouse to go forward, backwards, a line at
a time, or a page at a time. The number of pages I can store depends on the amount of storage.
All color screens have a color map, or color lookup table (LUT), to translate the pixel data to actual video color data.
On a color screen, there is a pipeline of data from the pixels of color memory to the video stream you see on your monitor. The translation follows the following path.
Picture Elements to Color Map to RGB Video to Screen
On a pseudo-color system, one 8-bit pixel is fed to all three red, green and blue color maps. In this mode, 256 colors are available at any time, and the current
color map contains a palette of 256 colors. The broadcast-resolution system has a 12-bit mode that behaves similarly, with a palette of 1024 colors.
On a direct color system, eight bits each of red, green, and blue pixels are independently translated by the red, green, and blue color maps. In this mode, the color maps are typically simple intensity ramps.
The color map contains several segments, only one of which actually drives the video at the moment. The map segment that is driving the video is called the Primary color map segment. Each map segment contains specifications for the red, green, and blue intensity to be associated with each pixel value. the specifications vary with the hardware configuration.
Four color map segments are implemented in hardware in a high-resolution system. Sixteen color map segments are implemented in hardware in a broadcast resolution system, and one color map is implemented in hardware in a digitizing frame grabber system and a CAD buffer system. On a digitizing frame grabbers system and a CAD buffer system, the software emulates eight hardware, color map segments. Normally, one map segment is driving the video portion of the color system at any one time.
Transparent overlay is a special translation to make it possible to create
special effects with the video using color maps. There are two types of overlays. One uses some number of bits of pixel data to select which color map to use, while the rest of the bits select which color to use in that map. For example, to test Piero’s Resurrection” to check out the idea that he painted it to appear at sunset,
I started with a small number of bits of pixel data using pure red, then slowly increased the number of bits of pixel data. In the final version the effect makes the image appear as if it is seen through a red filter.
The second type of overlay is the replacement overlay which uses some number of bits of pixel data to specify a value to use in some other color map segment. This type of translation makes it possible to insert blocks of graphics to replace the normal pixel data such as with Piero’s painting of Christ’s face where I flipped 1/2 of the image.
Most monitors adjust a nonlinear signal source by varying the brightness linearly. Therefore, the color maps map the digital color values (usually 0-255) directly into video values, with no gamma correction. However, if the amount of gamma correction performed by the monitor is inadequate, you can change the color map to perform gamma correction. Since the eye’s response to light is not linear a gamma correction process is valuable to correct the brightness values to appear linear.
A Color Model
A color model contains three components that give a color a unique specification.
The IHS Intensity, Hue, Saturation color model mixes colors much the same way an artist mixes paints on a palette. Intensity controls the brightness of the color. For example, black has the lowest intensity, white has the highest. The shades of grey have intermediate intensities.
Hue controls the frequency of the color. For example, red and blue are different hues.
Saturation controls the richness of the color. For example pure blue has a very high saturation. As pure blue becomes pastel and turns into grey the saturation decreases.
A fourth model is the RGB color model which combines colors much the same way the human eye combines the light frequencies of the three primary colors. The three components of the RGB color model represent these color primaries: red, green, and blue. The three components red, green, and blue also correspond to the three red, green, and blue electron guns in a color monitor. To mix colors, the three primaries are added together. For example, mixing red and green creates yellow, and mixing equal amounts of red, green, and blue produce a grey scale color.
The intensity, hue and saturation can be selected independently to change more than one aspect of a color at a time affording greater flexibility and control in combinations of color. I select colors by using a mouse or typing in the numbers for the colors, intensity, hue and saturation. Sometimes the latter method is more accurate.
Rate of Gradients
A procedure for gradating from one color to another is extremely valuable for electronic restoration where two main colors and their combinations are identified for a specific area. I can control the rate that one color gradates into another as well as the rate of gradients in either direction.
I can copy the color area that I made up into a shape of a different size. This is important when I am restoring a large area and then find that the same palette was used in a smaller area. When the brush is reduced in size the colors are adjusted accordingly. It is essential to check the colors again since the computer will sometimes take one color as the dominant color and adjust accordingly.
When I have difficulty identifying the colors I sometimes use the old masters technique of first selecting a grey palette and then adding the color to the greys.
Step-by-Step for Electronic Restoration
1. Magnify the area that needs restoration
2. Move the color select box over the color to be identified.
3. Select and store the color. Repeat this operation for as many times as there are
changes in the color.
4. Try different rates of gradation from one color to another.
5. Test the results by painting a patch to the side of the picture.
6. Make a cutout brush of this mixture and place it over the crack.
7. Enlarge the area again to test the match with the adjacent colors.
8. Check the colors against a read-out of the numbers of the surrounding colors.
9. Fix the patch and test at different magnifications.
10. Store this brush for future use on other parts of the canvas. An artist usually works with a given palette but in fresco painting the mixtures will change from day to day. The color identification will give a clue as to the sections painted from day to day.
11. Put a grid over the image and mark off the areas that hold the same color values. that allows me to change the number of colors in the picture. If I am examining the composition alone I like to reduce the number of colors. Sometimes I start with black and white, then grey, than slowly add colors until I have the full
palette. I can then study the composition at each step without the intrusion of color information. I can also get a better sense of what color does for the composition.