Contrast

Contrast - the difference between light and dark shades - is a key element in making a object that you draw look three-dimensional instead of flat. Without contrast, your buildings are going to look like cartoon outlines. With contrast, your buildings will have definition, detail, and depth. Here are two examples of what a difference good use of contrast can make (for once, they're not my own SCURKs...):
 

  I deliberately "defaced" an early Joel Garcia tile to show you what happens when you don't use contrast correctly. Even though the facade of this building is very interesting and complex, it comes out as a washed-out mush because of lack of contrast. At the bottom, the entrance-way looks cartoonish because there is no blending between colors to make the door look detailed. Joel's actual tiles are much, much better than this, take a look at his New York tileset. If you want to see this tile before I messed with it, click here.
 
This is the SHIROKIYA DEPARTMENT STORE, Tokyo, SCURKed by Lee Sojot. Notice the complex roof line, the cylinder, dome, and cone that make up the corner tower, and his effective detailing. Lee's excellent use of contrast makes these features stand out. Contrast makes Clarity.
 
What is the main "technical" difference between these two images? On my previous page on shade, we introduced the "Rule of Three," you need at least a 3-shade difference in order to have contrast. Lee's department store follows the rule, my messed-up tile doesn't. Notice that not only does Lee use use the "Rule of Three" to decide which colors to use on the shaded side of the building and around the bend of the corner tower (where he gradually bends shades to fool the eye into seeing rounded shapes), but he also goes to lighter shades for the flags which jut out over the entranceway at the bottom of the corner tower and for the line over the ground-floor windows. Below is my analysis of the corner tower of Lee's building. Of course, I don't know exactly what was going through Lee's head as he was SCURKing this....my analysis just looks at how the different shades of grey work together to clearly define the complex detail of this tower:
 
 
As you can see, the "Rule of Three," is not meant to strictly dictate what the difference should be between adjoining surfaces, but rather it's a guide for what shades you can use. Lee uses a +3 contrast or even more when he wants to sharply delineate two surfaces but he uses less when he wants to create a smooth blend, like in the different facets of the tower. Notice also that Lee uses a shade very close to black for the shadowed part of the dome and the steeple. Black helps to outline and define shapes. Another especially important detail is the way that Lee uses white for the flags over the entranceway. The flags are drawn with only 10 pixels but look like they jut out from the rest of the building.

These same guidelines apply well to anything else that you are SCURKing.  Keep careful track of the relative shades you are using. Use +3 or more contrast when you want to show a sharp difference between surfaces, and +1 or +2 for a blend such as in a curved surface.  Do not be afraid to use pure black (or one of its equivalent shades).  However, be sure to use it sensibly -  as a very dark shadow or the last color in a blend of shades - otherwise your drawing will look cartoonish. As always, these are just guidelines, but by keeping track of the differences between shades you'll be effectively employing contrast to give definition.
 

Careful use of contrast helps define the multiple facets of the building, as well as the rooftop and window details. As you can see, using black can be effective.
 
The manner in which contrast aids detail and depth is more complex. Small details (such as Lee's flags) can be made to stand out by using a high contrast, especially when you use brighter shades for the detail against a dark background. This happens because brighter shades actually appear bigger - look at the two dots in the circle below:
 
   which dot looks bigger?

Both dots are only one pixel big, but the one on the left looks bigger. In general, any small detail that you draw in a lighter shade than the surrounding surfaces will appear bigger than it really is (how much lighter? use the Rule of Three...). So Lee's white flags appear bigger than they really are, therefore they look like they are closer to the viewer (remember the rules of perspective), fooling the eye into seeing 3-D. I use the same technique for the balconies in the building below:
 

White balconies jut out from shaded side of the building
 
Strictly speaking, the balconies should be progressively darker from left to right because they are on the shaded side of the building. However, by making them very light, I get the same 3-D effect as in Lee's department store.  Darker details also give an impression of depth, especially since any indentations in a building are naturally shadowed.  As I wrote before, contrast gives depth.
The light on dark contrast has other uses also. You might exaggerate the brightness of an object purely avoid that it gets lost in the shadows:
 
Original Size
Notice how light the back edge of this ledge is in comparison to it's surroundings
 
Again, strictly speaking, the back edge of the ledge behind the rooftop tower should be very dark because it is in shadow. However, if I drew it with the correct dark shade, it would "disappear" into the dark rooftop:
Instead, I made it  very light in relation to it's surroundings so that this small little ledge is now a distinctive feature. I did not use "Pure White," as it would be too overpowering, but rather one of the lighter greys which is just right. As usual, I used the "Rule of Three" and some experimenting to chose the right shade - the ledge is -3 from the wall of the roof tower.  This is yet another example of contrast giving definition.

By drawing a much lighter shade around a darker object - highlighting - you can make the dark object look a bit smaller, even though you haven't changed the color of the object itself. Look at the row of windows on the stairwell tower on the small office building below:

These windows are long, thin slits in the wall. However, due to the size of the building, I have only two pixels to work with - at best I'd get a rectangular window. I need to get each black window pixel look like it is only about half a pixel high. How do I do this? Under each window (the lower ones are somewhat obscured by the palm tree), I colored the wall white instead of the pre-dominant light-greys. The result is to make this piece of wall, the "windowsill," look bigger than it really is (brighter is bigger) and kind of "squeeze" the window above it to look narrower and longer. Note that in the figures below that this effect is really only noticeable in the original size view, where the pixels all kind of blend together:
 

2x zoom
Original Size
With white lines under windows
White lines emphasize "windowsills," make the windows look thinner
Without white lines under windows
"Normal size" windows
 
Dark areas (such as shadows) drawn next to light areas have a similar effect - the light areas look bigger. In the tile below, I have only a one-pixel wide vertical strip on the darker building on the left to represent the part of the front wall that is not windows. I need to visually widen this strip so that it looks about 1.5 pixels wide, but I can't make the strip any lighter because then there would not be enough contrast in relationship to the left-hand building (I'd lose definition). To solve this problem, instead of exaggerating the brightness of the vertical strip, I exaggerate the darkness of the shadow next to it, thereby making the vertical strip of wall appear wider:
 
Original size
Zoomed-in
Dark blue strip exaggerates size of grey wall to the left. notice that the lighter building to the right uses the same dark-against light technique for the left wall. Of course, in both cases, the dark strips are also simply shadows of their respective walls.
 
Note that these effects that I've outlined work because of contrast, not brightness. In other words, a white pixel looks big not because it is so bright but because it contrasts so much in relation to the pixels around it. A very dark pixel will not look correspondingly small. It will also appear to be bigger because it is distinct, the viewer can see it more clearly than if there were less contrast. Once again, look at the two dots in the circle below:
Which dot looks bigger?

Again, both dots are only one pixel but the dot on the left looks bigger, even though it's "pure black"! So you can get some of the same detail effects with very dark shades that you can with very light shades - the key element is high contrast.

Below is a fragment of a building that I've drawn three times, once with low contrast windows, once with high contrast windows, and once with "blended" windows. Note the apparent sizes of the rows of windows on both sides of the building:
 

Low contrast
High contrast
"Blended"
The low contrast windows look small or even non-existent (even though some of them are six pixels big!); the high contrast windows are distinct, but sometimes they appear too large, especially in the upper right row of windows and the two windows on the rooftop tower. The solution is to to "blend" the windows, using high contrast (black) to define the window size and then lower contrast to avoid a big blotch of a window. Here is a close-up view so you can see exactly what I did:
 
 I drew the black pixels in where the shadow would naturally fall. Remember, black must be used sensibly to be effective. The lightest grey that I used in the windows is the same shade that I used in my "low-contrast" example above. By blending from high to low contrast, I've got the benefits of both techniques without the disadvantages. This technique is particularly effective when SCURKing windows.

These techniques of exaggerating contrast allow you to overcome the chief limitation of SCURK - pixelation. You don't have to be stuck with blocky, pixelated details. Many times while SCURKing, especially if you are working from a photo, you will want to draw something that is "just a little bit" bigger or smaller than a pixel. By carefully controlling the contrast, you can get "half pixels," and can make the details appear any size that you want. A low contrast-pixel looks smaller than it really is, and a high-contrast pixel (especially a brightly-colored one) looks larger than it really is. Combining the two, as I did in the example above, you can modulate the apparent size of your details.  I use this technique a lot. ;-)

 

 
 
 Choosing which exact shade to use in the "blend" is, as usual, a matter of trial and error. As a guideline, I count how many shades are between the detail that I want to blend and the base color of the surface behind it (i.e. between a window and the wall) and then I divide by two. This "halfway shade" is my guideline to start experimenting and will appear to be about "half a pixel" wide - I use a lighter shade if I want a smaller apparent pixel and darker if I want a larger apparent pixel. On the building below, I wanted to make the small windows on the extreme ends of the sunlit facade a little bit longer (up and down direction). The windows are actually drawn with only 1 pixel,but I used a blend with the wall to make each window appear to be 1.5 pixels long. Note that these windows are actually a dark blue, but as I illustrated on the page on color, this shade of blue has a corresponding shade of grey on the color palette. The base shade and the detail shade are 16 apart, so my "halfway shade" is +8, but I felt that +8 was too dark so I used +4. I did something similar on the roof detail, so that the roof would look like it has a slant to it:
 
 
 Do all windows and details have to be blended? Of course not! I only use these effects when I want to fool the viewer's eye into seeing a little bit more than is actually there in the pixels.
 
Phew! I'm finally done this contrast page... since there is a lot of material here, I'm going to give you a quick summary (at the risk of repeating myself (again)): Overall, the proper use of contrast is one of your most effective tools in making realistic images with SCURK.
 

 
saci@voicenet.com