To get a realistic image, a good looking surface map is as important as a well fitting atmosphere and a pleasing terrain. If you like, the surface map is responsible for the color coating of the landscape. How to create a surface map is described in detail in a couple of other tutorials (e.g. at Terra-Dreams, at Terragen-Web, or at Terragenigma). Due to this you will find in the following just some basics regarding this thematic. The focus of this tutorial is settings, switches and sliders which are easy to be overseen and/or difficult to handle during the creation of a surface map.
All thumbnails could be opened in a separate windows and a bigger size by clicking at them.
First three fundamental items:
1. Angle of Illumination: In fact a visible bumpiness of a surface is caused by shadows. Due to that the angle of illumination is very important. As long as possible, a frontal illumination should be avoided, because this would suppress the creation of shadows.
The following images are illustrating pretty much the importance of the angle of illumination.
 |
 |
| In this image the sun was roughly exact positioned behind the camera. The structures of the landscape are more or less invisible. | Here the sun is positioned at the left side with some back light. The structures of the landscape are shaped very well. |
If you are orientating at nature, you will see that blatant colors are found really rare. Most often there are mixtures of colors present. At a lot of images, I have looked at, the green of the vegetation is just more green than it should be. Indeed it is not really simple to mix the correct color using the 3 sliders. Additionally it is quite difficult to make some hints. You will need some sensitiveness for that.
In principal you can say : the closer the three color values are together the less intensive a color seems to be. Regarding green used for grass, bushes or woods the results are more realistic if the slider defining the red portion is closer to the green slider than the blue one. Likewise the base color of grey rock should not be really grey, but a low saturated color. To add grey rock to a green landscape I often use a color with more red than blue. To interrupt rocks by other rocks I add some green which leads to weathered rocks or rocks covered by lichen and moss.
At all color gradations could be worked out very well and realistic by using some surface layers with dark colors, medium colors and light colors based on the same hue, but varied by uniform increasement and decreasement of the luminance to get some differences in brightness.
 |
|
| In this image each green color has same amount of red and blue and the grey hues are "real" grey (which means that red, green and blue are at the same value). Although the result isn't extremly different compared to the image on the right hand, the colors are not as natural looking as the colors of the right image. | Here the green colors are mixed with an higher amount of red and less blue and grey was done as very low saturated orange (mostly red, middle green and less blue). These colors are looking very realistic. |
3. Render Resolution: In general terms you can say, that the higher the render resolution is the more details are visible, and that will be regarding the terrain as well as the surface map. You should be prepared to accept noticeable longer render cycles. The unregistered version of Terragen allows a max. resolution of 1280 x 960 pixels. I normally render my images using the registered version and a resolution of 3200 x 2400 pixels.
 |
These two images are placed together to demonstrate the difference resulting in different render resolutions. The left one was rendered with a resolution of 640 x 480 pixels and the right one with 3200 x 2400 pixels. Both are resized to 400 x 300 pixels and re-sharpened afterwards using an image processing program. Despite the small size the right image shows much more details. This effect will be more obvious at images with a size of 800 x 600 pixels or 1024 x 768 pixels. |
Creation of a Surface Map
After we have discussed the basics I would like to start looking at the details of the settings which are available in the "Surface Layer" dialogue. A tutorial covering a similiar thematic, but focussing more on functional background information, could be found by Engineer. Now let me explain the opportunities as well as the impacts first.
The upper part of the dialogue - tab "Base Surface" - is responsible for the structure of the layer. The second tab "Children Surfaces" includes just one option and further opportunities may come in further versions of Terragen. So we skip a discussion of this tab.
In the lower part there is the tab "Advanced Distribution" which controls the distribution of the layers color and the layer itself. The tab "Distribution Presets" is again related to further versions of Terragen and by that we skip this one, too.
To start, we look at the slider Bumpiness which is more or less self-explanatory. This slider controls how smooth or how bumpy the layer will be represented in the landscape, resp. how intense the bumpiness should be visible. Even if the functionality could be explained in just two sentences, the usage of this setting holds some opportunities to make mistakes very easily, because there is a strong relationship to the structure you want to achieve. I will come back to this issue during the discussion of the slider "Scale+Depth". Additionally there are some images later on which are dealing with the effects of different settings of the bumpiness.
Underneath there is a slider called Mimic Terrain. This one defines how many details of the base structure of the terrain should stay visible. This setting is mostly of minor significance because the effect would not be recognized in a terrain which is composed by more or less "flat" regions. Working with a terrain which includes a lot of bumpy areas (even small ones), you can absolutely recognize differences by increasing or decreasing the value. So the default setting could be used in most cases. If you want to represent some snow or some sand you should decrease the default setting or even set the slider completly to the left side, because e.g. a thick snow mantle covers normally the structure of a landscape and will disguise it. And exactly this effect could be achieved using this slider. To point it out again, this slider don't influence the bumpiness of the current layer, but changes merely the influences of the bumpiness of the landscape thereunder.
Now let's have a closer look at the slider Scale + Depth at the right side of the lower tab "Advanced Distribution". This slider defines the scaling of details of the landscape and is strongly related to the slider "Bumpiness". From my point of view this slider would fit into the upper part of the dialogue much better, but anyway. "Scale + Depth" influences the size of the bumpiness as well in size as in depth. To illustrate this: if you move the slider to the right (increasing the value) the bumps will appear bigger and higher. If you dcrease the value the bumps appear smaller and more flat. It almost looks like, that you magnify or reduce the structures and details. To show the relationship between "Bumpiness" and "Scale + Depth" resp. to show the differences in results, I have put some example images hereafter.
 |
 |
 |
| Scale + Depth at 3,072 Bumpiness reduced to 1/3 |
Scale + Depth at 3,072 Bumpiness at 50 % (default setting) |
Scale + Depth at 3,072 |
 |
 |
 |
|
Scale + Depth at 0,432 |
Scale + Depth at 0,432 |
Scale + Depth at 0,432 Bumpiness 100 % |
 |
 |
 |
| Scale + Depth at 20,172 Bumpiness reduced to 1/3 |
Scale + Depth at 20,172 Bumpiness at 50 % (default setting) |
Scale + Depth at 20,172 Bumpiness at 50 % (default setting) |
If you will look carefully at the images above it is easy to detect that smaller values of "Scale + Depth" will result in smaller structures of the grass. This seems to be more realistic than using higher values of "Scale + Depth". Small values are suited perfectly to be used for small structured parts like grass, pebbels or rocky ground if you have - like in the example above - a distinctive foreground and the camera is positioned close to the surface map. The visibility of details will decrease with increasing distance but that's normal (isn't it?). A mid-range value - about 3 up to 5 - is nicely useable for rocks because than the details will stay visible even in greater distances. A larger forest - especially at higher distance or height - will need an higher value for "Scale + Depth". Which setting you select will depend strongly on the usage the layer will be meant for (grass, stones, rocks, forests or snow) resp. whether you will have a distinctive foreground and a low height of the camera which will request a lot of visible details. The keyword here is "scaling"! And to repeat this - looking at the images above it is obvious that "Bumpiness" in practice defines the way a structure (which is defined in size by "Scale + Depth") is visible.
The next slider, we will look at, is labeled Coverage. This slider controls the total amount of the landscape the layer will cover. Left handed there is a small preview which will show this. In which way the coverage is distributed (regarding the area) - except altitude constraints or slope constraints - Terragen controls at random. I have had often a situation in which the coverage was concentrated at one side, but I wanted to have it well distributed in small areas about the whole landscape. If this will happen, the only opportunity is to delete the layer and to create it again afterwards with identical settings. The way Terragen will distribute the layer will change definitely. If needed you can repeat this procedure as often as you like. The following images were rendered with identical settings and they show the distribution at random.
 |
 |
 |
DTerragen generates the colors of a layer per default by using nuances and due to this it is mostly difficult to identify the size and the locations of a new layer at once. In this case it is helpful to define a really blatant color first to give this layer a better visibility in the landscape. If the coverage is well defined, which could be checked by some pre-renderings, the layer can get its correct color. The amount of the landscape the layer should cover depends certainly on the object you want simulate. A meadow will surely get an higher coverage than some lichen on a dark rock. To get a well-balanced relation of all layers in total you should look at nature. Looking at landscapes eyes wide open - at photos, too - will be a requirement, I guess.
Fractal Noise defines the precision the color of a layer will be visible. Reducing this value (slider to the left side) will result in less color intensity. Decreasing "Fractal Noise" also will smoothen the edges. It seems that the color will be separated into a lot of small areas - and due to the spacing the lower layer will become visible. Accordingly the color is intensified and more distinguished if the value of "Fractal Noise" will be increased. The affected areas seem to be more compact. The following images should illustrate this. Furthermore there is an additional issue, may be not obvious: may be you have been faced with the situation that a layer was defined with a max. coverage and without any constraints, but the color was rendered with some gaps (best to be seen by "snow" layers). In this case the "Fractal Noise" should be reduced perhaps downto 0.
 |
 |
 |
| For this image the default setting of "Fractal Noise" was reduced by four big steps. | Here the default setting of "Fractal Noise" was used which means the slider was centered. | In this example the default setting of "Fractal Noise" was increased by four big steps. |
Underneath the slider "Fractal Noise" there is button labeled "Tex:". Clicking on it will result in a dialogue to define the settings of mascs to control the layer distribution. Regarding this thematic an other tutorial is existing, made by Frank Basinski.
The next settings I would like to talk about, are again a little more complex to be used and to be explained. Let us start with Altitude Constraints: the selection boxes "Max Altitude" and "Min Altitude" together with the appropriate input areas will define to which altitude a layer will be visible resp. at which altitude a layer will start. By the way, the altitude of a point of the terrain could be measured by moving the cursor across the landscape preview. Underneath the preview the current altitude is given at "z" value and could be read there. By that the altitude can be choosen and entered in the particular input area.
Aside of the input areas there are sliders labeled "sharp/fuzzy". Using these sliders you can define the grade of precision with which the entered altitude will constraint the layer. If a slider is positioned to the left ("sharp"), the altitude will limit the layer very precisely - like drawn with a ruler. The more the slider will be positioned to the right ("fuzzy") the more gradual the end (for "Max. Altitude") resp. the start (for Min. Altitude) of the layer will be. This behaviour is absolutly the same for both of the sliders. If a slider will be set to the right edge, the transition will cover a great altitude difference between minimum and maximum distribution according to the set coverage. The entered altitude value of "Max. Altitude" has to be the higher the larger the area of changeover between minnimum and maximum coverage of the layer should be - which means the more the slider is positioned to the right (to "fuzzy"). Accordingly the entered altitude value of "Min. Altitude" should be lower. In any other case there is not enough altitude range available to display the whole range of changeover and by that large amounts of the layer are not visible - from time to time this effect could be really irritating. To check whether the choosen altitude is correct, you can enter the altitude and make a pre-render with the slider completly positioned to the left ("sharp"). Now it is obvious where the the layer is starting resp. ending. Afterwards you just need to adjust this by changing the altitude and/or the "sharp/fuzzy" slider in small steps til the optimum setting is reached.
By the way, both constraints could be used at a time, if e.g. you would like to define a layer which should be visible between the two altitudes only. The setting of the sliders could be determined in the same way like mentioned above. The following images should clarify the discussed statements by usage of different settings of the entered altitude combined with different settings of the "sharp/fuzzy" sliders. To point out the relationship of altitude and the slider, the coverage of the layer (in these examples snow) was set to maximum.
 |
 |
 |
| The choosen value of "Min.Altitude" is 300m. At this altitude the snow should start. The "sharp/fuzzy" slider is set to the left edge. The border of the snow layer is like drawn with a ruler. | The altitude is again 300m, but the slider is positioned in the middle because the border of the snow layer should change smoothly, but the layer is visible in the area of the top of the mountains only. | Will the altitude be set to 0m, the snow layer will start more or less exactly at 300m. This result is caused by reducing the entered altitude compared to the original value. |
Now we will look at the last settings - Slope Constraints: These settings are very the same as the altitude constraints. The only difference is, that these settings are not dealing with the coverage of a layer in altitude but in definable areas of the slope of the terrain. The white portion of the bent line to the left of the selection boxes will show the starting slope (for "Min. Slope") resp. the ending slope (for "Max. Slope") the layer will be distributed. The "sharp/fuzzy" sliders are controlling how exact the slope definition will be used.
Regarding the impacts of illumination and coloring at the surface map, we have already talked about in detail.
In respect of the bumpiness I would would like to say that "less might be more" which means it would not make any sense to assign plenty of bumpiness to the layers more or less at random, just to get some structures in the landscape. A correct position of the sun may help much better. Furthermore even in nature details in the landscape structure are becoming invisible more and more with increasing distance.
The embodiment of small structures like pebbles, small stones and grass will more convincing by usage of lower values of "Scale + Depth", especially if the distance to the camera is just small. Quite often it is helpfull to assign different values of "Scale + Depth" to layers which are intended to grade the coloring of e.g. grass.
The grading of the coverage to get a nice color shading of a rock face for example seems to be more or less a question of feeling and sensitivity, especially the amount of layers is mostly higher than just one or two. If you are using a couple of layers, you should look at mutual overlaying of the layers.
Pretty often a small value of "Fractal Noise" will result in very good outcomes if you wanted to get a coloring of the layer to fit in the landscape in a discreet way. And this is an application I have quite often.
Last, but not least, you should take your time for creating an image. In the beginning I have rendered a new image every half an hour. Meanwhile I need about 3, 4 or even more days to create an image, each day a couple of hours working at it, til I reach my common goal - a realistic and convincing image fitting to my imagination.
Coming to the end there is may be one question left open: how many layer are necessary to get a good surface map created? Well, there exist no common answer for that because the amount of layers is just depending on the landscape itself. But if my landscape will have some rocks, meadows and dusty grounds, I will need 3 to 4 layers for the rocks, aside of the main rock layer (a lighter layer, a darker one, some lichen and maybe some moss) - the meadows needs in a minimum 2 layers eventually 3 or more (light, middle toned, dark, lush, dried, etc.) - the ground might be realized by just one layer e.g. if the ground will shine through the grass only. Due to this my surface maps are build by 7 to 8 layers mainly, from time to time more then 10. On the other hand you should avoid the usage of to many layers because the overlaying of the layers could reach a point where some of the layers are not visible anymore. After all the render time will increase corresponding to the amount of layers.
