|
|
|
Cyberware designs, manufactures, and sells standard and custom 3D scanning systems and software. | |||||
The following discussion does not include any references to the use of the Surface Tools feature of Echo or the gradlim command in attempts to improve the appearance of the transition zone between the periphery of a relief image surface with a background surface treatment. Both tools have been tried more recently, in combination with various other tools and with varying degrees of success. A discussion of their use is in progress.
The relief operations available in display mode enable you to convert a particular rendered view of a subject into a resampled cartesian range map. If the original image has an associated color file (rgb or texture) and the number of latitudes in the image is equal to the number of longitudes, a resampled color image for the view will also be generated. If the latter is not the case only the range data is resampled.
The range surface of the relief image will be nearly identical to the range surface from which it is generated whenever the view is displayed and rendered in the orthographic projection mode. If the relief is generated from the same vantage point but with perspective, the resulting range surface will be somewhat distorted from the original due to whatever degree of perspective is used in order to achieve the more "natural" screen-projected appearance.
Relief images may be generated from a variety of image sizes with respect to number of latitudes and longitudes, however, the number of latitudes must be either equal to or exactly one half the number of longitudes. The relief generation process involves resizing the render window by selecting the Relief Resize menu option. The resulting Relief window will be sized to contain an array of screen pixels equal to the number of longitudes in the image. This Relief window will most commonly contain a 512 × 512 pixel array.
The view that would normally appear in the Render window will now appear in the resized Relief window at the appropriate size and aspect ratio. The same operations available to you in the Render window continue to be available in the resized Relief window. The relief view may be adjusted at will up to the time of initiating the relief scan by selecting the Relief Scan menu option.
The relief scan results in the depth value associated with the projected surface, relative to some reasonable datum, at each pixel of the Relief window being mapped to the corresponding range point of the relief range map. If this happens to be a color image, a corresponding color image is also generated in which case you will observe a noticeably longer resampling time resulting.
You will need to exit display mode prior to any attempt to display the resulting relief image. This allows the entire display process to be reinitialized with what is now a new image in its own right as if it had just been digitized.
In summary, the relief making operation involves:
In order to form a bas relief images from a relief image it will be necessary to determine the required dimensions of the finished bas relief. The preferred order is to first determine the scale factor required to achieve the desired latitudinal and longitudinal dimensions. The next step is to find the depth scaling factor required to further reduce the range of depths occupied by the highest and lowest features of interest by the desired amount.
The first scaling should be accomplished using the fscale command whereas the second scaling may be accomplished using the rlogprop command. This command allows for depth proportioning by a constant decimal percentage factor. It also allows for depth proportioning by applying decimal percentage factors, which vary either linearly or sinusoidally over a specified range of depths. One constant factor is applied above the specified range of depths and another below.
The simplest approach is to invoke the command
| echo > rlogprop fscale |
where fscale is the decimal scaling percentage required to reduce the depth by the desired amount.
The somewhat more involved approach involves selecting a band of depths on the relief image demarked by a datum_hi and datum_lo between which the decimal percentage scaling factor varies linearly from a factor rprop_hi to the factor rpop_lo associated with datum_hi and datum_lo respectively. This proportioning strategy is invoked by the command
| echo > rlogprop -r rprop_hi rprop_lo datum_hi datum_lo |
The factor rprop_hi is applied uniformly above datum_hi whereas rprop_lo is applied uniformly to depths below datum_l.
Using essentially the same command structure as immediately above except for replacing -r with -t invokes a proportioning mode wherein the variation from rprop_hi to rprop_lo varies sinusoidally with inflection occurring at mid-depth. The intended effect of both strategies is to reduce depths more dramatically through a range of depths containing features of somewhat lesser interest. The choice of options is largely a matter of judgment based upon experience with their use in resolving various depth compression objectives.
An example of an instance where one might want to use these commands would be the rapidly receding depths lying along the lateral surfaces of a face-on view of a person's head, i.e., behind the eyes but forward of the ears. In order to achieve a reasonably pleasing appearance of the features which tend to characterize a person, even though radically foreshortened by the required depth compression, it is desirable to minimize the foreshortening of these essential features while maximizing the foreshortening of less essential features.