5. Géneros Fotográficos
5.3 El género fotográfico en el contexto de la transformación fotográfica
Displaying Illustration Layer data in 3D space is challenging, especially when the infor- mation to be presented is text. Besides the challenge of maintaining correlations and readability, the 3D interface in itself is linked with a number of problems, for example: disorienting navigation, complex user actions, and annoying occlusions [Shn03].
Besides requesting Illustration Layer data, the exploration of a 3D visualization includes interacting with the camera for view control. But changing the view requires layout updat- ing when Illustration Layer data is displayed. According to the afore itemized demands, this comprises searching for empty or unimportant regions as well as verifying that there is a clear correlation between Background and Illustration Layer components. Also, abrupt changes in the layout should be avoided because they can be annoying and confusing. Figure 6.8 shows an annotation technique that directly attaches text to its corresponding scene component (published in [SCS04]). Besides object closeness, a semitransparent poly- gon provides further visual feedback regarding the object-text correlation. Images (a) and (b) illustrate design variations: (a) shows an annotation of rectangular shape that facili- tates reading whereas (b) shows an annotation whose shape is adapted to the curvature of the corresponding scene object, which links object and annotation even closer.
Figure 6.8: Object-attached annotations. The annotation remains attached to its scene object when the view changes. Images (a) and (b) show design variations; (c) illustrates occlusion, which is reduced due to semitransparent rendering.
The most important features of this technique are:
• Annotation reading is facilitated by exact screen-aligned text rendering.
• Correlation between scene component and annotation is clear due to object closeness. • Annotations remain attached and move smoothly with their objects when the view
changes during scene exploration.
• Annotation shape and size can be adapted so that scene occlusion can be reduced or the amount of displayed text can be adjusted.
The last point indicates that scene occlusion can be a drawback of this technique. Fig- ure 6.8(c) illustrates such a scenario. But it also shows that the semitransparent rendering style still lets the viewer guess the structures being occluded. An alternative approach would be to place annotations with a certain distance outside of the area covered by the rendition. In static layouts, this method is appropriate. But when the layout permanently changes, longer distances not only complicate identifying correlations, they also increase the distance an annotation has to move when its counterpart changes the position. Employing distortion tools for scene exploration
Distortion lenses can be considered as parameterizable objects which are typically employed to magnify a focused region while the context is scaled down. Their basic components are:
6.3. Presentation of Descriptive Metadata 125
• the focal point, or region, at which the distortion is maximal,
• the area, or scope of influence, in which the lens creates an effect, and • the manner in which this effect attenuates (drop-off function).
In the following, two applications of lens distortion are proposed which can assist in ex- ploring a 3D visualization [SCS04].
The first application serves to adjust certain parameters, which can be compared with the functionality of traditional sliders. But there are two fundamental differences: (1) slid- ers are in general one-dimensional controllers whereas distortion lenses can be two- or three-dimensional, and (2) sliders adjust attributes linearly whereas the effects caused by distortion lenses depend on the drop-off function applied.
In this approach, a distortion lens is attached to a particular object of the visualization so that the lens moves with the object or remains at a place when the object pauses. In addition, the lens which is two-dimensional covers the region the object takes when it is projected onto the screen. When the user now moves the mouse cursor within the lens’ scope of influence, certain parameters associated with the particular scene object can be adjusted. That is, moving the cursor towards the focal point yields a maximum value, moving it towards the peripheral region decreases the value.
Figure 6.9: Employing distortion for text display. The level of distortion affects the size of the annotation box and text scrolling.
Figure 6.9, for example, shows how a user can control the extent to which the annotation box grows towards the image boundary when lens distortion is employed. To this end, a scene object becomes transparent enough to reveal its centroid (focal point rendered as a sphere) once it is reached by the mouse cursor. Selecting the object with a mouse click starts to reveal the annotation. As the cursor is moved towards the centroid, the annotation expands. A second click freezes the annotation’s current state. Now, moving the cursor will scroll the text if it does not entirely fit into the annotation box. When the cursor leaves the object, it is no longer rendered as transparent. This technique is an example for
indirectly controlling two parameters—annotation extent and text scrolling—by moving the mouse cursor within the same small region.
Figure 6.10: Employing distortion for exploration tasks. A 3D lens (probe) is moved through the scene and creates an explosion diagram.
The second application temporarily affects the composition of the visualization. Here, the lens serves as a probe which can be moved through the scene to reveal occluded scene objects and to shed light on the scene structure (see Fig. 6.10). In this way, the user needs to handle only one object—the probe, which spares to select and reposition every single scene object involved.
In contrast to the first application, the lens (probe) is independent from particular scene objects, and it operates in three dimensions. It still has a focal point and a certain scope of influence. A user can control the probe with a cube which moves with mouse repositioning in the plane associated with the face selected. As the probe is moved into the scene, the line of sight between the probe and the user is kept clear. To this end, the probe affects
6.3. Presentation of Descriptive Metadata 127
scene objects by pushing them apart, which is illustrated in Figure 6.10. The image also shows that annotations attached to scene objects remain with their objects when they are moved. As soon as objects leave the scope of the probe, they are moved back to their initial positions.