Interactive Computer Graphics: A Top-Down Approach With Shader-Based OpenGL (6th Edition) Download.z
Jalisa Landgren
A 2D scene can be generated as particular case of a 3D scene bysetting the vertex Z coordinate to always zero, replacing the viewtransform matrix with identity matrix and setting the projection matrixas the ortographic projection matrix. This approach only works if thegraphics API default coordinate system is the same as the OpenGL one,where the Y and X axis are on the screen and the Z axis positivedirection goes from the screen towards the viewer (computer user).
Transfer function design is a difficult iterative procedure that requires significant insight into the underlying data set. Some information is provided by the histogram of data values, indicating which ranges of values should be emphasized. The user interface is an important component of the interactive design procedure. Typically, the interface consists of a 1D curve editor for specifying transfer functions via a set of control points. Another approach is to use direct manipulation widgets for painting directly into the transfer function texture (Kniss et al. 2002a). The lower portions of the images in Figure 39-7 illustrate the latter technique. The widgets provide a view of the joint distribution of data values, represented by the horizontal axis, and gradient magnitudes, represented by the vertical axis. Arches within the value and gradient magnitude distribution indicate the presence of material boundaries. A set of brushes is provided for painting into the 2D transfer function dependent texture, which assigns the resulting color and opacity to voxels with the corresponding ranges of data values and gradient magnitudes.
Interactive Computer Graphics: A Top-Down Approach with Shader-Based OpenGL (6th Edition) download.z
Download File https://oyndr.com/2yUEJJ
Dynamic spatial augmented reality (SAR) is expected to provide appearance editing by projecting images onto moving targets. Standard SAR for static objects [1], known as projection mapping, has already achieved realistic augmented reality (AR) and provided fantastic entertainment opportunities. A number of events use this projection technique to entertain people by changing the appearances of real objects such as buildings. As an advanced approach to this technique, dynamic SAR has huge potential to drastically extend the projection effects by supporting various projection targets, such as dancing people and their clothes, and interactions with these targets. This technique can also provide novel forms of amusement for near-future interactive AR games.
where \(s_\min \) and \(s_\max \) is the minimum and maximum load size in MB, respectively, l is the load factor clamped between 0 and 1, \(\,\mathrmsign\,\) is a function returning the sign of a real number, \(\Delta t\) is the frame latency, and x is the difference between the weighted maximum frame latency \(t_w\) and the maximum frame latency threshold \(t_m\). The adaptive load size s controls texture streaming rate and is computed every rendering cycle based on the relative change of frame latency. It has a fixed limit between \(s_\min \) and \(s_\max \) and is derived from a normalized load factor l clamped within the range of 0 and 1. During rendering, the load factor l is incremented or decremented until x is minimized; in other words, s will produce the optimal texture streaming rate as the weighted maximum frame latency \(t_w\) approaches the maximum frame latency threshold \(t_m\). To put things simply, one can think of \(t_w\) as the average frame latency and \(t_m\) as a user-defined threshold constant. The polynomial P defines the exponential change in l based on the magnitude of x. Through it, texture streaming rate will be exponentially reduced after a spike in frame latency, and likewise, steadily (or exponentially) increased when frame latency remains below the threshold (left of Fig. 4). The constants in P are empirically found to work well for s and with other parts of the adaptive streaming algorithm. At the end of the equation for l, we introduce the frame latency factor \(\Delta t\) to ensure a consistent rate of change across computer systems. In our implementation, we set the minimum load size \(s_\min \) to the size of a sparse image memory block (i.e., the granularity of the sparse image selected by the graphics API). We also set a large maximum load size \(s_\max = 64MB\) that will not be reached by most systems. One may also use this variable to hard limit the texture streaming rate to free up computing resources.
In the series of benchmarks (Tables 3, 4, 5), we analyze our adaptive texture streaming performance for the Berlin (atlas), Berlin, and Helsinki datasets, and compare them to a baseline without adaptive texture streaming. Without adaptive streaming, textures are streamed as fast as possible, producing the highest texture transfer rate measurement. We also observe a significant amount of frame stuttering after about 5 s into the benchmark, with maximum frame latencies surpassing 100 ms, resulting in a non-interactive rendering. With adaptive streaming (\(\alpha = 12\,\)ms in Table 3), the average texture streaming rate in MB/s is reduced by 47% to support a low, stutter-free average rendering latency of 5.3ms, which is an 88% reduction from the baseline. The number of mipmaps streamed per second remains similar to the baseline and even increases as \(\alpha \) approaches 16ms, indicating the deferment of streaming higher resolution mipmaps by our adaptive mipmap bias algorithm. Adaptive texture streaming performance is also similarly reflected for the Berlin and Helsinki datasets (\(\alpha = 14\,\)ms in Table 4, \(\alpha = 25\,\)ms in Table 5). In the texture-heavy Berlin dataset, we even see a 9% increase in mipmaps streamed per second which results in a quicker display of lower resolution textures for new camera views with fewer texture pop-ins. Additionally, frame stuttering is significantly reduced in all three datasets, both from measured maximum frame latency and from detailed observations of the rendering. We measure a 97%, 51%, and 80% reduction in maximum frame latency using the \(\alpha \) parameters of \(12\,\)ms, \(14\,\)ms, and \(25\,\)ms in the Berlin (atlas), Berlin, and Helsinki datasets, respectively. Also from the benchmarks, we can see that the texture transfer rate and frame latency increase as the target maximum frame latency \(\alpha \) increases, supporting the design of our adaptive streaming algorithms.