 This video will introduce you to processing raster data using both raster functions and geoprocessing tools. In general, raster functions are when you need quick, fast, temporary output, and geoprocessing tools. You'll use those when you want permanent data. There is not complete overlap in terms of the functionality between raster functions and geoprocessing tools, so it's important to keep that in mind. I'll be using two distinct datasets for this demo. The first is a DEM or digital elevation model. The second is a DSM or digital surface model. Both of these are 1.5 foot resolution datasets derived from LIDAR. The DEM is the bare earth or topographic surface, and the DSM is the true 3D surface that includes trees, buildings, and other above ground features. Raster functions are fantastic for cartographic effects like hillshade, because they execute so much faster than geoprocessing tools. Here I'm going in and creating hillshades for both my DEM and DSM. You'll notice that raster functions execute near-instantaneously. That's because they're not actually creating a new raster dataset, but rather they're creating a virtual representation of the existing raster data. With my new hillshade layers, I'm going through and making sure that any background values of 0 are not displayed as no data to ensure they're not showing through the layers underneath. I'm then organizing my raster surface models, putting the hillshade layers underneath, and setting some transparency for optimal display. I could have generated the hillshade layers using geoprocessing tools, but with raster functions, those layers were produced in a fraction of the time. There are dozens of raster functions available. In this next example, I'm taking the DEM and subtracting it from the DSM to create a normalized digital surface model or height above ground using the calculator function. In the calculator function, I set variables equal to existing layers, then perform mathematical operations using those variables. I'm simply creating two variables, one for my DEM, one for my DSM, and then subtracting the DEM from the DSM to produce a new raster layer. The new raster layers created virtually instantaneously. I'm then going to take the time to go in and adjust the symbology settings, changing the color ramp, and also making sure the background value of 0 is not set in their data. With the DSM or digital surface model, a tree in a valley with the same height as a tree in a hilltop would actually be at different elevations. But when we subtract the DEM from the DSM, we've normalized the heights relative to the ground. Now each pixel value represents the height above ground for buildings, trees, and other above ground features. Now I'm going to create an NDSM, but using geoprocessing tools. The geoprocessing tool I'm going to use is called the raster calculator. It's very similar to the function calculator. The key difference here is I don't have to set variables, and then that my output raster is actually a new raster dataset that's written to my hard disk. I'm storing this new raster dataset inside my geodatabase. The raster function calculator took less than a second to execute, but the raster calculator geoprocessing tool actually took two minutes. For the purposes of this video, I've sped that process up. Now let's take the time to dive in and have a closer look at the difference between a raster function layer and a raster output from geoprocessing. You can see that the function layer actually is not a new raster dataset. It's simply referencing the raster datasets that were the inputs to the function calculator. A raster function layer does not have any statistical information calculated by default, and its raster information is simply referenced from the original source layers that were used to generate it. The NDSM that we generated as part of the geoprocessing tool is quite different. This is because it's actually a raster dataset that's stored within the geodatabase. It's not referencing the inputs to the raster calculator. It's a brand new raster dataset with its own raster information and own statistics. When we go in and click on an individual pixel, we see that the pixel values for each layer are identical. What this tells us is that subtracting a DEM from a DSM is the same, whether you do it in a raster geoprocessing tool or as part of a raster function. That being said, the two layers are not identical. For example, you'll notice there's some strange striping within the function layer. This is because the DEM and the DSM did not align perfectly. As a result, this layer, which is a representation of those two layers, has a less than desirable appearance. The raster dataset that was the result of the geoprocessing tool, on the other hand, had to set a single cell size. Raster functions have the advantage that they execute extremely quickly, and in most cases, they'll suffice for the analysis that you're doing. Particularly if your goal is visualization. However, if you need absolute precision, it's a good idea to use the raster geoprocessing tools. If your goal is to distribute the raster data as a standalone dataset, you're probably better off using a raster geoprocessing tool as it produces new raster output. But with any raster function output, you can export the data to a new raster.