About using transparency in an image service definition

There are cases when the source raster datasets have been projected and saved to a file with a border of cells around the data representing invalid data (null data). It is recommended that you edit the footprints to remove these areas of unwanted pixels; however, you may prefer to use transparency to remove these pixels from the image service. There may also be other cases where a pixel value in your raster dataset represents a region you want to be transparent.

If you want to edit the footprints to remove unwanted areas from your image service, you can edit the footprints with the standard ArcMap editing tools or using the Recompute Footprint tools, such as Recompute Footprint By Radiometry.

Learn more about editing the raster dataset footprints

To define a transparent pixel value, you must define a NoData value. You can define the NoData value on the raster dataset or for each raster contained within the raster dataset; however, you will generally only need to define the NoData value for the raster dataset. You cannot define the NoData value at the service level. There are three ways to edit the NoData value for a raster dataset in your image service definition:

You must also set the resampling method for the clients to Nearest Neighbor; otherwise, the NoData value will become blended with the pixels next to it using bilinear interpolation or cubic convolution.

Optimization issues

When pixels in raster datasets are interpreted as transparent, the service provider has to perform additional computations to create the output image. Using transparency is an optional setting that is not always applied to an image service definition. Normally, raster datasets are clipped by their footprint, which minimizes the number of pixels that need to be read from each. When using transparency, every raster dataset needs to be read and every pixel needs to be examined so the transparency can be applied correctly. This can increase the processing time by up to 1,000 percent if there are many overlapping raster datasets. Therefore, it is recommended that footprints be modified wherever possible, such as around the edges of raster datasets, to minimize the use of transparency.

Examples

Without using transparency or editing the footprints

No transparency

Using 0 to define the transparent pixel value

With transparency

Using bilinear interpolation

Using Bilinear Interpolation
Using Bilinear Interpolation

Using nearest neighbor:

Using Nearest Neighbor
Using Nearest Neighbor

7/10/2012