You are here: Advanced Encoding and Tools > Floating Point Images > Floating Point Quantization and No-Data

Floating Point Quantization and No-Data

Many floating point images denote no-data values as an arbitrary minimum value. For example, an image may use -9999 as the no-data value even though there aren't any pixel values between -9999 and 0.

Such large gaps in pixel values create an artificially wide dynamic range that may cause output images to appear dark. For more information on dynamic range, see Generating Dynamic Range Metadata. To ensure that output images display correctly, ensure that the no-data value is specified in the metadata before you compress the image. If the no-data value is not in the metadata, you can manually specify the no-data value with the -alphasource switch. Alternatively, you can use the shortened -as switch.

For more information on specifying no-data values, see Alpha Bands and MrSID.