stores information in a floating point format, so pixel values are limitless and can hold negative numbers delivering much higher quality and more accurate data. 8 and 16 bit formats can hold only a fixed range.