Version: 2022.2

Mathf.FloatToHalf

切换到手册
public static ushort FloatToHalf (float val);

参数

val The floating point value to convert.

返回

ushort The converted half-precision float, stored in a 16-bit unsigned integer.

描述

Encode a floating point value into a 16-bit representation.

Converting a floating point value to a half causes it to lose precision and also reduces the maximum range of values it can represent. The new range is from -65,504 and 65,504. For more information on 16-bit floating-point numbers, and for information on how precision changes over the range of values, see Half-precision floating-point format.

If the converted floating point value falls exactly between two half-precision values, this method rounds it to the value furthest from zero (Round away from zero tie-break rule). This selects the value closer to positive or negative infinity, depending on the sign.

You should only use the returned ushort as a storage format. If you want to perform mathematical operations on it, first convert it back to a float with Mathf.HalfToFloat.