colour.io.convert_bit_depth

colour.io.convert_bit_depth(a: ArrayLike, bit_depth: Literal['uint8', 'uint16', 'float16', 'float32', 'float64', 'float128'] = 'float32') → NDArrayReal

Convert the specified array to the specified bit-depth.

The conversion path is determined by the current bit-depth of the input array and the target bit-depth. Supports conversions between unsigned integers, floating-point types, and mixed type conversions with appropriate scaling.

Parameters:

• a (ArrayLike) – Array to convert to the specified bit-depth.

• bit_depth (Literal['uint8', 'uint16', 'float16', 'float32', 'float64', 'float128']) – Target bit-depth. Supported types include unsigned integers (“uint8”, “uint16”) and floating-point (“float16”, “float32”, “float64”, “float128”).

Returns:

Array converted to the specified bit-depth.

Return type:

numpy.ndarray

Raises:

AssertionError – If the source or target bit-depth is not supported.

Examples

>>> a = np.array([0.0, 0.5, 1.0])
>>> convert_bit_depth(a, "uint8")
array([  0, 128, 255], dtype=uint8)
>>> convert_bit_depth(a, "uint16")
array([    0, 32768, 65535], dtype=uint16)
>>> convert_bit_depth(a, "float16")
array([ 0. ,  0.5,  1. ], dtype=float16)
>>> a = np.array([0, 128, 255], dtype=np.uint8)
>>> convert_bit_depth(a, "uint16")
array([    0, 32896, 65535], dtype=uint16)
>>> convert_bit_depth(a, "float32")
array([ 0.       ,  0.501960...,  1.       ], dtype=float32)