Module sample 

Type conversion of samples values.

Structs

ConversionResult

A conversion result, containing the resulting value as value and a boolean clipped indicating if the value was clipped during conversion.

F32BE

32 bit floating point, big endian. Stored as 4 bytes.

F32LE

32 bit floating point, little endian. Stored as 4 bytes.

F64BE

64 bit floating point, big endian. Stored as 8 bytes.

F64LE

64 bit floating point, little endian. Stored as 8 bytes.

I16BE

16 bit signed integer, big endian. Stored as 2 bytes.

I16LE

16 bit signed integer, little endian. Stored as 2 bytes.

I24BE

24 bit signed integer, big endian. 24 bits stored packed as as 3 bytes or padded as 4 bytes.

I24LE

24 bit signed integer, little endian. 24 bits stored packed as as 3 bytes or padded as 4 bytes.

I32BE

32 bit signed integer, big endian. Stored as 4 bytes.

I32LE

32 bit signed integer, little endian. Stored as 4 bytes.

I64BE

64 bit signed integer, big endian. Stored as 8 bytes.

I64LE

64 bit signed integer, little endian. Stored as 8 bytes.

U16BE

16 bit unsigned integer, big endian. Stored as 2 bytes.

U16LE

16 bit unsigned integer, little endian. Stored as 2 bytes.

U24BE

24 bit unsigned integer, big endian. 24 bits stored packed as as 3 bytes or padded as 4 bytes.

U24LE

24 bit unsigned integer, little endian. 24 bits stored packed as as 3 bytes or padded as 4 bytes.

U32BE

32 bit unsigned integer, big endian. Stored as 4 bytes.

U32LE

32 bit unsigned integer, little endian. Stored as 4 bytes.

U64BE

64 bit unsigned integer, big endian. Stored as 8 bytes.

U64LE

64 bit unsigned integer, little endian. Stored as 8 bytes.

Traits

BytesSample

A trait for converting samples stored as raw bytes into a numerical type. Each implementation defines the associated type NumericType, which is the nearest matching numeric type for the original format. If a direct match exists, this is used. For example signed 16 bit integer samples use i16. For formats that don’t have a direct match, the next larger numeric type is used. For example for 24 bit signed integers, this means i32. The values are scaled to use the full range of the NumericType associated type.

RawSample

A trait for converting a given sample type to and from floating point values. The floating point values use the range -1.0 to +1.0. When converting to/from signed integers, the range does not include +1.0. For example, an 8-bit signed integer supports the range -128 to +127. When these values are converted to float, 0 becomes 0.0, -128 becomes -1.0, and 127 becomes 127/128 ≈ 0.992. Unsigned integers are also converted to the same -1.0 to +1.0 range. For an 8-but unsigned integer, 128 is the center point and becomes 0.0. The value 0 becomes -1.0, and 255 becomes 127/128 ≈ 0.992.