hi::v1::audio_sample_format Struct Reference

#include <hikogui/audio/audio_sample_format.hpp>

Public Member Functions
constexpr	audio_sample_format (audio_sample_format const &) noexcept=default
constexpr	audio_sample_format (audio_sample_format &&) noexcept=default
constexpr audio_sample_format &	operator= (audio_sample_format const &) noexcept=default
constexpr audio_sample_format &	operator= (audio_sample_format &&) noexcept=default
constexpr	audio_sample_format (uint8_t num_bytes, uint8_t num_guard_bits, uint8_t num_bits, bool is_float, std::endian endian) noexcept
	Constructor of an audio sample format.
constexpr	operator bool () const noexcept
float	pack_multiplier () const noexcept
	How much to multiply float samples to create integer samples.
float	unpack_multiplier () const noexcept
	How much to multiply integer samples to create float samples.
std::size_t	num_samples_per_chunk (std::size_t stride) const noexcept
	The number of packed samples that are handled in a single 128 bit load or store.
std::size_t	chunk_stride (std::size_t stride) const noexcept
	The number of bytes to advance to the next chunk to be loaded or stored.
std::size_t	num_chunks_per_quad (std::size_t stride) const noexcept
	The number of chunks to load or store to handle 4 samples.
std::size_t	num_fast_quads (std::size_t stride, std::size_t num_samples) const noexcept
	Calculate the number of 4 sample-quads can be handled as chunked loads and stores.
i8x16	load_shuffle_indices (std::size_t stride) const noexcept
	Return a shuffle indices for loading samples into 32 bit integers.
i8x16	store_shuffle_indices (std::size_t stride) const noexcept
	Return a shuffle indices for storing 32 bit samples into packed samples.
i8x16	concat_shuffle_indices (std::size_t stride) const noexcept
	Return a shuffle indices to shift previous loaded samples for concatenation.
constexpr bool	holds_invariant () const noexcept
	Is the audio sample format valid.

Static Public Member Functions
static constexpr audio_sample_format	float32_le () noexcept
static constexpr audio_sample_format	float32_be () noexcept
static constexpr audio_sample_format	float32 () noexcept
static constexpr audio_sample_format	int16_le () noexcept
static constexpr audio_sample_format	int16_be () noexcept
static constexpr audio_sample_format	int16 () noexcept
static constexpr audio_sample_format	int20_le () noexcept
static constexpr audio_sample_format	int20_be () noexcept
static constexpr audio_sample_format	int20 () noexcept
static constexpr audio_sample_format	int24_le () noexcept
static constexpr audio_sample_format	int24_be () noexcept
static constexpr audio_sample_format	int24 () noexcept
static constexpr audio_sample_format	int32_le () noexcept
static constexpr audio_sample_format	int32_be () noexcept
static constexpr audio_sample_format	int32 () noexcept
static constexpr audio_sample_format	fix8_23_le () noexcept
static constexpr audio_sample_format	fix8_23_be () noexcept
static constexpr audio_sample_format	fix8_23 () noexcept

Data Fields
uint8_t	num_bytes
	The number of bytes of the container.
uint8_t	num_guard_bits
	The number of bits used for the integer part of a fixed point number.
uint8_t	num_bits
	The number of significant bits of the sample format.
bool	is_float
	The numeric type is floating point.
std::endian	endian
	The endian order of the bytes in the container.

Detailed Description

Audio sample format.

Audio samples described by this type can be in three different formats.

• Signed integer PCM, which will be treated like a fixed point where the num_integer_bits is set to zero. This is the format used in most audio file formats.
• Fixed point PCM, a more generic format than signed integers which has some head-room/guard bits above normalized signed integers. The Q8.23 fixed integer format is used by iOS as a sample format.
• Floating point PCM, the floating point format used internally in hikogui, useful for doing calculations in.

Sample values are aligned to the most significant bits of the container described by num_bytes. The bottom bits are set to zero.

Constructor & Destructor Documentation

audio_sample_format()

constexpr hi::v1::audio_sample_format::audio_sample_format ( uint8_t num_bytes, uint8_t num_guard_bits, uint8_t num_bits, bool is_float, std::endian endian )

inlineconstexprnoexcept

Constructor of an audio sample format.

Parameters

num_bytes	The number of bytes used for each sample in the stream of data.
num_guard_bits	The number of bits used beyond -1.0 and 1.0.
num_bits	The number of bits used to represent a normalized sample between 0.0 and 1.0 (without the sign bit).
is_float	True if the sample is float, otherwise integer or fixed point.
endian	The ordering of bytes in each sample

Member Function Documentation

chunk_stride()

std::size_t hi::v1::audio_sample_format::chunk_stride ( std::size_t stride ) const

inlinenoexcept

The number of bytes to advance to the next chunk to be loaded or stored.

concat_shuffle_indices()

i8x16 hi::v1::audio_sample_format::concat_shuffle_indices ( std::size_t stride ) const

inlinenoexcept

Return a shuffle indices to shift previous loaded samples for concatenation.

holds_invariant()

constexpr bool hi::v1::audio_sample_format::holds_invariant ( ) const

inlineconstexprnoexcept

Is the audio sample format valid.

load_shuffle_indices()

i8x16 hi::v1::audio_sample_format::load_shuffle_indices ( std::size_t stride ) const

inlinenoexcept

Return a shuffle indices for loading samples into 32 bit integers.

num_chunks_per_quad()

std::size_t hi::v1::audio_sample_format::num_chunks_per_quad ( std::size_t stride ) const

inlinenoexcept

The number of chunks to load or store to handle 4 samples.

num_fast_quads()

std::size_t hi::v1::audio_sample_format::num_fast_quads ( std::size_t stride, std::size_t num_samples ) const

inlinenoexcept

Calculate the number of 4 sample-quads can be handled as chunked loads and stores.

num_samples_per_chunk()

std::size_t hi::v1::audio_sample_format::num_samples_per_chunk ( std::size_t stride ) const

inlinenoexcept

The number of packed samples that are handled in a single 128 bit load or store.

Always one of: 1, 2 or 4.

pack_multiplier()

float hi::v1::audio_sample_format::pack_multiplier ( ) const

inlinenoexcept

How much to multiply float samples to create integer samples.

store_shuffle_indices()

i8x16 hi::v1::audio_sample_format::store_shuffle_indices ( std::size_t stride ) const

inlinenoexcept

Return a shuffle indices for storing 32 bit samples into packed samples.

unpack_multiplier()

float hi::v1::audio_sample_format::unpack_multiplier ( ) const

inlinenoexcept

How much to multiply integer samples to create float samples.

Field Documentation

endian

std::endian hi::v1::audio_sample_format::endian

The endian order of the bytes in the container.

is_float

bool hi::v1::audio_sample_format::is_float

The numeric type is floating point.

Otherwise it is a signed integer or fixed point number.

num_bits

uint8_t hi::v1::audio_sample_format::num_bits

The number of significant bits of the sample format.

This value is excluding the sign. (1 << num_bits) - 1 is the maximum sample value.

Examples:

• 16 bit signed PCM -> num_bits=15
• 24 bit signed PCM -> num_bits=23
• float PCM -> num_bits=23
• Q8.23 PCM -> num_bits=23

num_bytes

uint8_t hi::v1::audio_sample_format::num_bytes

The number of bytes of the container.

Must be either 1, 2, 3 or 4

num_guard_bits

uint8_t hi::v1::audio_sample_format::num_guard_bits

The number of bits used for the integer part of a fixed point number.

This value is zero for signed integer and float samples.

---

The documentation for this struct was generated from the following file:

• hikogui/src/hikogui/audio/audio_sample_format.hpp