The DXT/BC block codec. More...

#include <wowlib/formats/blp/codec.hpp>

Public Member Functions
	DxtCodec (PixelFormat format) noexcept
std::size_t	blockBytes () const noexcept
	The byte size of one compressed block (8 for BC1, 16 otherwise).
std::size_t	encodedSize (std::uint32_t width, std::uint32_t height) const noexcept
	The exact payload size of a level per the block formula ceil(w/4) * ceil(h/4) * blockBytes().
Result< std::vector< std::uint8_t > >	decode (std::span< const std::byte > src, std::uint32_t width, std::uint32_t height) const
	Decode one level's block stream to RGBA8 pixels.
Result< FileBuffer >	encode (std::span< const std::uint8_t > rgba, std::uint32_t width, std::uint32_t height) const
	Encode RGBA8 pixels to one level's block stream, exactly encodedSize() bytes.

Static Protected Member Functions
static void	_decodeColorBlock (const std::uint8_t *block, BlockPixels &out, bool punchThrough) noexcept
	Decode one BC1 color block (8 bytes) into 16 RGBA pixels.
static void	_decodeExplicitAlpha (const std::uint8_t *block, BlockPixels &out) noexcept
	Decode one BC2 explicit-alpha block (8 bytes, 4 bits per texel) onto the alpha channel of out.
static void	_decodeInterpolatedAlpha (const std::uint8_t *block, std::span< std::uint8_t, 16 > out) noexcept
	Decode one BC3/BC4 interpolated-alpha block (8 bytes) into 16 values.
static void	_encodePunchThroughBlock (const BlockPixels &pixels, std::uint8_t *dest) noexcept
	Encode one BC1 punch-through block (color0 <= color1 3-color mode, alpha < 128 texels as index 3) — the mode stb_dxt does not emit.
static BlockPixels	_extractBlock (std::span< const std::uint8_t > rgba, std::uint32_t width, std::uint32_t height, std::uint32_t bx, std::uint32_t by) noexcept
	Extract one 4x4 block from a level, clamping reads at the right/bottom edges (partial blocks repeat their last row/column, matching what the compressors expect).

Detailed Description

The DXT/BC block codec.

Constructed with the resolved block format (its state — Dxt1, Dxt3, Dxt5 or Bc5); decode()/encode() move a whole level between its block stream and RGBA8 pixels. Tolerates the known client quirk of understated mip sizes by decoding short streams into transparent-black padding.

Definition at line 40 of file codec.hpp.

Constructor & Destructor Documentation

◆ DxtCodec()

wowlib::formats::blp::detail::DxtCodec::DxtCodec ( PixelFormat format )

inlineexplicitnoexcept

Parameters

format the resolved block format (Dxt1, Dxt3, Dxt5 or Bc5).

Definition at line 43 of file codec.hpp.

Member Function Documentation

◆ _decodeColorBlock()

void wowlib::formats::blp::detail::DxtCodec::_decodeColorBlock	(	const std::uint8_t *	block,
		BlockPixels &	out,
		bool	punchThrough )

staticprotectednoexcept

Decode one BC1 color block (8 bytes) into 16 RGBA pixels.

Parameters

block	the 8 color-block bytes.
out	the 16 decoded pixels.
punchThrough	whether the color0 <= color1 3-color mode maps index 3 to transparent black (BC1 semantics; BC2/ BC3 color blocks always decode 4-color).

Definition at line 60 of file blp.cpp.

Referenced by decode().

◆ _decodeExplicitAlpha()

void wowlib::formats::blp::detail::DxtCodec::_decodeExplicitAlpha	(	const std::uint8_t *	block,
		BlockPixels &	out )

staticprotectednoexcept

Decode one BC2 explicit-alpha block (8 bytes, 4 bits per texel) onto the alpha channel of out.

Parameters

block	the 8 alpha bytes.
out	the 16 pixels to stamp alpha into.

Definition at line 96 of file blp.cpp.

Referenced by decode().

◆ _decodeInterpolatedAlpha()

void wowlib::formats::blp::detail::DxtCodec::_decodeInterpolatedAlpha	(	const std::uint8_t *	block,
		std::span< std::uint8_t, 16 >	out )

staticprotectednoexcept

Decode one BC3/BC4 interpolated-alpha block (8 bytes) into 16 values.

Parameters

block	the 8 alpha-block bytes.
out	the 16 decoded 8-bit values, row-major.

Definition at line 103 of file blp.cpp.

Referenced by decode().

◆ _encodePunchThroughBlock()

void wowlib::formats::blp::detail::DxtCodec::_encodePunchThroughBlock	(	const BlockPixels &	pixels,
		std::uint8_t *	dest )

staticprotectednoexcept

Encode one BC1 punch-through block (color0 <= color1 3-color mode, alpha < 128 texels as index 3) — the mode stb_dxt does not emit.

Parameters

pixels	the 16 RGBA texels.
dest	the 8 output bytes.

Definition at line 139 of file blp.cpp.

Referenced by encode().

◆ _extractBlock()

BlockPixels wowlib::formats::blp::detail::DxtCodec::_extractBlock	(	std::span< const std::uint8_t >	rgba,
		std::uint32_t	width,
		std::uint32_t	height,
		std::uint32_t	bx,
		std::uint32_t	by )

staticprotectednoexcept

Extract one 4x4 block from a level, clamping reads at the right/bottom edges (partial blocks repeat their last row/column, matching what the compressors expect).

Parameters

rgba	the level's RGBA bytes.
width	the level width in pixels.
height	the level height in pixels.
bx	the block column.
by	the block row.

Returns: the 16 RGBA texels.

Definition at line 122 of file blp.cpp.

Referenced by encode().

◆ blockBytes()

std::size_t wowlib::formats::blp::detail::DxtCodec::blockBytes ( ) const

inlinenodiscardnoexcept

The byte size of one compressed block (8 for BC1, 16 otherwise).

Returns: the block byte size.

Definition at line 48 of file codec.hpp.

References wowlib::formats::blp::Dxt1.

Referenced by decode(), encode(), and encodedSize().

◆ decode()

Result< std::vector< std::uint8_t > > wowlib::formats::blp::detail::DxtCodec::decode	(	std::span< const std::byte >	src,
		std::uint32_t	width,
		std::uint32_t	height ) const

nodiscard

Decode one level's block stream to RGBA8 pixels.

Parameters

src	the on-disk block bytes; may be shorter than encodedSize() (a known client quirk) — missing blocks decode transparent black.
width	the level width in pixels.
height	the level height in pixels.

Returns: width * height * 4 RGBA bytes.

Definition at line 207 of file blp.cpp.

References _decodeColorBlock(), _decodeExplicitAlpha(), _decodeInterpolatedAlpha(), wowlib::formats::blp::Bc5, blockBytes(), wowlib::formats::blp::Dxt1, wowlib::formats::blp::Dxt3, wowlib::formats::blp::Dxt5, wowlib::makeError(), and wowlib::NotSupported.

Referenced by wowlib::formats::blp::BLP::decode().

◆ encode()

Result< FileBuffer > wowlib::formats::blp::detail::DxtCodec::encode	(	std::span< const std::uint8_t >	rgba,
		std::uint32_t	width,
		std::uint32_t	height ) const

nodiscard

Encode RGBA8 pixels to one level's block stream, exactly encodedSize() bytes.

Parameters

rgba	width * height * 4 RGBA bytes.
width	the level width in pixels.
height	the level height in pixels.

Returns: the compressed block bytes.

Definition at line 264 of file blp.cpp.

References _encodePunchThroughBlock(), _extractBlock(), wowlib::formats::blp::Bc5, blockBytes(), wowlib::formats::blp::Dxt1, wowlib::formats::blp::Dxt3, wowlib::formats::blp::Dxt5, wowlib::makeError(), and wowlib::NotSupported.

Referenced by wowlib::formats::blp::BLP::encode().

◆ encodedSize()

std::size_t wowlib::formats::blp::detail::DxtCodec::encodedSize	(	std::uint32_t	width,
		std::uint32_t	height ) const

inlinenodiscardnoexcept

The exact payload size of a level per the block formula ceil(w/4) * ceil(h/4) * blockBytes().

Parameters

width	the level width in pixels.
height	the level height in pixels.

Returns: the encoded byte count.

Definition at line 58 of file codec.hpp.

References blockBytes().

Referenced by wowlib::formats::blp::BLP::validate().

The documentation for this class was generated from the following files:

src/wowlib/formats/blp/codec.hpp
src/wowlib/formats/blp/blp.cpp

Public Member Functions

Static Protected Member Functions

Detailed Description

Constructor & Destructor Documentation

◆ DxtCodec()

Member Function Documentation

◆ _decodeColorBlock()

◆ _decodeExplicitAlpha()

◆ _decodeInterpolatedAlpha()

◆ _encodePunchThroughBlock()

◆ _extractBlock()

◆ blockBytes()

◆ decode()

◆ encode()

◆ encodedSize()