dany/libwebp
1
0
Fork 0
mirror of https://github.com/webmproject/libwebp.git synced 2025-04-04 16:06:49 +02:00
Code Issues Actions Packages Projects Releases Wiki Activity

webp-lossless-bitstream-spec,cosmetics: reflow paragraphs

Browse Source 
Change-Id: Ifd7472fe0678b45efc62faa66de8e8dc2e9931e3
This commit is contained in:
James Zern 2022-11-22 10:55:38 -08:00
parent 0ceeeab987
commit e5fe2cfc1b
1 changed files with 227 additions and 251 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

468
doc/webp-lossless-bitstream-spec.txt
View File

@ -1,8 +1,8 @@
<!-- <!--
Although you may be viewing an alternate representation, this document Although you may be viewing an alternate representation, this document is
is sourced in Markdown, a light-duty markup scheme, and is optimized for sourced in Markdown, a light-duty markup scheme, and is optimized for the
the [kramdown](https://kramdown.gettalong.org/) transformer. [kramdown](https://kramdown.gettalong.org/) transformer.
See the accompanying specs_generation.md. External link targets are referenced See the accompanying specs_generation.md. External link targets are referenced
at the end of this file. at the end of this file.
@ -27,10 +27,10 @@ WebP lossless is an image format for lossless compression of ARGB images. The
lossless format stores and restores the pixel values exactly, including the lossless format stores and restores the pixel values exactly, including the
color values for pixels whose alpha value is 0. The format uses subresolution color values for pixels whose alpha value is 0. The format uses subresolution
images, recursively embedded into the format itself, for storing statistical images, recursively embedded into the format itself, for storing statistical
data about the images, such as the used entropy codes, spatial predictors, data about the images, such as the used entropy codes, spatial predictors, color
color space conversion, and color table. LZ77, prefix coding, and a color cache space conversion, and color table. LZ77, prefix coding, and a color cache are
are used for compression of the bulk data. Decoding speeds faster than PNG have used for compression of the bulk data. Decoding speeds faster than PNG have been
been demonstrated, as well as 25% denser compression than can be achieved using demonstrated, as well as 25% denser compression than can be achieved using
today's PNG format. today's PNG format.
@ -41,18 +41,18 @@ today's PNG format.
1 Introduction 1 Introduction
-------------- --------------
This document describes the compressed data representation of a WebP This document describes the compressed data representation of a WebP lossless
lossless image. It is intended as a detailed reference for the WebP lossless image. It is intended as a detailed reference for the WebP lossless encoder and
encoder and decoder implementation. decoder implementation.
In this document, we extensively use C programming language syntax to In this document, we extensively use C programming language syntax to describe
describe the bitstream, and assume the existence of a function for the bitstream, and assume the existence of a function for reading bits,
reading bits, `ReadBits(n)`. The bytes are read in the natural order of `ReadBits(n)`. The bytes are read in the natural order of the stream containing
the stream containing them, and bits of each byte are read in them, and bits of each byte are read in least-significant-bit-first order. When
least-significant-bit-first order. When multiple bits are read at the multiple bits are read at the same time, the integer is constructed from the
same time, the integer is constructed from the original data in the original data in the original order. The most significant bits of the returned
original order. The most significant bits of the returned integer are integer are also the most significant bits of the original data. Thus, the
also the most significant bits of the original data. Thus, the statement statement
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
b = ReadBits(2); b = ReadBits(2);
@ -66,20 +66,18 @@ b |= ReadBits(1) << 1;
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
We assume that each color component (e.g. alpha, red, blue and green) is We assume that each color component (e.g. alpha, red, blue and green) is
represented using an 8-bit byte. We define the corresponding type as represented using an 8-bit byte. We define the corresponding type as uint8. A
uint8. A whole ARGB pixel is represented by a type called uint32, an whole ARGB pixel is represented by a type called uint32, an unsigned integer
unsigned integer consisting of 32 bits. In the code showing the behavior consisting of 32 bits. In the code showing the behavior of the transformations,
of the transformations, alpha value is codified in bits 31..24, red in alpha value is codified in bits 31..24, red in bits 23..16, green in bits 15..8
bits 23..16, green in bits 15..8 and blue in bits 7..0, but and blue in bits 7..0, but implementations of the format are free to use another
implementations of the format are free to use another representation representation internally.
internally.
Broadly, a WebP lossless image contains header data, transform Broadly, a WebP lossless image contains header data, transform information and
information and actual image data. Headers contain width and height of actual image data. Headers contain width and height of the image. A WebP
the image. A WebP lossless image can go through four different types of lossless image can go through four different types of transformation before
transformation before being entropy encoded. The transform information being entropy encoded. The transform information in the bitstream contains the
in the bitstream contains the data required to apply the respective data required to apply the respective inverse transforms.
inverse transforms.
2 Nomenclature 2 Nomenclature
@ -92,53 +90,52 @@ ARGB image
: A two-dimensional array containing ARGB pixels. : A two-dimensional array containing ARGB pixels.
color cache color cache
: A small hash-addressed array to store recently used colors, to be able : A small hash-addressed array to store recently used colors, to be able to
to recall them with shorter codes. recall them with shorter codes.
color indexing image color indexing image
: A one-dimensional image of colors that can be indexed using a small : A one-dimensional image of colors that can be indexed using a small integer
integer (up to 256 within WebP lossless). (up to 256 within WebP lossless).
color transform image color transform image
: A two-dimensional subresolution image containing data about : A two-dimensional subresolution image containing data about correlations of
correlations of color components. color components.
distance mapping distance mapping
: Changes LZ77 distances to have the smallest values for pixels in 2D : Changes LZ77 distances to have the smallest values for pixels in 2D
proximity. proximity.
entropy image entropy image
: A two-dimensional subresolution image indicating which entropy coding : A two-dimensional subresolution image indicating which entropy coding should
should be used in a respective square in the image, i.e., each pixel be used in a respective square in the image, i.e., each pixel is a meta
is a meta prefix code. prefix code.
prefix code prefix code
: A classic way to do entropy coding where a smaller number of bits are : A classic way to do entropy coding where a smaller number of bits are used
used for more frequent codes. for more frequent codes.
LZ77 LZ77
: Dictionary-based sliding window compression algorithm that either : Dictionary-based sliding window compression algorithm that either emits
emits symbols or describes them as sequences of past symbols. symbols or describes them as sequences of past symbols.
meta prefix code meta prefix code
: A small integer (up to 16 bits) that indexes an element in the meta : A small integer (up to 16 bits) that indexes an element in the meta prefix
prefix table. table.
predictor image predictor image
: A two-dimensional subresolution image indicating which spatial : A two-dimensional subresolution image indicating which spatial predictor is
predictor is used for a particular square in the image. used for a particular square in the image.
prefix coding prefix coding
: A way to entropy code larger integers that codes a few bits of the : A way to entropy code larger integers that codes a few bits of the integer
integer using an entropy code and codifies the remaining bits raw. using an entropy code and codifies the remaining bits raw. This allows for
This allows for the descriptions of the entropy codes to remain the descriptions of the entropy codes to remain relatively small even when
relatively small even when the range of symbols is large. the range of symbols is large.
scan-line order scan-line order
: A processing order of pixels, left-to-right, top-to-bottom, starting : A processing order of pixels, left-to-right, top-to-bottom, starting from
from the left-hand-top pixel, proceeding to the right. Once a row is the left-hand-top pixel, proceeding to the right. Once a row is completed,
completed, continue from the left-hand column of the next row. continue from the left-hand column of the next row.
3 RIFF Header 3 RIFF Header
------------- -------------
@ -157,27 +154,26 @@ following 21 bytes:
lossless stream. lossless stream.
6. One byte signature 0x2f. 6. One byte signature 0x2f.
The first 28 bits of the bitstream specify the width and height of the The first 28 bits of the bitstream specify the width and height of the image.
image. Width and height are decoded as 14-bit integers as follows: Width and height are decoded as 14-bit integers as follows:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
int image_width = ReadBits(14) + 1; int image_width = ReadBits(14) + 1;
int image_height = ReadBits(14) + 1; int image_height = ReadBits(14) + 1;
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The 14-bit dynamics for image size limit the maximum size of a WebP The 14-bit dynamics for image size limit the maximum size of a WebP lossless
lossless image to 16384✕16384 pixels. image to 16384✕16384 pixels.
The alpha_is_used bit is a hint only, and should not impact decoding. The alpha_is_used bit is a hint only, and should not impact decoding. It should
It should be set to 0 when all alpha values are 255 in the picture, and be set to 0 when all alpha values are 255 in the picture, and 1 otherwise.
1 otherwise.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
int alpha_is_used = ReadBits(1); int alpha_is_used = ReadBits(1);
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The version_number is a 3 bit code that must be set to 0. Any other value The version_number is a 3 bit code that must be set to 0. Any other value should
should be treated as an error. \[AMENDED\] be treated as an error. \[AMENDED\]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
int version_number = ReadBits(3); int version_number = ReadBits(3);
@ -187,19 +183,18 @@ int version_number = ReadBits(3);
4 Transformations 4 Transformations
----------------- -----------------
Transformations are reversible manipulations of the image data that can Transformations are reversible manipulations of the image data that can reduce
reduce the remaining symbolic entropy by modeling spatial and color the remaining symbolic entropy by modeling spatial and color correlations.
correlations. Transformations can make the final compression more dense. Transformations can make the final compression more dense.
An image can go through four types of transformation. A 1 bit indicates An image can go through four types of transformation. A 1 bit indicates the
the presence of a transform. Each transform is allowed to be used only presence of a transform. Each transform is allowed to be used only once. The
once. The transformations are used only for the main level ARGB image: transformations are used only for the main level ARGB image: the subresolution
the subresolution images have no transforms, not even the 0 bit images have no transforms, not even the 0 bit indicating the end-of-transforms.
indicating the end-of-transforms.
Typically, an encoder would use these transforms to reduce the Shannon Typically, an encoder would use these transforms to reduce the Shannon entropy
entropy in the residual image. Also, the transform data can be decided in the residual image. Also, the transform data can be decided based on entropy
based on entropy minimization. minimization.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
while (ReadBits(1)) { // Transform present. while (ReadBits(1)) { // Transform present.
@ -212,8 +207,8 @@ while (ReadBits(1)) { // Transform present.
// Decode actual image data (Section 4). // Decode actual image data (Section 4).
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
If a transform is present then the next two bits specify the transform If a transform is present then the next two bits specify the transform type.
type. There are four types of transforms. There are four types of transforms.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
enum TransformType { enum TransformType {
@ -224,25 +219,23 @@ enum TransformType {
}; };
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The transform type is followed by the transform data. Transform data The transform type is followed by the transform data. Transform data contains
contains the information required to apply the inverse transform and the information required to apply the inverse transform and depends on the
depends on the transform type. Next we describe the transform data for transform type. Next we describe the transform data for different types.
different types.
### 4.1 Predictor Transform ### 4.1 Predictor Transform
The predictor transform can be used to reduce entropy by exploiting the The predictor transform can be used to reduce entropy by exploiting the fact
fact that neighboring pixels are often correlated. In the predictor that neighboring pixels are often correlated. In the predictor transform, the
transform, the current pixel value is predicted from the pixels already current pixel value is predicted from the pixels already decoded (in scan-line
decoded (in scan-line order) and only the residual value (actual - order) and only the residual value (actual - predicted) is encoded. The
predicted) is encoded. The _prediction mode_ determines the type of _prediction mode_ determines the type of prediction to use. We divide the image
prediction to use. We divide the image into squares and all the pixels into squares and all the pixels in a square use the same prediction mode.
in a square use the same prediction mode.
The first 3 bits of prediction data define the block width and height in The first 3 bits of prediction data define the block width and height in number
number of bits. The number of block columns, `block_xsize`, is used in of bits. The number of block columns, `block_xsize`, is used in indexing
indexing two-dimensionally. two-dimensionally.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
int size_bits = ReadBits(3) + 2; int size_bits = ReadBits(3) + 2;
@ -252,26 +245,24 @@ int block_height = (1 << size_bits);
int block_xsize = DIV_ROUND_UP(image_width, 1 << size_bits); int block_xsize = DIV_ROUND_UP(image_width, 1 << size_bits);
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The transform data contains the prediction mode for each block of the The transform data contains the prediction mode for each block of the image. All
image. All the `block_width * block_height` pixels of a block use same the `block_width * block_height` pixels of a block use same prediction mode. The
prediction mode. The prediction modes are treated as pixels of an image prediction modes are treated as pixels of an image and encoded using the same
and encoded using the same techniques described in techniques described in [Chapter 5](#image-data).
[Chapter 5](#image-data).
For a pixel _x, y_, one can compute the respective filter block address For a pixel _x, y_, one can compute the respective filter block address by:
by:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
int block_index = (y >> size_bits) * block_xsize + int block_index = (y >> size_bits) * block_xsize +
(x >> size_bits); (x >> size_bits);
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
There are 14 different prediction modes. In each prediction mode, the There are 14 different prediction modes. In each prediction mode, the current
current pixel value is predicted from one or more neighboring pixels pixel value is predicted from one or more neighboring pixels whose values are
whose values are already known. already known.
We choose the neighboring pixels (TL, T, TR, and L) of the current pixel We choose the neighboring pixels (TL, T, TR, and L) of the current pixel (P) as
(P) as follows: follows:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
O O O O O O O O O O O O O O O O O O O O O O
@ -282,12 +273,12 @@ X X X X X X X X X X X
X X X X X X X X X X X X X X X X X X X X X X
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
where TL means top-left, T top, TR top-right, L left pixel. where TL means top-left, T top, TR top-right, L left pixel. At the time of
At the time of predicting a value for P, all pixels O, TL, T, TR and L predicting a value for P, all pixels O, TL, T, TR and L have already been
have been already processed, and pixel P and all pixels X are unknown. processed, and pixel P and all pixels X are unknown.
Given the above neighboring pixels, the different prediction modes are Given the above neighboring pixels, the different prediction modes are defined
defined as follows. as follows.
| Mode | Predicted value of each channel of the current pixel | | Mode | Predicted value of each channel of the current pixel |
| ------ | ------------------------------------------------------- | | ------ | ------------------------------------------------------- |
@ -342,8 +333,8 @@ uint32 Select(uint32 L, uint32 T, uint32 TL) {
} }
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The functions `ClampAddSubtractFull` and `ClampAddSubtractHalf` are The functions `ClampAddSubtractFull` and `ClampAddSubtractHalf` are performed
performed for each ARGB component as follows: for each ARGB component as follows:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
// Clamp the input value between 0 and 255. // Clamp the input value between 0 and 255.
@ -365,30 +356,28 @@ int ClampAddSubtractHalf(int a, int b) {
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
There are special handling rules for some border pixels. If there is a There are special handling rules for some border pixels. If there is a
prediction transform, regardless of the mode \[0..13\] for these pixels, prediction transform, regardless of the mode \[0..13\] for these pixels, the
the predicted value for the left-topmost pixel of the image is predicted value for the left-topmost pixel of the image is 0xff000000, L-pixel
0xff000000, L-pixel for all pixels on the top row, and T-pixel for all for all pixels on the top row, and T-pixel for all pixels on the leftmost
pixels on the leftmost column. column.
\[AMENDED2\] \[AMENDED2\] Addressing the TR-pixel for pixels on the rightmost column is
Addressing the TR-pixel for pixels on the rightmost column is exceptional. The pixels on the rightmost column are predicted by using the modes
exceptional. The pixels on the rightmost column are predicted by using \[0..13\] just like pixels not on the border, but the leftmost pixel on the same
the modes \[0..13\] just like pixels not on the border, but the leftmost pixel row as the current pixel is instead used as the TR-pixel.
on the same row as the current pixel is instead used as the TR-pixel.
### 4.2 Color Transform ### 4.2 Color Transform
\[AMENDED2\] \[AMENDED2\]
The goal of the color transform is to decorrelate the R, G and B values The goal of the color transform is to decorrelate the R, G and B values of each
of each pixel. The color transform keeps the green (G) value as it is, pixel. The color transform keeps the green (G) value as it is, transforms red
transforms red (R) based on green and transforms blue (B) based on green (R) based on green and transforms blue (B) based on green and then based on red.
and then based on red.
As is the case for the predictor transform, first the image is divided As is the case for the predictor transform, first the image is divided into
into blocks and the same transform mode is used for all the pixels in a blocks and the same transform mode is used for all the pixels in a block. For
block. For each block there are three types of color transform elements. each block there are three types of color transform elements.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
typedef struct { typedef struct {
@ -398,11 +387,10 @@ typedef struct {
} ColorTransformElement; } ColorTransformElement;
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The actual color transformation is done by defining a color transform The actual color transformation is done by defining a color transform delta. The
delta. The color transform delta depends on the `ColorTransformElement`, color transform delta depends on the `ColorTransformElement`, which is the same
which is the same for all the pixels in a particular block. The delta is for all the pixels in a particular block. The delta is subtracted during the
subtracted during the color transform. The inverse color transform then is just color transform. The inverse color transform then is just adding those deltas.
adding those deltas.
The color transform function is defined as follows: The color transform function is defined as follows:
@ -424,9 +412,9 @@ void ColorTransform(uint8 red, uint8 blue, uint8 green,
} }
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
`ColorTransformDelta` is computed using a signed 8-bit integer `ColorTransformDelta` is computed using a signed 8-bit integer representing a
representing a 3.5-fixed-point number, and a signed 8-bit RGB color 3.5-fixed-point number, and a signed 8-bit RGB color channel (c) \[-128..127\]
channel (c) \[-128..127\] and is defined as follows: and is defined as follows:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
int8 ColorTransformDelta(int8 t, int8 c) { int8 ColorTransformDelta(int8 t, int8 c) {
@ -434,22 +422,20 @@ int8 ColorTransformDelta(int8 t, int8 c) {
} }
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
A conversion from the 8-bit unsigned representation (uint8) to the 8-bit A conversion from the 8-bit unsigned representation (uint8) to the 8-bit signed
signed one (int8) is required before calling `ColorTransformDelta()`. one (int8) is required before calling `ColorTransformDelta()`. It should be
It should be performed using 8-bit two's complement (that is: uint8 range performed using 8-bit two's complement (that is: uint8 range \[128..255\] is
\[128..255\] is mapped to the \[-128..-1\] range of its converted int8 value). mapped to the \[-128..-1\] range of its converted int8 value).
The multiplication is to be done using more precision (with at least The multiplication is to be done using more precision (with at least 16-bit
16-bit dynamics). The sign extension property of the shift operation dynamics). The sign extension property of the shift operation does not matter
does not matter here: only the lowest 8 bits are used from the result, here: only the lowest 8 bits are used from the result, and there the sign
and there the sign extension shifting and unsigned shifting are extension shifting and unsigned shifting are consistent with each other.
consistent with each other.
Now we describe the contents of color transform data so that decoding Now we describe the contents of color transform data so that decoding can apply
can apply the inverse color transform and recover the original red and the inverse color transform and recover the original red and blue values. The
blue values. The first 3 bits of the color transform data contain the first 3 bits of the color transform data contain the width and height of the
width and height of the image block in number of bits, just like the image block in number of bits, just like the predictor transform:
predictor transform:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
int size_bits = ReadBits(3) + 2; int size_bits = ReadBits(3) + 2;
@ -457,17 +443,15 @@ int block_width = 1 << size_bits;
int block_height = 1 << size_bits; int block_height = 1 << size_bits;
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The remaining part of the color transform data contains The remaining part of the color transform data contains `ColorTransformElement`
`ColorTransformElement` instances corresponding to each block of the instances corresponding to each block of the image. `ColorTransformElement`
image. `ColorTransformElement` instances are treated as pixels of an instances are treated as pixels of an image and encoded using the methods
image and encoded using the methods described in described in [Chapter 5](#image-data).
[Chapter 5](#image-data).
During decoding, `ColorTransformElement` instances of the blocks are During decoding, `ColorTransformElement` instances of the blocks are decoded and
decoded and the inverse color transform is applied on the ARGB values of the inverse color transform is applied on the ARGB values of the pixels. As
the pixels. As mentioned earlier, that inverse color transform is just mentioned earlier, that inverse color transform is just adding
adding `ColorTransformElement` values to the red and blue `ColorTransformElement` values to the red and blue channels. \[AMENDED3\]
channels. \[AMENDED3\]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
void InverseTransform(uint8 red, uint8 green, uint8 blue, void InverseTransform(uint8 red, uint8 green, uint8 blue,
@ -492,11 +476,10 @@ void InverseTransform(uint8 red, uint8 green, uint8 blue,
### 4.3 Subtract Green Transform ### 4.3 Subtract Green Transform
The subtract green transform subtracts green values from red and blue The subtract green transform subtracts green values from red and blue values of
values of each pixel. When this transform is present, the decoder needs each pixel. When this transform is present, the decoder needs to add the green
to add the green value to both red and blue. There is no data associated value to both red and blue. There is no data associated with this transform. The
with this transform. The decoder applies the inverse transform as decoder applies the inverse transform as follows:
follows:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
void AddGreenToBlueAndRed(uint8 green, uint8 *red, uint8 *blue) { void AddGreenToBlueAndRed(uint8 green, uint8 *red, uint8 *blue) {
@ -505,53 +488,47 @@ void AddGreenToBlueAndRed(uint8 green, uint8 *red, uint8 *blue) {
} }
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
This transform is redundant as it can be modeled using the color This transform is redundant as it can be modeled using the color transform, but
transform, but it is still often useful. Since it can extend the it is still often useful. Since it can extend the dynamics of the color
dynamics of the color transform and there is no additional data here, transform and there is no additional data here, the subtract green transform can
the subtract green transform can be coded using fewer bits than a be coded using fewer bits than a full-blown color transform.
full-blown color transform.
### 4.4 Color Indexing Transform ### 4.4 Color Indexing Transform
If there are not many unique pixel values, it may be more efficient to If there are not many unique pixel values, it may be more efficient to create a
create a color index array and replace the pixel values by the array's color index array and replace the pixel values by the array's indices. The color
indices. The color indexing transform achieves this. (In the context of indexing transform achieves this. (In the context of WebP lossless, we
WebP lossless, we specifically do not call this a palette transform specifically do not call this a palette transform because a similar but more
because a similar but more dynamic concept exists in WebP lossless dynamic concept exists in WebP lossless encoding: color cache).
encoding: color cache).
The color indexing transform checks for the number of unique ARGB values The color indexing transform checks for the number of unique ARGB values in the
in the image. If that number is below a threshold (256), it creates an image. If that number is below a threshold (256), it creates an array of those
array of those ARGB values, which is then used to replace the pixel ARGB values, which is then used to replace the pixel values with the
values with the corresponding index: the green channel of the pixels are corresponding index: the green channel of the pixels are replaced with the
replaced with the index; all alpha values are set to 255; all red and index; all alpha values are set to 255; all red and blue values to 0.
blue values to 0.
The transform data contains color table size and the entries in the The transform data contains color table size and the entries in the color table.
color table. The decoder reads the color indexing transform data as The decoder reads the color indexing transform data as follows:
follows:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
// 8 bit value for color table size // 8 bit value for color table size
int color_table_size = ReadBits(8) + 1; int color_table_size = ReadBits(8) + 1;
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The color table is stored using the image storage format itself. The The color table is stored using the image storage format itself. The color table
color table can be obtained by reading an image, without the RIFF can be obtained by reading an image, without the RIFF header, image size, and
header, image size, and transforms, assuming a height of one pixel and transforms, assuming a height of one pixel and a width of `color_table_size`.
a width of `color_table_size`. The color table is always The color table is always subtraction-coded to reduce image entropy. The deltas
subtraction-coded to reduce image entropy. The deltas of palette colors of palette colors contain typically much less entropy than the colors
contain typically much less entropy than the colors themselves, leading themselves, leading to significant savings for smaller images. In decoding,
to significant savings for smaller images. In decoding, every final every final color in the color table can be obtained by adding the previous
color in the color table can be obtained by adding the previous color color component values by each ARGB component separately, and storing the least
component values by each ARGB component separately, and storing the significant 8 bits of the result.
least significant 8 bits of the result.
The inverse transform for the image is simply replacing the pixel values The inverse transform for the image is simply replacing the pixel values (which
(which are indices to the color table) with the actual color table are indices to the color table) with the actual color table values. The indexing
values. The indexing is done based on the green component of the ARGB is done based on the green component of the ARGB color.
color.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
// Inverse transform // Inverse transform
@ -561,13 +538,12 @@ argb = color_table[GREEN(argb)];
If the index is equal or larger than `color_table_size`, the argb color value If the index is equal or larger than `color_table_size`, the argb color value
should be set to 0x00000000 (transparent black). \[AMENDED\] should be set to 0x00000000 (transparent black). \[AMENDED\]
When the color table is small (equal to or less than 16 colors), several When the color table is small (equal to or less than 16 colors), several pixels
pixels are bundled into a single pixel. The pixel bundling packs several are bundled into a single pixel. The pixel bundling packs several (2, 4, or 8)
(2, 4, or 8) pixels into a single pixel, reducing the image width pixels into a single pixel, reducing the image width respectively. Pixel
respectively. Pixel bundling allows for a more efficient joint bundling allows for a more efficient joint distribution entropy coding of
distribution entropy coding of neighboring pixels, and gives some neighboring pixels, and gives some arithmetic coding-like benefits to the
arithmetic coding-like benefits to the entropy code, but it can only be entropy code, but it can only be used when there are 16 or fewer unique values.
used when there are 16 or fewer unique values.
`color_table_size` specifies how many pixels are combined: `color_table_size` specifies how many pixels are combined:
@ -621,9 +597,9 @@ We use image data in five different roles:
[meta prefix codes](#decoding-of-meta-prefix-codes). The red and green [meta prefix codes](#decoding-of-meta-prefix-codes). The red and green
components of a pixel define the meta prefix code used in a particular components of a pixel define the meta prefix code used in a particular
block of the ARGB image. block of the ARGB image.
1. Predictor image: Stores the metadata for [Predictor 1. Predictor image: Stores the metadata for
Transform](#predictor-transform). The green component of a pixel defines [Predictor Transform](#predictor-transform). The green component of a pixel
which of the 14 predictors is used within a particular block of the defines which of the 14 predictors is used within a particular block of the
ARGB image. ARGB image.
1. Color transform image. It is created by `ColorTransformElement` values 1. Color transform image. It is created by `ColorTransformElement` values
(defined in [Color Transform](#color-transform)) for different blocks of (defined in [Color Transform](#color-transform)) for different blocks of
@ -683,8 +659,8 @@ while the extra bits are stored as they are (without an entropy code).
**Rationale**: This approach reduces the storage requirement for the entropy **Rationale**: This approach reduces the storage requirement for the entropy
code. Also, large values are usually rare, and so extra bits would be used for code. Also, large values are usually rare, and so extra bits would be used for
very few values in the image. Thus, this approach results in better very few values in the image. Thus, this approach results in better compression
compression overall. overall.
The following table denotes the prefix codes and extra bits used for storing The following table denotes the prefix codes and extra bits used for storing
different ranges of values. different ranges of values.
@ -709,8 +685,8 @@ values. For distance values, however, all the 40 prefix codes are valid.
| 524289..786432 | 38 | 18 | | 524289..786432 | 38 | 18 |
| 786433..1048576 | 39 | 18 | | 786433..1048576 | 39 | 18 |
The pseudocode to obtain a (length or distance) value from the prefix code is The pseudocode to obtain a (length or distance) value from the prefix code is as
as follows: follows:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
if (prefix_code < 4) { if (prefix_code < 4) {
@ -729,8 +705,8 @@ previously seen pixel, from which the pixels are to be copied. This subsection
defines the mapping between a distance code and the position of a previous defines the mapping between a distance code and the position of a previous
pixel. pixel.
Distance codes larger than 120 denote the pixel-distance in scan-line Distance codes larger than 120 denote the pixel-distance in scan-line order,
order, offset by 120. offset by 120.
The smallest distance codes \[1..120\] are special, and are reserved for a close The smallest distance codes \[1..120\] are special, and are reserved for a close
neighborhood of the current pixel. This neighborhood consists of 120 pixels: neighborhood of the current pixel. This neighborhood consists of 120 pixels:
@ -770,8 +746,8 @@ neighboring pixel, that is, the pixel above the current pixel (0 pixel
difference in the X-direction and 1 pixel difference in the Y-direction). difference in the X-direction and 1 pixel difference in the Y-direction).
Similarly, the distance code `3` indicates the left-top pixel. Similarly, the distance code `3` indicates the left-top pixel.
The decoder can convert a distance code `i` to a scan-line order distance The decoder can convert a distance code `i` to a scan-line order distance `dist`
`dist` as follows: as follows:
\[AMENDED3\] \[AMENDED3\]
@ -812,16 +788,16 @@ int color_cache_size = 1 << color_cache_code_bits;
`color_cache_code_bits` is \[1..11\]. Compliant decoders must indicate a `color_cache_code_bits` is \[1..11\]. Compliant decoders must indicate a
corrupted bitstream for other values. corrupted bitstream for other values.
A color cache is an array of size `color_cache_size`. Each entry A color cache is an array of size `color_cache_size`. Each entry stores one ARGB
stores one ARGB color. Colors are looked up by indexing them by color. Colors are looked up by indexing them by (0x1e35a7bd * `color`) >> (32 -
(0x1e35a7bd * `color`) >> (32 - `color_cache_code_bits`). Only one `color_cache_code_bits`). Only one lookup is done in a color cache; there is no
lookup is done in a color cache; there is no conflict resolution. conflict resolution.
In the beginning of decoding or encoding of an image, all entries in all In the beginning of decoding or encoding of an image, all entries in all color
color cache values are set to zero. The color cache code is converted to cache values are set to zero. The color cache code is converted to this color at
this color at decoding time. The state of the color cache is maintained decoding time. The state of the color cache is maintained by inserting every
by inserting every pixel, be it produced by backward referencing or as pixel, be it produced by backward referencing or as literals, into the cache in
literals, into the cache in the order they appear in the stream. the order they appear in the stream.
6 Entropy Code 6 Entropy Code
@ -871,9 +847,9 @@ stream. This may be inefficient, but it is allowed by the format.
\[AMENDED2\] \[AMENDED2\]
This variant is used in the special case when only 1 or 2 prefix symbols are This variant is used in the special case when only 1 or 2 prefix symbols are in
in the range \[0..255\] with code length `1`. All other prefix code lengths the range \[0..255\] with code length `1`. All other prefix code lengths are
are implicitly zeros. implicitly zeros.
The first bit indicates the number of symbols: The first bit indicates the number of symbols:
@ -882,10 +858,11 @@ int num_symbols = ReadBits(1) + 1;
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Following are the symbol values. Following are the symbol values.
This first symbol is coded using 1 or 8 bits depending on the value of This first symbol is coded using 1 or 8 bits depending on the value of
`is_first_8bits`. The range is \[0..1\] or \[0..255\], respectively. `is_first_8bits`. The range is \[0..1\] or \[0..255\], respectively. The second
The second symbol, if present, is always assumed to be in the range \[0..255\] symbol, if present, is always assumed to be in the range \[0..255\] and coded
and coded using 8 bits. using 8 bits.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
int is_first_8bits = ReadBits(1); int is_first_8bits = ReadBits(1);
@ -897,13 +874,12 @@ if (num_symbols == 2) {
} }
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
**Note:** Another special case is when _all_ prefix code lengths are _zeros_ **Note:** Another special case is when _all_ prefix code lengths are _zeros_ (an
(an empty prefix code). For example, a prefix code for distance can be empty empty prefix code). For example, a prefix code for distance can be empty if
if there are no backward references. Similarly, prefix codes for alpha, red, there are no backward references. Similarly, prefix codes for alpha, red, and
and blue can be empty if all pixels within the same meta prefix code are blue can be empty if all pixels within the same meta prefix code are produced
produced using the color cache. However, this case doesn't need special using the color cache. However, this case doesn't need special handling, as
handling, as empty prefix codes can be coded as those containing a single empty prefix codes can be coded as those containing a single symbol `0`.
symbol `0`.
**(ii) Normal Code Length Code:** **(ii) Normal Code Length Code:**
@ -940,8 +916,8 @@ int length_nbits = 2 + 2 * ReadBits(3);
int max_symbol = 2 + ReadBits(length_nbits); int max_symbol = 2 + ReadBits(length_nbits);
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
A prefix table is then built from `code_length_code_lengths` and used to read A prefix table is then built from `code_length_code_lengths` and used to read up
up to `max_symbol` code lengths. to `max_symbol` code lengths.
* Code \[0..15\] indicates literal code lengths. * Code \[0..15\] indicates literal code lengths.
* Value 0 means no symbols have been coded. * Value 0 means no symbols have been coded.
@ -964,8 +940,8 @@ distance) is formed using their respective alphabet sizes:
#### 6.2.2 Decoding of Meta Prefix Codes #### 6.2.2 Decoding of Meta Prefix Codes
As noted earlier, the format allows the use of different prefix codes for As noted earlier, the format allows the use of different prefix codes for
different blocks of the image. _Meta prefix codes_ are indexes identifying different blocks of the image. _Meta prefix codes_ are indexes identifying which
which prefix codes to use in different parts of the image. prefix codes to use in different parts of the image.
Meta prefix codes may be used _only_ when the image is being used in the Meta prefix codes may be used _only_ when the image is being used in the
[role](#roles-of-image-data) of an _ARGB image_. [role](#roles-of-image-data) of an _ARGB image_.
@ -1019,8 +995,8 @@ int num_prefix_groups = max(entropy image) + 1;
where `max(entropy image)` indicates the largest prefix code stored in the where `max(entropy image)` indicates the largest prefix code stored in the
entropy image. entropy image.
As each prefix code group contains five prefix codes, the total number of As each prefix code group contains five prefix codes, the total number of prefix
prefix codes is: codes is:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
int num_prefix_codes = 5 * num_prefix_groups; int num_prefix_codes = 5 * num_prefix_groups;
@ -1037,8 +1013,8 @@ PrefixCodeGroup prefix_group = prefix_code_groups[meta_prefix_code];
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
where, we have assumed the existence of `PrefixCodeGroup` structure, which where, we have assumed the existence of `PrefixCodeGroup` structure, which
represents a set of five prefix codes. Also, `prefix_code_groups` is an array represents a set of five prefix codes. Also, `prefix_code_groups` is an array of
of `PrefixCodeGroup` (of size `num_prefix_groups`). `PrefixCodeGroup` (of size `num_prefix_groups`).
The decoder then uses prefix code group `prefix_group` to decode the pixel The decoder then uses prefix code group `prefix_group` to decode the pixel
(x, y) as explained in the [next section](#decoding-entropy-coded-image-data). (x, y) as explained in the [next section](#decoding-entropy-coded-image-data).