dany/libwebp
1
0
Fork 0
mirror of https://github.com/webmproject/libwebp.git synced 2024-11-19 20:08:28 +01:00
Code Issues Actions Packages Projects Releases Wiki Activity

Update the WebP Container Spec.

Browse Source 
- Add details about the VP8L chunk support.
- Also add new example bitsteams containing VP8L chunk.
- Add back a section describing the VP8 chunk.
- Restrict some fields to 16 or 24 bits instead of 32 bits.
- Fields whose values are always positive are stored 1-based
instead of 0-based.
- Unknown chunks can only occur at certain places in the file.
- Remove the restriction for some fields to be divisible by 32 or 16.
  Instead they are restricted to be even.
- Add a restriction for (canvas_width * canvas_height) product.
- Add 3 bits for rotation & symmetry in VP8X flags.
- Add some new example layouts.
- Add/clarify some nitty-gritties throughout the doc.
- Use the terms frame/tile more consistently and logically.
- Update related TODOs.

Change-Id: I611c1f58ecc3ee87546ca31bad1609376fad251e
This commit is contained in:
Urvang Joshi 2012-06-21 16:03:20 -07:00
parent 31b68fe639
commit 7f8472a610
2 changed files with 281 additions and 211 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
doc/TODO
View File

@ -3,10 +3,6 @@
* Determine that normative RFC 2119 terms (MUST, SHOULD, MAY, etc.) are * Determine that normative RFC 2119 terms (MUST, SHOULD, MAY, etc.) are
truly intended in all cases where capitalized. truly intended in all cases where capitalized.
* Document hierarchy WRT headings has a flaw, in that topics related to
animated WebPs are discussed under subheads of "Single-image WebP
Files".
* Several passages could be made clearer. * Several passages could be made clearer.
* Overall edit for scope. Portions are phrased as an introduction to * Overall edit for scope. Portions are phrased as an introduction to
@ -15,10 +11,3 @@
* To wit, suggest s/[spec|specification]/guide/g . "Spec" can imply a * To wit, suggest s/[spec|specification]/guide/g . "Spec" can imply a
standards track; in any case it's too formal for a work in progress. standards track; in any case it's too formal for a work in progress.
* Sections and passages re "multi-image" should likely be suppressed
until multi-image drops.
* Improve the term "Mux-Container".

481
doc/webp-container-spec.txt
View File

@ -13,7 +13,7 @@ end of this file.
WebP Container Specification WebP Container Specification
============================ ============================
_Working Draft, v0.3, 20120523_ _Working Draft, v0.4, 20120613_
* TOC placeholder * TOC placeholder
@ -23,40 +23,42 @@ _Working Draft, v0.3, 20120523_
Introduction Introduction
------------ ------------
WebP is a still image format that uses the VP8 key frame encoding, and WebP is an image format that uses either (i) the VP8 key frame encoding
possibly other encodings in the future, to compress image data in a to compress image data in a lossy way, or (ii) the WebP lossless encoding
lossy way. The VP8 encoding should make it more efficient than currently (and possibly other encodings in the future). These encoding schemes should
used formats. It is optimized for fast image transfer over the network make it more efficient than currently used formats. It is optimized for fast
(e.g., for websites). However, it also aims for feature parity image transfer over the network (e.g., for websites). However, it also aims
(color profile, XMP metadata, animation, etc.) with other formats. This for feature parity (color profile, metadata, animation, etc.) with other
document describes the structure of a WebP file. formats. This document describes the structure of a WebP file.
The first version of WebP handled only the basic use case: a file The WebP container (i.e., RIFF container for WebP) allows feature support over
containing a single image (being one VP8 key frame), with no metadata. and above the basic use case of WebP (i.e., a file containing a single image
The use of a RIFF container permits additional feature support. This encoded as a VP8 key frame). The WebP container provides additional support
document describes additional support for: for:
* **Metadata and color profiles.** We specify chunks that can contain * **Lossless compression.** An image can be losslessly compressed, using the
this information, as other popular formats do. WebP Lossless Format.
* **Transparency.** An image may have transparency, i.e., an alpha channel
for each frame/tile.
* **Metadata.** An image can have metadata stored in any of the popular
metadata formats.
* **Color profiles.** An image can have an ICC profile characterizing a color
input or output device.
* **Animation.** An image may have pauses between frames, making it
an animation.
* **Tiling.** A single VP8 frame has an inherent limitation for width * **Tiling.** A single VP8 frame has an inherent limitation for width
or height of 2^14 pixels, and a 512kB limit on the size of the first or height of 2^14 pixels, and a 512 KiB limit on the size of the first
compressed partition. To support larger images, we support images compressed partition. To support larger images, we support images
that are composed of multiple tiles, each encoded as a separate VP8 that are composed of multiple tiles, each encoded as a separate VP8
frame. All tiles form logically a single image: they have common frame. All tiles form logically a single image: they have common
metadata, color profile, etc. Tiling may also improve efficiency for metadata, color profile, etc. Tiling may also improve efficiency for
larger images, e.g., grass can be encoded differently than sky. larger images, e.g., grass can be encoded differently than sky.
* **Animation.** An image may have pauses between frames, making it
an animation.
* **Transparency.** An image may have transparency, i.e., an alpha channel for
each frame/tile.
Files not using these new features are backward compatible with the
original format. Use of these features will produce files that are not
compatible with older programs.
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
document are to be interpreted as described in [RFC 2119][]. document are to be interpreted as described in [RFC 2119][].
@ -71,9 +73,9 @@ metadata, etc. In case we need to refer only to the matrix of pixels,
we will call it the _canvas_ of the image. we will call it the _canvas_ of the image.
The canvas of an image is built from one or multiple tiles. Each tile The canvas of an image is built from one or multiple tiles. Each tile
is a separately encoded VP8 key frame (other encodings are possible in is a separately encoded VP8 key frame or a WebP lossless bitstream. Building
the future). Building an image from several tiles allows us to overcome an image from several tiles allows the format to overcome the size limitations
the size limitations of a single VP8 frame. Tiles are an internal detail of a single VP8 frame / WebP lossless bitstream. Tiles are an internal detail
of the file: they are not supposed to be exposed to the user. of the file: they are not supposed to be exposed to the user.
Below are additional terms used throughout this document: Below are additional terms used throughout this document:
@ -85,10 +87,18 @@ _uint16_
: A 16-bit, little-endian, unsigned integer. : A 16-bit, little-endian, unsigned integer.
_uint24_
: A 24-bit, little-endian, unsigned integer.
_uint32_ _uint32_
: A 32-bit, little-endian, unsigned integer. : A 32-bit, little-endian, unsigned integer.
_1-based_
: An unsigned integer field storing values offset by `-1`. e.g., Such a field
would store value _25_ as _24_.
The basic element of a RIFF file is a _chunk_. It consists of: The basic element of a RIFF file is a _chunk_. It consists of:
0 1 2 3 0 1 2 3
@ -180,15 +190,28 @@ File Size: 32 bits (_uint32_)
: The ASCII characters 'W' 'E' 'B' 'P'. : The ASCII characters 'W' 'E' 'B' 'P'.
Simple file format Simple file format (lossy)
------------------ --------------------------
Simple WebP file header:
This layout SHOULD be used if the image requires _lossy_ encoding and does not
require transparency or other advanced features provided by the extended format.
Files with this layout are smaller and supported by older software.
Simple WebP (lossy) file header:
0 1 2 3 0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| WebP file header (12 bytes) | | WebP file header (12 bytes) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| VP8 chunk |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
VP8 chunk:
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| ChunkHeader('VP8 ') | | ChunkHeader('VP8 ') |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| VP8 data | | VP8 data |
@ -198,18 +221,50 @@ VP8 data: _Chunk Size_ bytes
: VP8 bitstream data. : VP8 bitstream data.
The content of a 'VP8 ' chunk (note the last character is a space) MUST be one The VP8 bitstream format specification can be found at [VP8 Data Format and
VP8 key frame (with optional padding). Decoding Guide][vp8spec]. Note that the VP8 frame header contains the VP8 frame
width and height. That is assumed to be the width and height of the canvas.
The current [VP8 Data Format and Decoding Guide][vp8spec] can be found
at the IETF website, <http://www.ietf.org/>.
The VP8 specification describes how to decode the image into Y'CbCr The VP8 specification describes how to decode the image into Y'CbCr
format. To convert to RGB, Rec. 601 SHOULD be used. format. To convert to RGB, Rec. 601 SHOULD be used.
This layout SHOULD be used if the image does not require advanced Simple file format (lossless)
features: color profiles, XMP metadata, animation or tiling. Files with -----------------------------
this layout are smaller and supported by older software.
**Note:** Older readers may not support files using the lossless format.
This layout SHOULD be used if the image requires _lossless_ encoding (with an
optional transparency channel) and does not require advanced features provided
by the extended format.
Simple WebP (lossless) file header:
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| WebP file header (12 bytes) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| VP8L chunk |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
VP8L chunk:
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| ChunkHeader('VP8L') |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| VP8L data |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
VP8L data: _Chunk Size_ bytes
: VP8L bitstream data.
The current specification of the VP8L bitstream can be found at
[WebP Lossless Bitstream Format][webpllspec]. Note that the VP8L header
contains the VP8L image width and height. That is assumed to be the width
and height of the canvas.
Extended file format Extended file format
-------------------- --------------------
@ -222,14 +277,14 @@ An extended format file consists of:
* An optional 'ICCP' chunk with color profile. * An optional 'ICCP' chunk with color profile.
* Optionally, some other unknown chunk types that may be defined by future
specifications.
* An optional 'LOOP' chunk with animation control data. * An optional 'LOOP' chunk with animation control data.
* Data for all the frames. * Data for all the frames.
* An optional 'META' chunk with XMP metadata. * An optional 'META' chunk with metadata.
* Some other chunk types may be defined by future specifications and
placed anywhere in the file.
As will be described in the 'VP8X' chunk description, by checking a As will be described in the 'VP8X' chunk description, by checking a
flag one can distinguish animated and non-animated images. A flag one can distinguish animated and non-animated images. A
@ -242,18 +297,21 @@ multiple frames. Data for each frame consists of:
present in non-animated images. present in non-animated images.
* An optional 'TILE' chunk with tile position metadata. It MUST be * An optional 'TILE' chunk with tile position metadata. It MUST be
present at the beginning of data for an image that's represented as present at the beginning of each tile for a frame that is represented as
multiple tile images. multiple tile images.
* An optional 'ALPH' chunk with alpha bitstream of the tile. It MUST be * An optional 'ALPH' chunk with alpha bitstream of the frame/tile. If the
present for an image containing transparency. It MUST NOT be present file contains transparency, this chunk MUST be present for each frame/tile
in non-transparent images. containing a 'VP8 ' chunk. It MUST NOT be present otherwise.
* A 'VP8 ' chunk with the bitstream of the tile. * A 'VP8 ' or a 'VP8L' chunk containing compressed image data of the
frame/tile.
All chunks SHOULD be placed in the same order as listed above (except * An optional unknown chunk type that may be defined by future
for unknown chunks, which MAY appear anywhere). If a chunk appears specifications.
in the wrong place, the file is invalid, but readers MAY parse the
All chunks SHOULD be placed in the same order as listed above. If a chunk
appears in the wrong place, the file is invalid, but readers MAY parse the
file, ignoring the chunks that come too late. file, ignoring the chunks that come too late.
**Rationale:** Setting the order of chunks should allow quicker file **Rationale:** Setting the order of chunks should allow quicker file
@ -271,12 +329,12 @@ Extended WebP file header:
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| ChunkHeader('VP8X') | | ChunkHeader('VP8X') |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Rsv |L|M|I|A|T| Reserved | | R |L|M|I|A|T| Reserved |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Canvas Width | | Canvas Width | ...
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Canvas Height |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
... Canvas Height |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Tiling (T): 1 bit Tiling (T): 1 bit
@ -299,25 +357,48 @@ Alpha (L): 1 bit
: Set if the file contains images with transparency information ("alpha"). : Set if the file contains images with transparency information ("alpha").
Reserved (Rsv): 3 bits Rotation and Symmetry (R): 3 bits
: SHOULD be `0`. : Specify an isometry to be applied to every bitstream chunk decoded.
The table below specifies into what coordinates a point (x,y) in the original
coordinate system has to be transformed into:
| Value | Name | New coordinates |
|:------|:--------------------------:|:--------------------------------------: |
| 0 | Identify | (x,y) |
|-------
| 1 | Horizontal symmetry | (x, CanvasHeight-1-y) |
|-------
| 2 | Vertical symmetry | (CanvasWidth-1-x, y) |
|-------
| 3 | Rotation 180 degrees | (CanvasWidth-1-x, CanvasHeight-1-y) |
|-------
| 4 | Diagonal symmetry 1 | (y, x) |
|-------
| 5 | Rotation clockwise | (CanvasHeight-1-y, x) |
|-------
| 6 | Rotation counter-clockwise | (y, CanvasWidth-1-x) |
|-------
| 7 | Diagonal symmetry 2 | (CanvasHeight-1-y, CanvasWidth-1-x) |
|-------
{: rules="groups"}
Reserved: 24 bits Reserved: 24 bits
: SHOULD be `0`. : SHOULD be `0`.
Canvas Width: 32 bits Canvas Width: 24 bits
: Width of the canvas in pixels. : _1-based_ width of the canvas in pixels.
Canvas Height: 32 bits Canvas Height: 24 bits
: Height of the canvas in pixels. : _1-based_ height of the canvas in pixels.
Future specifications MAY add more fields. If a chunk of larger size is found, The product of _Canvas Width_ and _Canvas Height_ SHOULD be at most `2^32 - 1`.
programs MUST ignore the extra bytes but SHOULD preserve them when modifying
the file. Future specifications MAY add more fields.
### Chunks ### Chunks
@ -344,63 +425,48 @@ This chunk MUST appear if the _Animation_ flag in chunk VP8X is set.
If the _Animation_ flag is not set and this chunk is present, it If the _Animation_ flag is not set and this chunk is present, it
SHOULD be ignored. SHOULD be ignored.
Per-frame parameters of the animation: Frame chunk:
0 1 2 3 0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| ChunkHeader('FRM ') | | ChunkHeader('FRM ') |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Frame X | | Frame X | ...
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Frame Y | ... Frame Y | Frame Width ...
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Frame Width | ... | Frame Height |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Frame Height |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Frame Duration | | Frame Duration |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Frame X: 32 bits (_uint32_) Frame X: 24 bits (_uint24_)
: The X coordinate of the upper left corner of the frame. : The X coordinate of the upper left corner of the frame is `Frame X * 2`
Frame Y: 32 bits (_uint32_) Frame Y: 24 bits (_uint24_)
: The Y coordinate of the upper left corner of the frame. : The Y coordinate of the upper left corner of the frame is `Frame Y * 2`
Frame Width: 32 bits (_uint32_) Frame Width: 24 bits (_uint24_)
: The width of the frame. : The _1-based_ width of the frame.
Frame Height: 32 bits (_uint32_) Frame Height: 24 bits (_uint24_)
: The height of the frame. : The _1-based_ height of the frame.
Frame Duration: 16 bits (_uint16_) Frame Duration: 24 bits (_uint24_)
: Time to wait before displaying the next tile, in 1 millisecond units. : Time to wait before displaying the next frame, in 1 millisecond units
(_1-based_).
Notes for frames containing VP8 data: For images that are animations, this chunk contains information about a single
frame, and describes the (optional) alpha chunk and the bitstream chunk that
follows it. If the _Animation flag_ is not set and this chunk is present,
it SHOULD be ignored.
* _Frame X_ and _Frame Y_ values MUST be divisible by `32`.
**Rationale:** This ensures that pixels on U and V planes are aligned to a
16-byte boundary (even after a rotation), which may help with vector
instructions on some architectures. This also makes the tiles align to
16-pixel macroblock boundaries.
* _Frame Width_ MUST be divisible by `16` or
`Frame X + Frame Width == Canvas Width` MUST be true.
* _Frame Height_ MUST be divisible by `16` or
`Frame Y + Frame Height == Canvas Height` MUST be true.
**Rationale:** The width and height constraints simplify the handling of
macroblocks that are on the edge of a tile. VP8 decoders can overwrite
pixels outside the boundary in such a macroblock, and this guarantees they
won't overwrite any data.
#### Tiling #### Tiling
@ -409,35 +475,23 @@ Notes for frames containing VP8 data:
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| ChunkHeader('TILE') | | ChunkHeader('TILE') |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Tile Canvas X | | Tile X | ...
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Tile Canvas Y |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
... Tile Y |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Tile Canvas X: 32 bits (_uint32_) Tile X: 24 bits (_uint24_)
: X coordinate of the upper left corner of the tile. : X coordinate of the upper left corner of the tile is `Tile X * 2`
Tile Canvas Y: 32 bits (_uint32_) Tile Y: 24 bits (_uint24_)
: Y coordinate of the upper left corner of the tile. : Y coordinate of the upper left corner of the tile is `Tile Y * 2`
This chunk contains information about a single tile and describes the For images that contain tiling, this chunk contains information about a single
(optional) alpha chunk and bitstream chunk that follows it. tile and describes the (optional) alpha chunk and bitstream chunk that follow
it. If the _Tile flag_ is not set and this chunk is present, it SHOULD be
Notes for tiles containing VP8 data: ignored.
* _Tile Canvas X_ and _Tile Canvas Y_ values MUST be
divisible by `32`.
* The _Tile Width_ and _Tile Height_ can be extracted from the VP8 data.
See 'Section 9' in the [VP8 RFC][vp8spec].
* The width of a tile MUST be divisible by `16` or
`Tile Canvas X + Tile Width == Canvas Width` MUST be true.
* The height of a tile MUST be divisible by `16` or
`Tile Canvas Y + Tile Height == Canvas Height` MUST be true.
#### Alpha #### Alpha
@ -455,7 +509,7 @@ Compression method (C): 2 bits
: The compression method used: : The compression method used:
* `0`: No compression. * `0`: No compression.
* `1`: Backward reference counts encoded with arithmetic encoder. * `1`: Compressed using the WebP lossless format.
Filtering method (F): 2 bits Filtering method (F): 2 bits
@ -485,20 +539,23 @@ Alpha bitstream: _Chunk Size_ - `1` bytes
: Encoded alpha bitstream. : Encoded alpha bitstream.
This optional chunk contains encoded alpha data for a single tile. This optional chunk contains encoded alpha data for a single frame/tile.
Either **ALL or NONE** of the tiles must contain this chunk. Either **ALL or NONE** of the frame/tiles must contain this chunk. However,
there is one exception to this rule: a frame/tile containing a 'VP8L' chunk
SHOULD NOT contain this chunk. **Rationale**: the transparency information of
a frame/tile is already part of the 'VP8L' chunk.
The alpha channel data is losslessly stored as raw data (when The alpha channel data is losslessly stored as raw data (when
compression method is '0') or compressed using the lossless format compression method is '0') or compressed using the lossless format
(when the compression method is '1'). (when the compression method is '1').
* raw data: consists of a byte sequence of length width * height, * Raw data: consists of a byte sequence of length width * height,
containing all the 8bit transparency values in scan order. containing all the 8bit transparency values in scan order.
* lossless format compression: the byte sequence is a compressed * Lossless format compression: the byte sequence is a compressed
image-stream (as described in the specification for lossless format) image-stream (as described in the [WebP Lossless Bitstream Format]
of implicit dimension width x height. Note that this image-stream byte [webpllspec]) of implicit dimension width x height. (That is, this
sequence does NOT contain any headers describing the image dimension. image-stream does NOT contain any headers describing the image dimension).
**Rationale**: the dimension is already known from other sources, **Rationale**: the dimension is already known from other sources,
so storing it again would be redundant and error-prone. so storing it again would be redundant and error-prone.
@ -512,6 +569,20 @@ compression method is '0') or compressed using the lossless format
steps in the specification -unlike the other channels- that can steps in the specification -unlike the other channels- that can
improve compression. improve compression.
#### Bitstream (VP8/VP8L)
This chunk contains compressed image data. As described earlier, images
with a simple file format (lossy/lossless) have a single bitstream chunk
as the first subchunk of RIFF, while images with extended file format may
contain several of them, one for each frame/tile.
A bitstream chunk may be either (i) a VP8 chunk, using "VP8 " (note the
significant fourth-character space) as its tag _or_ (ii) a VP8L chunk , using
"VP8L" as its tag.
The formats of VP8 and VP8L chunks are as described in sections _Simple file
format (lossy)_ and _Simple file format (lossless)_ respectively.
#### Color profile #### Color profile
0 1 2 3 0 1 2 3
@ -519,21 +590,17 @@ compression method is '0') or compressed using the lossless format
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| ChunkHeader('ICCP') | | ChunkHeader('ICCP') |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Compression | Color Profile | | Color Profile |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Compression: 8 bits Color Profile: _Chunk Size_ bytes
: Compression method used:
* `0`: None.
* `1`: Deflate/inflate.
Color Profile: _Chunk Size_ - `1` bytes
: ICC profile. : ICC profile.
There SHOULD be at most one 'ICCP' chunk. This chunk MUST appear before data for all the frames.
There SHOULD be at most one such chunk. If there are more such chunks, readers
MAY ignore all except the first one.
See <http://www.color.org> for specifications. See <http://www.color.org> for specifications.
If this chunk is not present, sRGB SHOULD be assumed. If this chunk is not present, sRGB SHOULD be assumed.
@ -545,38 +612,25 @@ If this chunk is not present, sRGB SHOULD be assumed.
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| ChunkHeader('META') | | ChunkHeader('META') |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Compression | XMP Metadata | | Metadata |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Compression: 8 bits Metadata: _Chunk Size_ bytes
: Compression method used: : image metadata.
* `0`: None. This chunk MUST appear after data for all the frames.
* `1`: Deflate/inflate.
XMP Metadata: _Chunk Size_ - `1` bytes
: XMP metadata.
There SHOULD be at most one such chunk. If there are more such chunks, readers There SHOULD be at most one such chunk. If there are more such chunks, readers
MAY ignore all except the first one. MAY ignore all except the first one.
XMP packets are XML text as specified in the [XMP Specification Part
1][xmpspec]. The chunk tag is different from the one specified by Adobe
for WAV and AVI (also RIFF formats), because we have the option of
compression.
Additional guidance about handling metadata can be found in the Additional guidance about handling metadata can be found in the
Metadata Working Group's [Guidelines for Handling Metadata][metadata]. Metadata Working Group's [Guidelines for Handling Metadata][metadata].
Note that the sections of the document about reconciliation of EXIF,
XMP and IPTC-IIM don't apply to WebP. As WebP supports only XMP, no
reconciliation is necessary.
#### Other Chunks #### Unknown Chunks
A file MAY contain other chunks. Readers SHOULD be ignore these chunks. Writers A file MAY contain other unknown chunks. Readers SHOULD be ignore these chunks.
SHOULD preserve them in their original order. Writers SHOULD preserve them in their original order.
### Assembling the Canvas from Tiles and Animation ### Assembling the Canvas from Tiles and Animation
@ -584,63 +638,62 @@ Here we provide an overview of how 'TILE' chunks and 'FRM '/'LOOP' chunks are
used to assemble the canvas. The notation _VP8X.field_ means the field in used to assemble the canvas. The notation _VP8X.field_ means the field in
the 'VP8X' chunk with the same description. the 'VP8X' chunk with the same description.
Decoding a non-animated canvas MUST be equivalent to the following Decoding a _non-animated_ canvas MUST be equivalent to the following
pseudocode: pseudocode:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
assert not VP8X.flags.haveAnimation assert not VP8X.flags.haveAnimation
canvas ← new black image of size canvas ← new black image of size
VP8X.canvasWidth x VP8X.canvasHeight. VP8X.canvasWidth x VP8X.canvasHeight.
tile_params.tileCanvasX = tile_params.tileCanvasY = 0 tile_params.tileX = tile_params.tileY = 0
for chunk in data_for_all_frames: for chunk in data_for_all_frames:
if chunk.tag == "TILE": if chunk.tag == "TILE":
assert No other TILE chunk after the last "VP8 " chunk assert No other TILE chunk after the last Bitstream chunk
assert No ALPH chunk after the last "VP8 " chunk assert No ALPH chunk after the last Bitstream chunk
tile_params = chunk tile_params = chunk
assert VP8X.canvasWidth >=
tile_params.tileX + tile_params.tileWidth
assert VP8X.canvasHeight >=
tile_params.tileY + tile_params.tileHeight
if chunk.tag == "ALPH": if chunk.tag == "ALPH":
assert No other ALPH chunk after the last "VP8 " chunk assert No other ALPH chunk after the last Bitstream chunk
tile_params.alpha = alpha_data tile_params.alpha = alpha_data
if chunk.tag == "VP8 ": if chunk.tag == "VP8 " OR chunk.tag == "VP8L":
render image in chunk in canvas with top-left corner in render image in chunk on canvas with top-left corner in
(tile_params.tileCanvasX, tile_params.tileCanvasY). (tile_params.tileX, tile_params.tileY).
tile_params.tileCanvasX = tile_params.tileCanvasY = 0
Ignore unknown chunks Ignore unknown chunks
canvas contains the decoded canvas. canvas contains the decoded canvas.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Decoding an animated canvas MUST be equivalent to the following Decoding an _animated_ canvas MUST be equivalent to the following
pseudocode: pseudocode:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
assert VP8X.flags.haveAnimation assert VP8X.flags.haveAnimation
canvas ← new black image of size canvas ← new black image of size
VP8X.canvasWidth x VP8X.canvasHeight. VP8X.canvasWidth x VP8X.canvasHeight.
if LOOP.loopCount==0: loop_count ← LOOP.loopCount
LOOP.loopCount=∞ if loop_count == 0:
current_FRM ← nil loop_count = ∞
frame_params ← nil
for LOOP.loop = 0, ..., LOOP.loopCount-1 for LOOP.loop = 0, ..., LOOP.loopCount-1
assert First chunk in data_for_all_frames is a FRM assert First chunk in data_for_all_frames is FRM
for chunk in data_for_all_frames: for chunk in data_for_all_frames:
if chunk.tag == "FRM ": if chunk.tag == "FRM ":
assert No other FRM chunk after the last "VP8 " chunk assert No other FRM chunk after the last Bitstream chunk
assert No ALPH chunk after the last "VP8 " chunk assert No ALPH chunk after the last Bitstream chunk
if current_FRM != nil: frame_params = chunk
Show the contents of canvas for assert VP8X.canvasWidth >=
current_FRM.frameDuration * 1ms. frame_params.frameX + frame_params.frameWidth
current_FRM = chunk assert VP8X.canvasHeight >=
frame_params.frameY + frame_params.frameHeight
if chunk.tag == "ALPH": if chunk.tag == "ALPH":
assert No other ALPH chunk after the last "VP8 " chunk assert No other ALPH chunk after the last Bitstream chunk
tile_params.alpha = alpha_data frame_params.alpha = alpha_data
if chunk.tag == "VP8 ": if chunk.tag == "VP8 " OR chunk.tag == "VP8L":
assert tile_params.tileCanvasX >= current_FRM.frameX render image in chunk on canvas with top-left corner in
assert tile_params.tileCanvasY >= current_FRM.frameY (frame_params.frameX, frame_params.frameY). Show the contents
assert tile_params.tileCanvasX + chunk.tileWidth >= of the image for frame_params.frameDuration * 1ms.
current_FRM.frameX + current_FRM.frameWidth
assert tile_params.tileCanvasY + chunk.tileHeight >=
current_FRM.frameX + current_FRM.frameHeight
render image in chunk in canvas with top-left corner in
(tile_params.tileCanvasX, tile_params.tileCanvasY).
tile_params.tileCanvasX = tile_params.tileCanvasY = 0
Ignore unknown chunks Ignore unknown chunks
canvas contains the decoded canvas. canvas contains the decoded canvas.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
@ -651,7 +704,7 @@ MAY ignore the badly-ordered chunks instead of failing to decode the file.
Example file layouts Example file layouts
-------------------- --------------------
A non-animated, tiled image without transparency may look as follows: A tiled image without transparency may look as follows:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
RIFF/WEBP RIFF/WEBP
@ -665,7 +718,7 @@ RIFF/WEBP
+- VP8 (bitstream - third tile) +- VP8 (bitstream - third tile)
+- TILE (fourth tile parameters) +- TILE (fourth tile parameters)
+- VP8 (bitstream - fourth tile) +- VP8 (bitstream - fourth tile)
+- META (XMP metadata) +- META (metadata)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
An animated image with transparency may look as follows: An animated image with transparency may look as follows:
@ -675,15 +728,43 @@ RIFF/WEBP
+- VP8X (descriptions of features used) +- VP8X (descriptions of features used)
+- LOOP (animation control parameters) +- LOOP (animation control parameters)
+- FRM (first animation frame parameters) +- FRM (first animation frame parameters)
+- ALPH (alpha bitstream - first image frame) +- ALPH (alpha bitstream - first frame)
+- VP8 (bitstream - first image frame) +- VP8 (bitstream - first frame)
+- FRM (second animation frame parameters) +- FRM (second animation frame parameters)
+- ALPH (alpha bitstream - second image frame) +- ALPH (alpha bitstream - second frame)
+- VP8 (bitstream - second image frame) +- VP8 (bitstream - second frame)
+- META (XMP metadata) +- META (metadata)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
A losslessly encoded non-animated non-tiled image may
look as follows:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
RIFF/WEBP
+- VP8X (descriptions of features used)
+- ICCP (color profile)
+- XYZW (unknown chunk)
+- VP8L (lossless bitstream)
+- META (metadata)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
An animated image may have a mix of lossy and lossless
bitstreams as follows:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
RIFF/WEBP
+- VP8X (descriptions of features used)
+- FRM (first animation frame parameters)
+- VP8 (lossy bitstream - first frame)
+- FRM (second animation frame parameters)
+- VP8L (lossless bitstream - second frame)
+- ABCD (unknown chunk)
+- FRM (third animation frame parameters)
+- VP8 (lossy bitstream - third frame)
+- EFGH (unknown chunk)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
[vp8spec]: http://tools.ietf.org/html/rfc6386 [vp8spec]: http://tools.ietf.org/html/rfc6386
[xmpspec]: http://www.adobe.com/content/dam/Adobe/en/devnet/xmp/pdfs/XMPSpecificationPart1.pdf [webpllspec]: https://gerrit.chromium.org/gerrit/gitweb?p=webm/libwebp.git;a=blob;f=doc/webp-lossless-bitstream-spec.txt;hb=master
[metadata]: http://www.metadataworkinggroup.org/pdf/mwg_guidance.pdf [metadata]: http://www.metadataworkinggroup.org/pdf/mwg_guidance.pdf
[rfc 2119]: http://tools.ietf.org/html/rfc2119 [rfc 2119]: http://tools.ietf.org/html/rfc2119