Babl is a dynamic, any to any, pixel format conversion library. It provides conversions between the myriad of buffer types images can be stored in. Babl doesn't only help with existing pixel formats, but also facilitates creation of new and uncommon ones.
News
- xxxx-xx-xx babl-0.0.2 Green fish release
-
DRAFT Initial release.
- 2005-08-xx babl-0.0.1
- Development release of babl, minimal working configuration for RGB, CMYK and YCbCr for 8bit 16bit float and double. Emphasis is on precision and clarity of code not speed.
See the ChangeLog in Subversion.
Background
When creating various libraries and small tools that deal with computer graphics I have often created functions for converting raster images between different representations.
Libraries to help with such conversions already do exist, but no extendable candidates have surfaced in my searches.
The vocabulary developed in babl will most likely be part of GEGL's vocabulary to describe image buffers.
How babl works
When using BablFishes to do your conversions, you request a fish to convert between two formats, and an optimal fish to babls capability is provided that you can use to do your conversions. Babl also provides the capability to describe new formats based on a vocabulary of user registered color models and data types.
Babl provides a base vocabulary in BablBase and some extensions that are thought to be generally useful.
When performing further extensions to the vocabulary of babl, the internal consistency is governed by reference conversions that operate on double (64 bit floating point values). The only color model created during BablCore bootstrap is RGBA (linear light RGB, 0.0 - 1.0, with a linear 0.0 - 1.0 opacity channel) backed by the double datatype.
If babls conversion isn't fast enough, you can provide your own conversion shortcut between two formats. The registered shortcut might also be used by babl as an intermediate conversion when constructing BablFishes for other conversions.
Babl extensions are shared objects. If you have already developed some fast conversion functions, wrapping them as babl extensions should not take much time and will speed up babl for other users as well.
Features
Core
- Small API for simple use cases
- Planar and linear buffers
- Thread safety for processing.
- Extendable and introspectable formats, color models, components and datatypes
- Reference 64bit floating point conversions for data types and color models.
Vocabulary
+-
Extensions
At compile, load, and runtime; babl is extendable with:
- data types.
- color models.
- pixel formats.
- optimized conversion functions:
- SIMD instructions
- MMX, SSE, Altivec ...
- External libraries
- liboil, hermes, libavcodec, lcms, ...
Shortcut coverage
The diagram shown below visualizes the coverage of current shortcut conversions. A DHTML version is also available.
·· ·▁····· ▁ ▂▂ ·· · ── 0 RGBA double
· ▁ ▁▂▁▁▁▁▁ · ▂▂▁▁ ▁▁ ▁ ── 1 R'G'B'A double
·▁ ▁▂▁▁▁▁▁ ▂ ▁▁ ▁ ── 2 R'aG'aB'aA double
··▁ ▁ ▁ ·· · ── 3 R'G'B' u8
▂· ·▁▂ · ▁▁·▁ ── 4 R'G'B'A u8
·▁▁ ·▁▁▁▁▁ ▂ · ▁▁ ▁ ── 5 RGBA float
▁▂▂ · ▂▂▂▂▂ · ▂▂ ▂ ── 6 RGB float
·▁▁ ▁▂ ▁▁▁▁ ▂ ▂ ▁▁ ▁ ── 7 Y'aA double
·▁▁ ▁▂▁ ▁▁▁ ▂ ▂ ▁▁ ▁ ── 8 Y double
·▁▁ ▁▂▁▁ ▁▁ ▂ ▁▁ ▁ ── 9 YA double
·▁▁ ▁▂▁▁▁ ▁ ▂ ▂ ▁▁ ▁ ──10 Y'CbCr double
·▁▁ ▁▂▁▁▁▁ ▂ ▁▁ ▁ ──11 Y'CbCrA double
──12 Y'CbCr u8
▁·▂ ▂ ▂▂▂▂▂ ·▁·· ▂▂ ▂ ──13 R'G'B'A float
· ▂▂▁▁ ──14 R'G'B'A u16
· ▁·▁ ──15 R'aG'aB'aA float
▁ · ──16 R'aG'aB'aA u16
▂▂ ▁ ·· ▂ ──17 R'aG'aB'aA u8
▂▁ · · ──18 R'G'B' float
▁ ▁ · ──19 R'G'B' u16
▂▁ · ▁▂▁▁ ·· ▂ ▁ ──20 Y'A float
▁ ·▂▁▂ · ▁ ▁ · ──21 Y'aA float
▂ ▁ ▁ · ·· ──22 Y' float
▁ ▂▂ · ▁ ▂ ──23 Y'A u16
▁ ▂ · ▁ ──24 Y'aA u16
▂ ▂ ▁ ▁▁· ──25 Y' u16
▁ ▂ ▂▂ ·▁▁ ▂ ──26 Y'A u8
▂ ▁ ▂ ▁·▂ · ──27 Y'aA u8
▂ ▂ ▁ ▁▁· ──28 Y' u8
──29 Y'CbCr float
──30 Y'CbCr (Y' as float) (Cb as float) (Cr as float) (A as float)
▁▂▂ ·▁▂▂▂▂▂ ▁ ▂▂ ▂ ──31 RGBA u16
▁▂▂ ·▁▂▂▂▂▂ · ▂▂ ▂ ──32 RGBA u8
──33 RaGaBaA float
· ──34 RaGaBaA u16
· ──35 RaGaBaA u8
▂ ▁· ▁ ──36 RGB u16
▂ ▁▂ ·▁ · ──37 RGB u8
▁▂▂ ·▁▂▂▂▂▂ ▁ ·· ▂▂ ▂ ──38 YA float
▂ ▁▂ · · ▁ ──39 YaA float
▂ ▁· ▁ ·· ──40 Y float
▁▂ · ▁ ──41 YA u16
▁ · ──42 YaA u16
▂▁ ▂ ▁▁· ──43 Y u16
▁▂ · ▁ ──44 YA u8
▂ ▁ ▁·▂ ──45 YaA u8
▂▁ ▂ ▁▁· ──46 Y u8
──47 Y'CbCrA float
──48 CIE Lab float
──49 CIE Lab alpha float
──50 CIE Lab u8
──51 CIE Lab u16
·▁▁ ▁▂▁▁▁▁▁ ▂ ▂ ▁ ▁ ──52 CIE Lab double
·▁▁ ▁▂▁▁▁▁▁ ▂ ▁ ▁ ──53 CIE Lab alpha double
──54 B'aG'aR'aA u8
──55 B'aG'aR'aPAD u8
·▁▁ ▁▂▁▁▁▁▁ ▂ ▂ ▁▁ ──56 CMYK double
──57 CMYK float
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
||||||||||111111111122222222223333333333444444444455555555
0123456789012345678901234567890123456789012345678901234567
total length: 829
total cost : 799939
Dependencies
Babl is intended to be a low level infrastructure library and thus strives for few dependencies and thus portability.
- libdl (for loading extensions, babl is also reported to work on win32)
Download
The latest versioned development version of babl can be found in ftp://ftp.gtk.org/pub/babl/.
Subversion
The latest development version is available in the babl module in GNOME Subversion.
Usage
The simplest scenario for using babl is converting between linear buffers represented by an existing BablPixelFormat.
babl_process (babl_fish ("srgb", "CIE Lab float"), srgb_buffer, lab_buffer, pixel_count);
If the existing pixel formats are not sufficient for your conversion needs, new ones can be created on the fly. The constructor will provide the prior created one if duplicates are registered.
babl_format_new (babl_model ("R'G'B'"), babl_type ("u8"), babl_component ("B'"), babl_component ("G'"), babl_component ("R'"), NULL);
Extending
For samples of how the current internal API specification of data types, color models, and conversions look in the extensions/ directory. The tables in this HTML file is directly generated based on the data registered by BablCore (double and RGBA), BablBase (core datatypes, and RGB models), extensions (CIE Lab, naive CMYK, various shortcut conversions). The API's used are very similar in style to the API's described under the Usage section.
Color Management
Babl is not a color management system, since it doesn't deal with ICC profiles. Babl is designed primarily for internal use when the color space is already known (sRGB, CIE Lab, Luminance or similar).
It is also possible to register new color models that are managed by a color management system like lcms, take a look at the lcms-lab extensions for an example.
TODO
- Validate that a created BablFishPath is actually faster than the BablFishReference it is a replacement for.
- Compile-time instead of runtime profiling / storage of profiling and loss data.. (The speed of conversions is bound to 'wobble', according to architecture, compiler flags, concurrently running processes and phase of moon.)
- Horizontal chroma subsampling (4:2:2 and 4:1:1)
- Thread safety (locking) for mutations of the type system.
- Support for datatypes that are not a multiple of 8bit.
- dithering
Copyright
Babl is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version.
Authors
- Øyvind Kolås pippin@gimp.org
- Lead developer.
- Sven Neumann sven@gimp.org
- Build sanity.
- Michael Natterer mitch@gimp.org
- Build sanity.
- Kevin Cozens
- Build sanity.
- Tim Mooney
- Portability fixes.
- Michael Schumacher schumaml@cvs.gnome.org
- win32 support for dynamic extensions.
- Portability fixes.
