</style>
<script type='text/javascript'>
- function hide(id)
- {
- (document.getElementById(id)).style.display = "none";
- }
- function show(id)
- {
- (document.getElementById(id)).style.display = "block";
- }
+ function hide(id)
+ {
+ (document.getElementById(id)).style.display = "none";
+ }
+ function show(id)
+ {
+ (document.getElementById(id)).style.display = "block";
+ }
function get_visible (id)
{
var element = document.getElementById(id);
return true;
return false;
}
-
function set_visible (id, visible)
{
var element = document.getElementById(id);
else
set_visible(id,true);
}
-
</script>
</head>
<body>
<h3>Contents</h3>
</div>
<ul>
- <li><a href='#About'>About</a></li>
+ <li><a href='#Babl'>Babl</a></li>
<li><a href='#NEWS'>News</a></li>
+ <li><a href='#Background'>Background</a></li>
+ <li><a href='#How'>How?</a></li>
<li><a href='#Features'>Features</a></li>
+ <li><a href='#Vocabulary'>Vocabulary</a></li>
<li><a href='#Data-types'> Data types</a></li>
<li><a href='#Color-models'> Color models</a></li>
<li><a href='#Pixel-formats'> Pixel formats</a></li>
+ <li><a href='#Extensions'>Extensions</a></li>
<li><a href='#Dependencies'>Dependencies</a></li>
<li><a href='#Source'>Source</a></li>
<li><a href='#Usage'>Usage</a></li>
<li><a href='#Extending'>Extending</a></li>
+ <li><a href='#ColorManagement'>Color Management</a></li>
<li><a href='#TODO'>Todo</a></li>
<li><a href='#Copyright'>Copyright</a></li>
<li><a href='#Authors'> Authors</a></li>
</ul>
</div>
- <a name='About'></a>
+ <a name='Babl'></a>
<div class='paper'>
<div class='content'>
- <h2>About</h2>
+ <!--<h1>Babl</h1>-->
<p>Babl is a <em>dynamic, any to any, pixel format conversion
library</em> It provides conversions between the myriad of buffer
types images can be stored in. Babl doesn't only help with existing
<a name='NEWS'></a>
<h2>News</h2>
+<!--
<!--NEWS-->
- <p>Also see the <a href='http://cvs.gnome.org/viewcvs/*checkout*/babl/ChangeLog'>ChangeLog in CVS</a>.</p>
+-->
+ <p>See the <a href='http://cvs.gnome.org/viewcvs/*checkout*/babl/ChangeLog'>ChangeLog in CVS</a>.</p>
+
+ <a name='Background'></a>
+ <h2>Background</h2>
+ <p> When creating various libraries and small tool that deal with
+ computer graphics I often have created conversion functions
+ transforming raster images between different representations. </p>
+
+ <p>Libraries to help with such conversions already do exist, but no
+ extendable candidate have surfaced in searches when looking.
+ </p>
+
+ <p>The vocabulary developed in babl will most likely be part of GEGL's
+ vocabulary to describe image buffers.
+ </p>
+
+ <a name='How'></a>
+ <h2>How babl works</h2>
+ <p>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 vocbulary of user
+ registered color models and data types.
+ </p>
+
+ <p> The need to provide your own reference conversions should be
+ unneccesary, since babl provides a good base vocabulary in BablBase,
+ as well as extensions.</p>
+
+ <p>When performing further extensions to the vocabulary of babl, the
+ internal consistency is governed by reference conversions that operate
+ on <em>double</em> (64 bit floating point values.) The only color model
+ created during BablCore bootstrap is <em>RGBA</em> (linear light RGB,
+ 0.0 - 1.0, with a linear 0.0 - 1.0 opacity channel.) backed by
+ <em>double</em>.
+ </p>
+
+ <p>If what babls built in conversion isn't fast enough, you can provide
+ your own conversion between two formats, your registered conversion
+ will be taken into account when constructing other BablFishes as well.
+ </p>
+
+ <p>Babl extensions are shared objects, if you are already using a set
+ of conversion functions and you'd like to keep using them, as well as
+ the conversions babl provide, wrap your existing functions as an
+ extension. The code needed to wrap existing conversions is small.
+ </p>
<a name='Features'></a>
<h2>Features</h2>
<li class='unstable'>Extendable and <a href='http://pippin.gimp.org/babl/introspect.txt'>introspectable</a> formats,
color models, components and datatypes</li>
<li class='unstable'>Reference 64bit floating point conversions for
- data types and color models.
-<!--BablBase-->
- </li>
+ data types and color models.</li>
</ul>
+ <a name='Vocabulary'></a>
+ <h3 style='margin-bottom:0em;'>Vocabulary</h3>
+<!--BablBase-->
+
+ <a name='Extensions'></a>
<h3>Extensions</h3>
<p>
At compile, load and runtime; babl is extendable with:
</li>
</ul>
+ <h3>Shortcut coverage</h3>
+ <p>The diagram shown below is the coverage of shortcut conversions that already exist,
+ prior to the BablFishPath actually being operational, due to lack of testing of loss
+ against the reference conversions this is the most optimistic estimate. (<a href='BablFishPath.html'>DHTML version</a>)
+ </p>
+<pre style='font-size:50%;height:30em;overflow:scroll'>
+<!--BablFishPath-->
+</pre>
<a name='Dependencies'></a>
<h2>Dependencies</h2>
buffers represented by an existing BablPixelFormat.
</p>
<pre
-><span class='function'>babl_process</span> <span class='paren'>(</span><span class='function'>babl_fish</span> <span class='paren'>(</span><span class='string'>"srgb"</span>, <span class='string'>"lab-float"</span><span class='paren'>)</span>,
+><span class='function'>babl_process</span> <span class='paren'>(</span><span class='function'>babl_fish</span> <span class='paren'>(</span><span class='string'>"srgb"</span>, <span class='string'>"CIE Lab float"</span><span class='paren'>)</span>,
srgb_buffer, lab_buffer,
pixel_count<span class='paren'>);</span></pre>
needs, new ones can be created on the fly. The constructor
will provide the prior created one if duplicates are registered. </p>
<pre
-><span class='function'>babl_format_new</span> <span class='paren'>(</span><span class='function'>babl_model</span> <span class='paren'>(</span><span class='string'>"rgb"</span><span class='paren'>)</span>,
+><span class='function'>babl_format_new</span> <span class='paren'>(</span><span class='function'>babl_model</span> <span class='paren'>(</span><span class='string'>"R'G'B'"</span><span class='paren'>)</span>,
<span class='function'>babl_type</span> <span class='paren'>(</span><span class='string'>"u8"</span><span class='paren'>)</span>,
- <span class='function'>babl_component</span> <span class='paren'>(</span><span class='string'>"B"</span><span class='paren'>)</span>,
- <span class='function'>babl_component</span> <span class='paren'>(</span><span class='string'>"G"</span><span class='paren'>)</span>,
- <span class='function'>babl_component</span> <span class='paren'>(</span><span class='string'>"R"</span><span class='paren'>)</span>,
+ <span class='function'>babl_component</span> <span class='paren'>(</span><span class='string'>"B'"</span><span class='paren'>)</span>,
+ <span class='function'>babl_component</span> <span class='paren'>(</span><span class='string'>"G'"</span><span class='paren'>)</span>,
+ <span class='function'>babl_component</span> <span class='paren'>(</span><span class='string'>"R'"</span><span class='paren'>)</span>,
<span class='NULL'>NULL</span><span class='paren'>);</span></pre>
<p>Instead of a linear buffer you can an image descriptor which desribes the start of the memory segment, the pitch in bytes between samples, and the rowstride (not used yet,
use 0 to indicate unlimited for compatibility with future API.)
</p>
<pre
-><span class='function'>babl_process</span> <span class='paren'>(</span><span class='function'>babl_fish</span> <span class='paren'>(</span><span class='string'>"srgb"</span>, <span class='string'>"y'cbcr420p"</span><span class='paren'>)</span>,
+><span class='function'>babl_process</span> <span class='paren'>(</span><span class='function'>babl_fish</span> <span class='paren'>(</span><span class='string'>"srgb"</span>, <span class='string'>"Y'CbCr u8"</span><span class='paren'>)</span>,
srgb_buffer,
<span class='function'>babl_image</span> <span class='paren'>(</span><span class='function'><span class='string'>"Y'"</span>, luma_buffer, 1, 0,
- <span class='string'>"Cb"</span>, cb_buffer, 1, 0,
- <span class='string'>"Cr"</span>, cr_buffer, 1, 0,
+ <span class='string'>"Cb"</span>, cb_buffer, 1, 0,
+ <span class='string'>"Cr"</span>, cr_buffer, 1, 0,
<span class='NULL'>NULL</span><span class='paren'>);</span>
</pre>
similar in style to the API's described under the <a href='#Usage'>Usage section</a>.
</p>
+ <a name='ColorManagement'></a>
+ <h2>Color Management</h2>
+ <p> Babl is not a color management system, since it doesn't deal with
+ ICC profiles. One way to deal with this is to register a custom color
+ model, that is backed by for instance <em>lcms</em>. (see the
+ <em>lcms-lab</em> extension) Then when using the newly registered color
+ model to create pixel formats, lcms will be used behind the scenes.
+ </p>
<a name='TODO'></a>
<h2>TODO</h2>
<ul>
- <li>Compile and/or runtime profiling. <em>(The speed of conversions
+ <li>Compile-time instead of runtime profiling / storae of profiling
+ and loss data.. <em>(The speed of conversions
is bound to 'wobble', according to architecture, compiler flags,
concurrently running processes and phase of moon.)</em></li>
<li class='unstable'>Horizontal and vertical subsampling (for
implementing 4:2:2 4:2:0 4:1:1 etc. chroma subsampling)</li>
<li>Thread safety for mutations of the type system.</li>
<li>Support for datatypes that are not a multiple of 8bit.</li>
- <!--<li>Relative error is the difference between the two vaues
- divided by the actual value, for instance, if the actual result of a
- calculation should be 3.13159, but the result obtained was 3.14x10^0,
- the relative error is 0.00159/3.14159 => 0.0005 (source: 'A new book
- on c (cook1_0a.pdf)')</li>-->
-<!--
-cvs:
-
-cp babl/babl-pixel-format.c babl/babl-format.c
-cp babl/babl-pixel-format.h babl/babl-format.h
-cp babl/babl-classes.c babl/babl-internal.c
-
-
--->
+ <li>dithering</li>
+ <li>image resampling?</li>
</ul>
-
<a name='Copyright'></a>
<h2>Copyright</h2>
projects / babl.git / commitdiff