Merge branch 'master' of github.com:jcupitt/libvips

Conflicts:
	ChangeLog
	TODO
This commit is contained in:
John Cupitt 2014-08-17 09:58:24 +01:00
commit c646356b4f
15 changed files with 691 additions and 79 deletions

View File

@ -1,6 +1,10 @@
25/7/14 started 7.41.0
- start working on --disable-deprecated
- fix pngload with libpng >1.6.1
- add vips_resize()
12/8/14 started 7.40.5
- more doc fixes
25/7/14 started 7.40.5
- fix a race in im_maxpos_avg()
@ -10,9 +14,12 @@
- argh fix affine, again, there were sometimes black bars with nohalo and the
vips8 interface
- pngsave in interlaced mode makes a copy of the image, so it's always seq
- vipsthumbnail shrinks to 1/2 window_size: faster and better quality
- vipsthumbnail defaults to bicubic + nosharpen
- vipsthumbnail shrinks to 1/2 window_size
- vipsthumbnail has an anti-alias filter between shrink and affine
- vipsthumbnail defaults to bicubic
- better rounding behaviour for fixed-point bicubic reduces noise
- fix pngload with libpng >=1.6.11
- fix colour for openslide read associated
4/7/14 started 7.40.4
- fix vips_rawsave_fd(), thanks aferrero2707

12
TODO
View File

@ -1,3 +1,8 @@
- can we pick the vipsthumbnail int shrink factor more intelligently?
- did we include the exif patch in the latest windows build? check on laptop
we don't seem to have this fix in the repository!
- vips_object_unref_outputs() needs docs ... bindings will need it
@ -5,10 +10,15 @@
finally finish --disable-deprecated option
- copy-paste blog post about writing vips8 operations into 'extending'
- vips_init() has a deprecation warning, bizarre
- vips_affine() for simple resize is slow at the command-line, since it has to
be non-seq ... make a seq version that can only up/down size
- check and fix up docs

View File

@ -38,7 +38,7 @@ VIPS_VERSION_STRING=$VIPS_VERSION-`date`
# binary interface changes not backwards compatible?: reset age to 0
LIBRARY_CURRENT=38
LIBRARY_REVISION=3
LIBRARY_REVISION=4
LIBRARY_AGE=0
# patched into include/vips/version.h

View File

@ -19,6 +19,18 @@
<para>
Write this section once the vips8 Python binding is done.
</para>
<para>
If you are writing a language binding, you won't need these. Instead, make
a new operation with vips_operation_new() (all it does is look up the
operation by name with vips_type_find(), then call g_object_new() for you),
then use vips_argument_map() and friends to loop over the operation's
arguments setting them. Once you have set all arguments, use
vips_cache_operation_build() to look up the operation in the cache and
either build or dup it. If something goes wrong, you'll need to use
vips_object_unref_outputs() and g_object_unref() to free the
partially-built object.
</para>
</refsect1>
</refentry>

View File

@ -14,23 +14,354 @@
<refpurpose>How to add operations to VIPS</refpurpose>
</refnamediv>
<refsect1 id="extending-subclassing">
<title>Adding operations to VIPS</title>
<refsect1 id="extending-pointtopoint">
<title>A simple point-to-point operation</title>
<para>
All about subclassing. Copy-paste from the blog post about extending
vips8.
All operations are subclasses of #VipsOperation, which in turn
subclasses #VipsObject and then %GObject. You need to define a new
instance struct and a new class struct.
<programlisting language="C">
typedef struct _Negative {
VipsOperation parent_instance;
VipsImage *in;
VipsImage *out;
int image_max;
} Negative;
typedef struct _NegativeClass {
VipsOperationClass parent_class;
/* No new class members needed for this op.
*/
} NegativeClass;
</programlisting>
</para>
<para>
If you are writing a language binding, you won't need these. Instead, make
a new operation with vips_operation_new() (all it does is look up the
operation by name with vips_type_find(), then call g_object_new() for you),
then use vips_argument_map() and friends to loop over the operation's
arguments setting them. Once you have set all arguments, use
vips_cache_operation_build() to look up the operation in the cache and
either build or dup it. If something goes wrong, you'll need to use
vips_object_unref_outputs() and g_object_unref() to free the
partially-built object.
This operation will find the photographic negative of an unsigned
8-bit image, optionally letting you specify the value which the pixels
"pivot" about. It doesn't need any class members (ie. values common
to all operations of this type), so the second struct is empty. See
vips_invert() for a more complete version of this operation that's
actually in the library.
</para>
<para>
GObject has a handy macro to write some of the boilerplate for you.
<programlisting language="C">
G_DEFINE_TYPE( Negative, negative, VIPS_TYPE_OPERATION );
</programlisting>
This defines a function called negative_get_type(),
which registers this new class and returns its #GType (a
pointer-sized integer). negative_get_type() in turn needs two
functions, negative_init(), to initialise a new instance, and
negative_class_init(), to initialise a new class.
</para>
<para>
negative_init() is very simple, it just sets the default value for
our optional class parameter.
<programlisting language="C">
static void
negative_init( Negative *negative )
{
negative->image_max = 255;
}
</programlisting>
</para>
<para>
negative_class_init() is more complicated: it has to set various
fields in various superclasses.
<programlisting language="C">
static void
negative_class_init( NegativeClass *class )
{
GObjectClass *gobject_class = G_OBJECT_CLASS( class );
VipsObjectClass *object_class = VIPS_OBJECT_CLASS( class );
gobject_class->set_property = vips_object_set_property;
gobject_class->get_property = vips_object_get_property;
object_class->nickname = "negative";
object_class->description = "photographic negative";
object_class->build = negative_build;
VIPS_ARG_IMAGE( class, "in", 1,
"Input",
"Input image",
VIPS_ARGUMENT_REQUIRED_INPUT,
G_STRUCT_OFFSET( Negative, in ) );
VIPS_ARG_IMAGE( class, "out", 2,
"Output",
"Output image",
VIPS_ARGUMENT_REQUIRED_OUTPUT,
G_STRUCT_OFFSET( Negative, out ) );
VIPS_ARG_INT( class, "image_max", 4,
"Image maximum",
"Maximum value in image: pivot about this",
VIPS_ARGUMENT_OPTIONAL_INPUT,
G_STRUCT_OFFSET( Negative, image_max ),
0, 255, 255 );
}
</programlisting>
</para>
<para>
In %GObject, it needs to set the getters and setters for this class. vips
has a generic get/set system, so any subclass of #VipsObject needs to
use the vips ones.
</para>
<para>
In #VipsObject, it needs to set the operation @nickname and @description,
and set a build function (see below). @nickname is used to refer to
this operation in the API, @description is used to explain this
operation to users and will be translated into their language.
</para>
<para>
Finally, it needs to set the arguments this class constructor
takes. There are a set of handy macros for doing this. The first few
parameters are always the same and mean: class pointer for argument,
argument name, argument priority (bindings expect required arguments in
order of priority), long argument name (this one is internationalised
and displayed to users), description (again, users can see this),
some flags describing the argument, and finally the position of the
member in the struct.
</para>
<para>
Integer arguments take three more values: the minimum, maximum and
default value for the argument.
</para>
<para>
The build function is the thing VipsObject calls after supplying
arguments. It checks that all required arguments have been set and are
valid and constructs the object. After build, the object is expected
to be ready for use.
<programlisting language="C">
static int
negative_build( VipsObject *object )
{
VipsObjectClass *class = VIPS_OBJECT_GET_CLASS( object );
Negative *negative = (Negative *) object;
if( VIPS_OBJECT_CLASS( negative_parent_class )->build( object ) )
return( -1 );
if( vips_check_uncoded( class->nickname, negative->in ) ||
vips_check_format( class->nickname, negative->in, VIPS_FORMAT_UCHAR ) )
return( -1 );
g_object_set( object, "out", vips_image_new(), NULL );
if( vips_image_pipelinev( negative->out,
VIPS_DEMAND_STYLE_THINSTRIP, negative->in, NULL ) )
return( -1 );
if( vips_image_generate( negative->out,
vips_start_one,
negative_generate,
vips_stop_one,
negative->in, negative ) )
return( -1 );
return( 0 );
}
</programlisting>
</para>
<para>
negative_build() first chains up to the superclass: this will check
that all input arguments have been supplied and are sane.
</para>
<para>
Next, it adds its own checks. This is a demo operation, so we just
work for uncoded, unsigned 8-bit images.
</para>
<para>
Next, it creates the output image. This needs to be set with
g_object_set() so that vips can see that it has been assigned. vips
will also handle the reference counting for you.
</para>
<para>
vips_image_pipelinev() links our new image onto the input image and
notes that this operation prefers to work in lines.
</para>
<para>
Finally, vips_image_generate() attaches a set of callbacks to the
output image to generate chunks of it on request. vips_start_one()
and vips_stop_one() are convenience functions that make the input
region for you.
</para>
<para>
And then the actual image processing.
<programlisting language="C">
static int
negative_generate( VipsRegion *or,
void *vseq, void *a, void *b, gboolean *stop )
{
/* The area of the output region we have been asked to make.
*/
VipsRect *r = &amp;or-&gt;valid;
/* The sequence value ... the thing returned by vips_start_one().
*/
VipsRegion *ir = (VipsRegion *) vseq;
Negative *negative = (Negative *) b;
int line_size = r-&gt;width * negative-&gt;in-&gt;Bands;
int x, y;
/* Request matching part of input region.
*/
if( vips_region_prepare( ir, r ) )
return( -1 );
for( y = 0; y &lt; r-&gt;height; y++ ) {
unsigned char *p = (unsigned char *)
VIPS_REGION_ADDR( ir, r-&gt;left, r-&gt;top + y );
unsigned char *q = (unsigned char *)
VIPS_REGION_ADDR( or, r-&gt;left, r-&gt;top + y );
for( x = 0; x &lt; line_size; x++ )
q[x] = negative-&gt;image_max - p[x];
}
return( 0 );
}
</programlisting>
</para>
<para>
This has to calculate a section of the output image. The output
#VipsRegion, @or, contains a #VipsRect called @valid which is the
area needing calculation. negative_generate() asks for the
corresponding pixels from the input region, then loops over the
area. VIPS_REGION_ADDR() is a simple macro that does pointer arithmetic
for you: you need to stay within the valid area.
</para>
<para>
To add the operation to vips, just call negative_get_type(). You
can then use @negative from any of the vips interfaces. For example,
in Python you'd use it like this:
<programlisting language="python">
out = in.negative(image_max = 128)
</programlisting>
</para>
<para>
From the command-line it'd look like this:
<programlisting language="bash">
$ vips negative in.png out.tif --image-max 128
</programlisting>
</para>
<para>
And from C like this:
<programlisting language="C">
VipsImage *in;
VipsImage *out;
if( vips_call( "negative", in, &amp;out, "image_max", 128, NULL ) )
... error
</programlisting>
</para>
<para>
Unfortunately that will do almost no compile-time type checking,
so all vips operations have a tiny extra wrapper to add a bit of
safety. For example:
<programlisting language="C">
static int
negative( VipsImage *in, VipsImage **out, ... )
{
va_list ap;
int result;
va_start( ap, out );
result = vips_call_split( "negative", ap, in, out );
va_end( ap );
return( result );
}
</programlisting>
</para>
<para>
And now you can write:
<programlisting language="C">
if( negative( in, &amp;out, "image_max", 128, NULL ) )
... error
</programlisting>
and it's at least a bit safer.
</para>
</refsect1>
<refsect1 id="extending-othertypes">
<title>Other types of operation</title>
<para>
Change the _build() function to make other types of operation.
</para>
<para>
Use vips_image_generate() with vips_start_many() to make operations
which demand pixels from more than one image at once, such as image
plus image.
</para>
<para>
Use vips_sink() instead of vips_image_generate() to loop over an image
and calculate a value. vips uses this for the statistics operations,
like vips_avg().
</para>
<para>
Use vips_image_wio_input() to get an entire image into memory so you
can read it with a pointer. This will obviously not scale well to
very large images, but some operations, like FFTs or flood-fill, need
the whole image to be available at once.
</para>
<para>
Make area operations, like filters, by enlarging the #VipsRect that
_generate() is given before calling vips_image_prepare(). You can
enlarge the input image, so that the output image is the same size as
the original input, by using vips_embed() within the _build() function.
</para>
<para>
Make things like flips and rotates by making larger changes to the
#VipsRect in _generate().
</para>
</refsect1>

View File

@ -157,8 +157,8 @@ vips_gaussmat_class_init( VipsGaussmatClass *class )
vobject_class->build = vips_gaussmat_build;
VIPS_ARG_DOUBLE( class, "sigma", 2,
_( "Radius" ),
_( "Radius of Gaussian" ),
_( "Sigma" ),
_( "Sigma of Gaussian" ),
VIPS_ARGUMENT_REQUIRED_INPUT,
G_STRUCT_OFFSET( VipsGaussmat, sigma ),
0.000001, 10000.0, 1.0 );

View File

@ -43,6 +43,8 @@
* - use openslide_detect_vendor() on >= 3.4.0
* 30/7/14
* - add autocrop toggle
* 9/8/14
* - do argb -> rgba for associated as well
*/
/*
@ -390,6 +392,45 @@ vips__openslide_read_header( const char *filename, VipsImage *out,
return( 0 );
}
/* Convert from ARGB to RGBA and undo premultiplication.
*
* We throw away transparency. Formats like Mirax use transparent + bg
* colour for areas with no useful pixels. But if we output
* transparent pixels and then convert to RGB for jpeg write later, we
* would have to pass the bg colour down the pipe somehow. The
* structure of dzsave makes this tricky.
*
* We could output plain RGB instead, but that would break
* compatibility with older vipses.
*/
static void
argb2rgba( uint32_t *buf, int n, uint32_t bg )
{
int i;
for( i = 0; i < n; i++ ) {
uint32_t *p = buf + i;
uint32_t x = *p;
uint8_t a = x >> 24;
VipsPel *out = (VipsPel *) p;
if( a != 0 ) {
out[0] = 255 * ((x >> 16) & 255) / a;
out[1] = 255 * ((x >> 8) & 255) / a;
out[2] = 255 * (x & 255) / a;
out[3] = 255;
}
else {
/* Use background color.
*/
out[0] = (bg >> 16) & 255;
out[1] = (bg >> 8) & 255;
out[2] = bg & 255;
out[3] = 255;
}
}
}
static int
vips__openslide_generate( VipsRegion *out,
void *_seq, void *_rslide, void *unused, gboolean *stop )
@ -401,7 +442,6 @@ vips__openslide_generate( VipsRegion *out,
uint32_t *buf = (uint32_t *) VIPS_REGION_ADDR( out, r->left, r->top );
const char *error;
int i;
VIPS_DEBUG_MSG( "vips__openslide_generate: %dx%d @ %dx%d\n",
r->width, r->height, r->left, r->top );
@ -434,39 +474,9 @@ vips__openslide_generate( VipsRegion *out,
return( -1 );
}
/* Convert from ARGB to RGBA and undo premultiplication. Since we are
* inside a cache, we know buf must be continuous.
*
* We throw away transparency. Formats like Mirax use transparent + bg
* colour for areas with no useful pixels. But if we output
* transparent pixels and then convert to RGB for jpeg write later, we
* would have to pass the bg colour down the pipe somehow. The
* structure of dzsave makes this tricky.
*
* We could output plain RGB instead, but that would break
* compatibility with older vipses.
/* Since we are inside a cache, we know buf must be continuous.
*/
for( i = 0; i < n; i++ ) {
uint32_t *p = buf + i;
uint32_t x = *p;
uint8_t a = x >> 24;
VipsPel *out = (VipsPel *) p;
if( a != 0 ) {
out[0] = 255 * ((x >> 16) & 255) / a;
out[1] = 255 * ((x >> 8) & 255) / a;
out[2] = 255 * (x & 255) / a;
out[3] = 255;
}
else {
/* Use background color.
*/
out[0] = (bg >> 16) & 255;
out[1] = (bg >> 8) & 255;
out[2] = bg & 255;
out[3] = 255;
}
}
argb2rgba( buf, n, bg );
return( 0 );
}
@ -519,6 +529,7 @@ vips__openslide_read_associated( const char *filename, VipsImage *out,
{
ReadSlide *rslide;
VipsImage *raw;
uint32_t *buf;
const char *error;
VIPS_DEBUG_MSG( "vips__openslide_read_associated: %s %s\n",
@ -532,14 +543,15 @@ vips__openslide_read_associated( const char *filename, VipsImage *out,
if( !(rslide = readslide_new( filename, raw, 0, FALSE, associated )) ||
vips_image_write_prepare( raw ) )
return( -1 );
openslide_read_associated_image( rslide->osr, rslide->associated,
(uint32_t *) VIPS_IMAGE_ADDR( raw, 0, 0 ) );
buf = (uint32_t *) VIPS_IMAGE_ADDR( raw, 0, 0 );
openslide_read_associated_image( rslide->osr, rslide->associated, buf );
error = openslide_get_error( rslide->osr );
if( error ) {
vips_error( "openslide2vips",
_( "reading associated image: %s" ), error );
return( -1 );
}
argb2rgba( buf, raw->Xsize * raw->Ysize, rslide->bg );
if( vips_image_write( raw, out ) )
return( -1 );

View File

@ -48,6 +48,8 @@
* still set color_type to alpha
* 16/7/13
* - more robust error handling from libpng
* 9/8/14
* - don't check profiles, helps with libpng >=1.6.11
*/
/*
@ -191,7 +193,8 @@ read_new( VipsImage *out, gboolean readbehind )
#ifdef PNG_SKIP_sRGB_CHECK_PROFILE
/* Prevent libpng (>=1.6.11) verifying sRGB profiles.
*/
png_set_option( read->pPng, PNG_SKIP_sRGB_CHECK_PROFILE, PNG_OPTION_ON );
png_set_option( read->pPng,
PNG_SKIP_sRGB_CHECK_PROFILE, PNG_OPTION_ON );
#endif /*PNG_SKIP_sRGB_CHECK_PROFILE*/
/* Catch PNG errors from png_create_info_struct().
@ -726,7 +729,8 @@ write_new( VipsImage *in )
#ifdef PNG_SKIP_sRGB_CHECK_PROFILE
/* Prevent libpng (>=1.6.11) verifying sRGB profiles.
*/
png_set_option( write->pPng, PNG_SKIP_sRGB_CHECK_PROFILE, PNG_OPTION_ON );
png_set_option( write->pPng,
PNG_SKIP_sRGB_CHECK_PROFILE, PNG_OPTION_ON );
#endif /*PNG_SKIP_sRGB_CHECK_PROFILE*/
/* Catch PNG errors from png_create_info_struct().

View File

@ -43,6 +43,9 @@ int vips_shrink( VipsImage *in, VipsImage **out,
__attribute__((sentinel));
int vips_similarity( VipsImage *in, VipsImage **out, ... )
__attribute__((sentinel));
int vips_resize( VipsImage *in, VipsImage **out,
double h_scale, double v_scale, ... )
__attribute__((sentinel));
int vips_affine( VipsImage *in, VipsImage **out,
double a, double b, double c, double d, ... )
__attribute__((sentinel));

View File

@ -9,6 +9,7 @@ libresample_la_SOURCES = \
quadratic.c \
resample.c \
similarity.c \
resize.c \
presample.h \
shrink.c \
interpolate.c \

View File

@ -1,4 +1,4 @@
/* im_affine() ... affine transform with a supplied interpolator.
/* affine transform with a supplied interpolator.
*
*
* Copyright N. Dessipris

View File

@ -115,10 +115,12 @@ vips_resample_operation_init( void )
extern GType vips_quadratic_get_type( void );
extern GType vips_affine_get_type( void );
extern GType vips_similarity_get_type( void );
extern GType vips_resize_get_type( void );
vips_shrink_get_type();
vips_quadratic_get_type();
vips_affine_get_type();
vips_similarity_get_type();
vips_resize_get_type();
}

196
libvips/resample/resize.c Normal file
View File

@ -0,0 +1,196 @@
/* resize an image ... a simple wrapper over affine
*
* 13/8/14
* - from affine.c
*/
/*
This file is part of VIPS.
VIPS 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.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
02110-1301 USA
*/
/*
These files are distributed with VIPS - http://www.vips.ecs.soton.ac.uk
*/
/*
#define DEBUG_VERBOSE
#define DEBUG
*/
#ifdef HAVE_CONFIG_H
#include <config.h>
#endif /*HAVE_CONFIG_H*/
#include <vips/intl.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <math.h>
#include <limits.h>
#include <vips/vips.h>
#include <vips/debug.h>
#include <vips/internal.h>
#include <vips/transform.h>
#include "presample.h"
typedef struct _VipsResize {
VipsResample parent_instance;
double h_scale;
double v_scale;
VipsInterpolate *interpolate;
double idx;
double idy;
} VipsResize;
typedef VipsResampleClass VipsResizeClass;
G_DEFINE_TYPE( VipsResize, vips_resize, VIPS_TYPE_RESAMPLE );
static int
vips_resize_build( VipsObject *object )
{
VipsResample *resample = VIPS_RESAMPLE( object );
VipsResize *resize = (VipsResize *) object;
VipsImage **t = (VipsImage **)
vips_object_local_array( object, 4 );
double a, b, c, d;
if( VIPS_OBJECT_CLASS( vips_resize_parent_class )->build( object ) )
return( -1 );
a = resize->h_scale;
b = 0.0;
c = 0.0;
d = resize->v_scale;
if( vips_affine( resample->in, &t[0], a, b, c, d,
"interpolate", resize->interpolate,
"idx", resize->idx,
"idy", resize->idy,
NULL ) ||
vips_image_write( t[0], resample->out ) )
return( -1 );
return( 0 );
}
static void
vips_resize_class_init( VipsResizeClass *class )
{
GObjectClass *gobject_class = G_OBJECT_CLASS( class );
VipsObjectClass *vobject_class = VIPS_OBJECT_CLASS( class );
VipsOperationClass *operation_class = VIPS_OPERATION_CLASS( class );
VIPS_DEBUG_MSG( "vips_resize_class_init\n" );
gobject_class->set_property = vips_object_set_property;
gobject_class->get_property = vips_object_get_property;
vobject_class->nickname = "resize";
vobject_class->description = _( "resize an image" );
vobject_class->build = vips_resize_build;
operation_class->flags = VIPS_OPERATION_SEQUENTIAL;
VIPS_ARG_DOUBLE( class, "h_scale", 113,
_( "Horizontal scale factor" ),
_( "Scale image by this factor in the horizontal axis" ),
VIPS_ARGUMENT_REQUIRED_INPUT,
G_STRUCT_OFFSET( VipsResize, h_scale ),
0, 10000000, 0 );
VIPS_ARG_DOUBLE( class, "v_scale", 114,
_( "Vertical scale factor" ),
_( "Scale image by this factor in the vertical axis" ),
VIPS_ARGUMENT_REQUIRED_INPUT,
G_STRUCT_OFFSET( VipsResize, v_scale ),
0, 10000000, 0 );
VIPS_ARG_INTERPOLATE( class, "interpolate", 2,
_( "Interpolate" ),
_( "Interpolate pixels with this" ),
VIPS_ARGUMENT_OPTIONAL_INPUT,
G_STRUCT_OFFSET( VipsResize, interpolate ) );
VIPS_ARG_DOUBLE( class, "idx", 115,
_( "Input offset" ),
_( "Horizontal input displacement" ),
VIPS_ARGUMENT_OPTIONAL_INPUT,
G_STRUCT_OFFSET( VipsResize, idx ),
-10000000, 10000000, 0 );
VIPS_ARG_DOUBLE( class, "idy", 116,
_( "Input offset" ),
_( "Vertical input displacement" ),
VIPS_ARGUMENT_OPTIONAL_INPUT,
G_STRUCT_OFFSET( VipsResize, idy ),
-10000000, 10000000, 0 );
}
static void
vips_resize_init( VipsResize *resize )
{
}
/**
* vips_resize:
* @in: input image
* @out: output image
* @h_scale: horizontal scale factor
* @v_scale: vertical scale factor
*
* Optional arguments:
*
* @interpolate: interpolate pixels with this
* @idx: input horizontal offset
* @idy: input vertical offset
*
* @interpolate defaults to bilinear.
*
* @idx, @idy default to zero.
*
* See also: vips_shrink(), vips_affine(), #VipsInterpolate.
*
* Returns: 0 on success, -1 on error
*/
int
vips_resize( VipsImage *in, VipsImage **out,
double h_scale, double v_scale, ... )
{
va_list ap;
int result;
va_start( ap, v_scale );
result = vips_call_split( "affine", ap, in, out, h_scale, v_scale );
va_end( ap );
return( result );
}

View File

@ -1,8 +1,8 @@
/* simple wrapper over vips_similarity() to make scale / rotate easy from the
/* simple wrapper over vips_affine() to make scale / rotate easy from the
* command-line
*
* 3/10/13
* - from similarity.c
* - from affine.c
* 25/10/13
* - oops, reverse rotation direction to match the convention used in the
* rest of vips

View File

@ -52,9 +52,12 @@
* 30/6/14
* - fix interlaced thumbnail output, thanks lovell
* 3/8/14
* - box shrink less, use interpolator more, if the window_size is large
* - box shrink less, use interpolator more, if window_size is large
* enough
* - default to bicubic + nosharpen if bicubic is available
* - default to bicubic if available
* - add an anti-alias filter between shrink and affine
* - support CMYK
* - use SEQ_UNBUF for a memory saving
*/
#ifdef HAVE_CONFIG_H
@ -72,7 +75,7 @@
#define ORIENTATION ("exif-ifd0-Orientation")
/* Default settings. We change the default to bicubic + nosharpen in main() if
/* Default settings. We change the default to bicubic in main() if
* this vips has been compiled with bicubic support.
*/
@ -316,7 +319,7 @@ thumbnail_open( VipsObject *process, const char *filename )
jpegshrink );
if( !(im = vips_image_new_from_file( filename,
"access", VIPS_ACCESS_SEQUENTIAL,
"access", VIPS_ACCESS_SEQUENTIAL_UNBUFFERED,
"shrink", jpegshrink,
NULL )) )
return( NULL );
@ -325,7 +328,7 @@ thumbnail_open( VipsObject *process, const char *filename )
/* All other formats.
*/
if( !(im = vips_image_new_from_file( filename,
"access", VIPS_ACCESS_SEQUENTIAL,
"access", VIPS_ACCESS_SEQUENTIAL_UNBUFFERED,
NULL )) )
return( NULL );
}
@ -344,7 +347,7 @@ thumbnail_interpolator( VipsObject *process, VipsImage *in )
calculate_shrink( in, &residual, NULL );
/* For images smaller than the thumbnail, we upscale with nearest
* neighbor. Otherwise we makes thumbnails that look fuzzy and awful.
* neighbor. Otherwise we make thumbnails that look fuzzy and awful.
*/
if( !(interp = VIPS_INTERPOLATE( vips_object_new_from_string(
g_type_class_ref( VIPS_TYPE_INTERPOLATE ),
@ -397,6 +400,7 @@ thumbnail_shrink( VipsObject *process, VipsImage *in,
int tile_width;
int tile_height;
int nlines;
double sigma;
/* RAD needs special unpacking.
*/
@ -411,12 +415,16 @@ thumbnail_shrink( VipsObject *process, VipsImage *in,
}
/* In linear mode, we import right at the start.
*
* We also have to import the whole image if it's CMYK, since
* vips_colourspace() (see below) doesn't know about CMYK.
*
* This is only going to work for images in device space. If you have
* an image in PCS which also has an attached profile, strange things
* will happen.
*/
if( linear_processing &&
if( (linear_processing ||
in->Type == VIPS_INTERPRETATION_CMYK) &&
in->Coding == VIPS_CODING_NONE &&
(in->BandFmt == VIPS_FORMAT_UCHAR ||
in->BandFmt == VIPS_FORMAT_USHORT) &&
@ -474,6 +482,7 @@ thumbnail_shrink( VipsObject *process, VipsImage *in,
* has been used ... but it never will, since thread1 will block on
* this cache lock.
*/
vips_get_tile_size( in,
&tile_width, &tile_height, &nlines );
if( vips_tilecache( in, &t[4],
@ -482,12 +491,39 @@ thumbnail_shrink( VipsObject *process, VipsImage *in,
"max_tiles", (nlines * 2) / 10,
"access", VIPS_ACCESS_SEQUENTIAL,
"threaded", TRUE,
NULL ) ||
vips_affine( t[4], &t[5], residual, 0, 0, residual,
"interpolate", interp,
NULL ) )
NULL ) )
return( NULL );
in = t[5];
in = t[4];
/* If the final affine will be doing a large downsample, we can get
* nasty aliasing on hard edges. Blur before affine to smooth this out.
*
* Don't blur for very small shrinks, blur with radius 1 for x1.5
* shrinks, blur radius 2 for x2.5 shrinks and above, etc.
*/
sigma = ((1.0 / residual) - 0.5) / 1.5;
if( residual < 1.0 &&
sigma > 0.1 ) {
if( vips_gaussmat( &t[9], sigma, 0.2,
"separable", TRUE,
"integer", TRUE,
NULL ) ||
vips_convsep( in, &t[5], t[9], NULL ) )
return( NULL );
vips_info( "vipsthumbnail", "anti-alias, sigma %g",
sigma );
#ifdef DEBUG
printf( "anti-alias blur matrix is:\n" );
vips_matrixprint( t[9], NULL );
#endif /*DEBUG*/
in = t[5];
}
if( vips_affine( in, &t[6], residual, 0, 0, residual,
"interpolate", interp,
NULL ) )
return( NULL );
in = t[6];
vips_info( "vipsthumbnail", "residual scale by %g", residual );
vips_info( "vipsthumbnail", "%s interpolation",
@ -511,10 +547,10 @@ thumbnail_shrink( VipsObject *process, VipsImage *in,
}
else {
vips_info( "vipsthumbnail", "converting to sRGB" );
if( vips_colourspace( in, &t[6],
if( vips_colourspace( in, &t[7],
VIPS_INTERPRETATION_sRGB, NULL ) )
return( NULL );
in = t[6];
in = t[7];
}
}
else if( export_profile &&
@ -530,13 +566,13 @@ thumbnail_shrink( VipsObject *process, VipsImage *in,
vips_info( "vipsthumbnail",
"exporting with profile %s", export_profile );
if( vips_icc_transform( in, &t[6], export_profile,
if( vips_icc_transform( in, &t[7], export_profile,
"input_profile", import_profile,
"embedded", TRUE,
NULL ) )
return( NULL );
in = t[6];
in = t[7];
}
/* If we are upsampling, don't sharpen, since nearest looks dumb
@ -687,13 +723,11 @@ main( int argc, char **argv )
textdomain( GETTEXT_PACKAGE );
setlocale( LC_ALL, "" );
/* Does this vips have bicubic? Default to that + nosharpen if it
/* Does this vips have bicubic? Default to that if it
* does.
*/
if( vips_type_find( "VipsInterpolate", "bicubic" ) ) {
if( vips_type_find( "VipsInterpolate", "bicubic" ) )
interpolator = "bicubic";
convolution_mask = "none";
}
context = g_option_context_new( _( "- thumbnail generator" ) );