1pod::Prima::Image(3) User Contributed Perl Documentation pod::Prima::Image(3)
2
6 Prima::Image - Bitmap routines
7
9 use Prima qw(Application);
10 # create a new image from scratch
12 my $i = Prima::Image-> new(
13 width => 32,
14 height => 32,
15 type => im::BW, # same as im::bpp1 | im::GrayScale
16 );
17 # draw something
19 $i-> begin_paint;
20 $i-> color( cl::White);
21 $i-> ellipse( 5, 5, 10, 10);
22 $i-> end_paint;
23 # mangle
25 $i-> size( 64, 64);
26 # file operations
28 $i-> save('a.gif') or die "Error saving:$@\n";
29 $i-> load('a.gif') or die "Error loading:$@\n";
30 # draw on screen
32 $::application-> begin_paint;
33 # an image is drawn as specified by its palette
35 $::application-> put_image( 100, 100, $i);
36 # a bitmap is drawn as specified by destination device colors
38 $::application-> set( color => cl::Red, backColor => cl::Green);
39 $::application-> put_image( 200, 100, $i-> bitmap);
40
42 Prima::Image, Prima::Icon and Prima::DeviceBitmap are classes for
43 bitmap handling, including file and graphic input and output.
44 Prima::Image and Prima::DeviceBitmap are descendants of Prima::Drawable
45 and represent bitmaps, stored in memory. Prima::Icon is a descendant
46 of Prima::Image and contains a transparency mask along with the regular
47 data.
48
50 Images usually are represented as a memory area, where pixel data are
51 stored row-wise. The Prima toolkit is no exception, however, it does
52 not assume that the GUI system uses the same memory format. The
53 implicit conversion routines are called when Prima::Image is about to
54 be drawn onto the screen, for example. The conversions are not always
55 efficient, therefore the Prima::DeviceBitmap class is introduced to
56 represent a bitmap, stored in the system memory in the system pixel
57 format. These two basic classes serve the different needs, but can be
58 easily converted to each other, with "image" and "bitmap" methods.
59 Prima::Image is a more general bitmap representation, capable of file
60 and graphic input and output, plus it is supplied with number of
61 conversion and scaling functions. The Prima::DeviceBitmap class has
62 almost none of additional functionality, and is targeted to efficient
63 graphic input and output.
64 Note: If you're looking for information how to display an image, this
66 is not the manual page. Look either at Prima::ImageViewer, or use
67 "put_image" / "stretch_image" ( Prima::Drawable ) inside your widget's
68 onPaint.
69 Graphic input and output
71 As descendants of Prima::Drawable, all Prima::Image, Prima::Icon and
72 Prima::DeviceBitmap objects are subject to three-state painting mode -
73 normal ( disabled ), painting ( enabled ) and informational.
74 Prima::DeviceBitmap is, however, exists only in the enabled state, and
75 can not be switched to the other two.
76 When an object enters the enabled state, it serves as a canvas, and all
78 Prima::Drawable operations can be performed on it. When the object is
79 back to the disabled state, the graphic information is stored into the
80 object associated memory, in the pixel format, supported by the
81 toolkit. This information can be visualized by using one of
82 "Prima::Drawable::put_image" group methods. If the object enters the
83 enabled state again, the graphic information is presented as an initial
84 state of a bitmap.
85 It must be noted, that if an implicit conversion takes place after an
87 object enters and before it leaves the enabled state, as it is with
88 Prima::Image and Prima::Icon, the bitmap is converted to the system
89 pixel format. During such conversion some information can be lost, due
90 to down-sampling, and there is no way to preserve the information. This
91 does not happen with Prima::DeviceBitmap.
92 Image objects can be drawn upon images, as well as on the screen and
94 Prima::Widget objects. This operation is performed via one of
95 Prima::Drawable::put_image group methods ( see Prima::Drawable), and
96 can be called with the image object disregarding the paint state. The
97 following code illustrates the dualism of an image object, where it can
98 serve both as a drawing surface and as a drawing tool:
99 my $a = Prima::Image-> create( width => 100, height => 100, type => im::RGB);
101 $a-> begin_paint;
102 $a-> clear;
103 $a-> color( cl::Green);
104 $a-> fill_ellipse( 50, 50, 30, 30);
105 $a-> end_paint;
106 $a-> rop( rop::XorPut);
107 $a-> put_image( 10, 10, $a);
108 $::application-> begin_paint;
109 $::application-> put_image( 0, 0, $a);
110 $::application-> end_paint;
111 It must be noted, that "put_image", "stretch_image" and
113 "put_image_indirect" only allow "Prima::Image" descendants to be passed
114 as a source image object. This functionality does not imply that the
115 image is internally switched to the paint-enabled state and back; on
116 the contrary, the painting is performed without switching and using
117 only Prima's own code, without using the system's graphical layer.
118 Another special case is a 1-bit ( monochrome ) DeviceBitmap. When it is
120 drawn upon a drawable with bit depth greater than 1, the drawable's
121 color and backColor properties are used to reflect 1 and 0 bits,
122 respectively. On a 1-bit drawable this does not happen, and the color
123 properties are not used.
124 File input and output
126 Depending on the toolkit configuration, images can be read and written
127 in different formats. This functionality in accessible via "load()" and
128 "save()" methods. Prima::image-load is dedicated to the description of
129 loading and saving parameters, that can be passed to the methods, so
130 they can handle different aspects of file format-specific options, such
131 as multi-frame operations, auto conversion when a format does not
132 support a particular pixel format etc. In this document, "load()" and
133 "save()" methods are illustrated only in their basic, single-frame
134 functionality. When called with no extra parameters, these methods fail
135 only if a disk I/O error occurred or an unknown image format was used.
136 When an image is loaded, the old bitmap memory content is discarded,
138 and the image attributes are changed accordingly to the loaded image.
139 Along with these, an image palette is loaded, if available, and a pixel
140 format is assigned, closest or identical to the pixel format in the
141 image file.
142 Pixel formats
144 Prima::Image supports a number of pixel formats, governed by the
145 "::type" property. It is reflected by an integer value, a combination
146 of "im::XXX" constants. The whole set of pixel formats is represented
147 by colored formats, like, 16-color, 256-color and 16M-color, and by
148 gray-scale formats, mapped to C data types - unsigned char, unsigned
149 short, unsigned long, float and double. The gray-scale formats are
150 further subdivided to real-number formats and complex-number format;
151 the last ones are represented by two real values per pixel, containing
152 the real and the imaginary values.
153 Prima::Image can also be initialized from other formats, that it does
155 not support, but can convert data from. Currently these are represented
156 by a set of permutations of 32-bit RGBA format, and 24-bit BGR format.
157 These formats can only be used in conjunction with "::data" property.
158 The conversions can be performed between any of the supported formats (
160 to do so, "::type" property is to be set-called ). An image of any of
161 these formats can be drawn on the screen, but if the system can not
162 accept the pixel format ( as it is with non-integer or complex formats
163 ), the bitmap data are implicitly converted. The conversion does not
164 change the data if the image is about to be drawn; the conversion is
165 performed only when the image is about to be served as a drawing
166 surface. If, by any reason, it is desired that the pixel format is not
167 to be changed, the "::preserveType" property must be set to 1. It does
168 not prevent the conversion, but it detects if the image was implicitly
169 converted inside "end_paint()" call, and reverts it to its previous
170 pixel format.
171 There are situations, when pixel format must be changed together while
173 down-sampling the image. One of four down-sampling methods can be
174 selected - no halftoning, 8x8 ordered halftoning, error diffusion, and
175 error diffusion combined with optimized palette. These can be set to
176 the "::conversion" property with one of "ict::XXX" constants. When
177 there is no information loss, "::conversion" property is not used.
178 Another special case of conversion is a conversion with a palette. The
180 following calls,
181 $image-> type( im::bpp4);
183 $image-> palette( $palette);
184 and
186 $image-> palette( $palette);
188 $image-> type( im::bpp4);
189 produce different results, but none of these takes into account
191 eventual palette remapping, because "::palette" property does not
192 change bitmap pixel data, but overwrites palette information. A proper
193 call syntax here would be
194 $image-> set(
196 palette => $palette,
197 type => im::bpp4,
198 );
199 This call produces also palette pixel mapping. This syntax is most
201 powerful when conversion is set to those algorithms that can take in
202 the account the existing image pixels, to produce an optimized palette.
203 These are "ict::Optimized" ( by default ) and "ict::Posterization".
204 This syntax not only allows remapping or downsampling to a predefined
205 colors set, but also can be used to limit palette size to a particular
206 number, without knowing the actual values of the final color palette.
207 For example, for an 24-bit image,
208 $image-> set( type => im::bpp8, palette => 32);
210 call would calculate colors in the image, compress them to an optimized
212 palette of 32 cells and finally converts to a 8-bit format.
213 Instead of "palette" property, "colormap" can also be used.
215 Data access
217 The pixel values can be accessed in Prima::Drawable style, via
218 "::pixel" property. However, Prima::Image introduces several helper
219 functions on its own.
220 The "::data" property is used to set or retrieve a scalar
222 representation of bitmap data. The data are expected to be lined up to
223 a 'line size' margin ( 4-byte boundary ), which is calculated as
224 $lineSize = int(( $image->width * ( $image-> type & im::BPP) + 31) / 32) * 4;
226 or returned from the read-only property "::lineSize".
228 This is the line size for the data as lined up internally in memory,
230 however "::data" should not necessarily should be aligned like this,
231 and can be accompanied with a write-only flag 'lineSize' if pixels are
232 aligned differently:
233 $image-> set( width => 1, height=> 2);
235 $image-> type( im::RGB);
236 $image-> set(
237 data => 'RGB----RGB----',
238 lineSize => 7,
239 );
240 print $image-> data, "\n";
241 output: RGB-RGB-
243 Internally, Prima contains images in memory so that the first scanline
245 is the farthest away from the memory start; this is consistent with
246 general Y-axis orientation in Prima drawable terminology, but might be
247 inconvenient when importing data organized otherwise. Another write-
248 only boolean flag "reverse" can be set to 1 so data then are treated as
249 if the first scanline of the image is the closest to the start of data:
250 $image-> set( width => 1, height=> 2, type => im::RGB);
252 $image-> set(
253 data => 'RGB-123-',
254 reverse => 1,
255 );
256 print $image-> data, "\n";
257 output: RGB-123-
259 Although it is possible to perform all kinds of calculations and
261 modification with the pixels, returned by "::data", it is not advisable
262 unless the speed does not matter. Standalone PDL package with help of
263 PDL::PrimaImage package, and Prima-derived IPA package provide routines
264 for data and image analysis. Also, Prima::Image::Magick connects
265 ImageMagick with Prima. Prima::Image itself provides only the simplest
266 statistic information, namely: lowest and highest pixel values, pixel
267 sum, sum of square pixels, mean, variance, and standard deviation.
268 Standalone usage
270 Some of image functionality can be used standalone, with all other
271 parts of the toolkit being uninitialized. The functionality is limited
272 to loading and saving files, and reading and writing pixels (outside
273 begin_paint only). All other calls are ignored. Example:
274 my $i = Prima::Image->new( size => [5,5]);
276 $i->color(cl::Red);
277 $i->bar(0,0,$i->size);
278 $i->save('1.bmp');
279 This feature is useful in non-interactive programs, running in
281 environments with no GUI access, a cgi-script with no access to X11
282 display, for example. Normally, Prima fails to start in such
283 situations, but can be told not to initialize its GUI part by
284 explicitly operating system-dependent options. To do so, invoke
285 use Prima::noX11;
287 in the beginning of your program. See Prima::noX11 for more.
289 Generally the standalone methods support all the OS-specific functions
291 (i.e. color, region, etc), plus they support alpha blending using "rop"
292 property and "rop::alpha($rop, $src_alpha[, $dst_alpha]))" value.
293 See individual methods and properties in API that support standalone
295 usage, and how they differ from system-dependent implementation.
296 Prima::Icon
298 Prima::Icon inherits all properties of Prima::Image, and it also
299 provides a transparency mask of either 1 or 8 bits. This mask can also
300 be loaded and saved into image files, if the format supports
301 transparency information.
302 Similar to Prima::Image::data property, Prima::Icon::mask property
304 provides access to the binary mask data. The mask can be updated
305 automatically, after an icon object was subject to painting, resizing,
306 or other destructive change. The auxiliary properties "::autoMasking"
307 and "::maskColor"/"::maskIndex" regulate mask update procedure. For
308 example, if an icon was loaded with the color ( vs. bitmap )
309 transparency information, the binary mask will be generated anyway, but
310 it will be also recorded that a particular color serves as a
311 transparent indicator, so eventual conversions can rely on the color
312 value, instead of the mask bitmap.
313 If an icon is drawn upon a graphic canvas, the image output is
315 constrained to the mask. On raster displays it is typically simulated
316 by a combination of and- and xor- operation modes, therefore attempts
317 to put an icon with "::rop", different from "rop::CopyPut", usually
318 fail.
319 Layering
321 The term layered window is borrowed from Windows world, and means a
322 window with transparency. In Prima, the property layered is used to
323 select this functionality. The call to
324 "$::application->get_system_value(sv::LayeredWidgets)" can check
325 whether this functionality is available; if not, the property is
326 ignored. By default, widgets can not use layering.
327 A layered drawable uses an extra alpha channel to designate the
329 transparency. Drawing on widgets will also look different - for
330 example, drawing with black color will make the black pixels fully
331 transparent, while other colors will blend with the underlying
332 background, but never in full. Prima provides no functions to draw with
333 alpha effects, and scarce image functions to address the alpha
334 surfaces. Drawing lines, text, etc with blending is delegated to
335 Prima::Cairo which is a separate module. Prima only provides alpha
336 surfaces and bitmaps with an additional alpha channel to draw upon.
337 However, "put_image" / "stretch_image" functions can operate on
338 surfaces with alpha as source and destination drawables. To address the
339 alpha channel on a drawable with Prima, one has to send either an
340 "Prima::Icon" with "maskType(im::bpp8)", or a layered "DeviceBitmap" to
341 these functions.
342 The corresponding "Prima::DeviceBitmap" type is "dbt::Layered", and is
344 fully compatible with layered widgets in the same fashion as
345 "DeviceBitmap" with type "dbt::Pixmap" is fully compatible with normal
346 widgets. One of ways to put a constant alpha value over a rectangle is
347 this, for example:
348 my $a = Prima::Icon->new(
350 width => 1,
351 height => 1,
352 type => im::RGB,
353 maskType => im::bpp8,
354 data => "\0\0\0",
355 mask => chr( $constant_alpha ),
356 );
357 $drawable-> stretch_image( 0, 0, 100, 100, $a, rop::SrcOver );
358 If displaying a picture with pre-existing alpha channel, you'll need to
360 call premultiply_alpha, because picture renderer assumes that pixel
361 values are premultiplied.
362 Even though addressing alpha values of pixels when drawing on layered
364 surfaces is not straighforward, the conversion between images and
365 device bitmaps fully supports alpha pixels. This means that:
366 * When drawing on an icon with 8-bit alpha channel (argb icon), any
368 changes to alpha values of pixels will be transferred back to the mask
369 property after "end_paint"
370 * Calls to "icon" function on DeviceBitmap with type "dbt::Layered"
372 produce identical argb icons. Calls to "bitmap" on argb icos produce
373 identical layered device bitmaps.
374 * Putting argb icons and layered device bitmap on other drawables
376 yields identical results.
377 Putting of argb source surfaces can be only used with two rops,
379 "rop::SrcOver" (default) and "rop::SrcCopy". The former produces
380 blending effect, while the latter copies alpha bits over to the
381 destination surface. Prima internal implementation of "put_image" and
382 "stretch_image" functions extends the allowed set of rops when
383 operating on images outside the begin_paint/end_paint brackets. These
384 rops support 12 Porter-Duff operators, some more "photoshop" operators,
385 and special flags to specify constant alpha values to override the
386 existing alpha channel, if any. See more in "Raster operations" in
387 Prima::Drawable.
388 Caveats: In Windows, mouse events will not be delivered to the layered
390 widget if the pixel under the mouse pointer is fully transparent.
391 See also: examples/layered.pl.
393
395 Prima::Image properties
396 colormap @PALETTE
397 A color palette, used for representing 1, 4, and 8-bit bitmaps,
398 when an image object is to be visualized. @PALETTE contains
399 individual colors component triplets, in RGB format. For example,
400 black-and-white monochrome image may contain colormap as
401 "0,0xffffff".
402 See also "palette".
404 conversion TYPE
406 Selects the type of dithering algorithm to be used for pixel down-
407 sampling. TYPE is one of "ict::XXX" constants:
408 ict::None - no dithering, with static palette or palette optimized by source palette
410 ict::Posterization - no dithering, with optimized palette by source pixels
411 ict::Ordered - fast 8x8 ordered halftone dithering with static palette
412 ict::ErrorDiffusion - error diffusion dithering with static palette
413 ict::Optimized - error diffusion dithering with optimized palette
414 As an example, if a 4x4 color image with every pixel set to
416 RGB(32,32,32), converted to a 1-bit image, the following results
417 occur:
418 ict::None, ict::Posterization:
420 [ 0 0 0 0 ]
421 [ 0 0 0 0 ]
422 [ 0 0 0 0 ]
423 [ 0 0 0 0 ]
424 ict::Ordered:
426 [ 0 0 0 0 ]
427 [ 0 0 1 0 ]
428 [ 0 0 0 0 ]
429 [ 1 0 0 0 ]
430 ict::ErrorDiffusion, ict::Ordered:
432 [ 0 0 1 0 ]
433 [ 0 0 0 1 ]
434 [ 0 0 0 0 ]
435 [ 0 0 0 0 ]
436 Values of these constants are made from "ictp::" in Prima::Const
438 and "ictd::" in Prima::Const constansts.
439 data SCALAR
441 Provides access to the bitmap data. On get-call, returns all bitmap
442 pixels, aligned to 4-byte boundary. On set-call, stores the
443 provided data with same alignment. The alignment can be altered by
444 submitting 'lineSize' write-only flag to set call; the ordering of
445 scan lines can be altered by setting 'reverse' write-only flag (
446 see "Data access" ).
447 height INTEGER
449 Manages the vertical dimension of the image data. On set-call, the
450 image data are changed accordingly to the new height, and depending
451 on "::vScaling" property, the pixel values are either scaled or
452 truncated.
453 lineSize INTEGER
455 A read-only property, returning the length of an image row in
456 bytes, as represented internally in memory. Data returned by
457 "::data" property are aligned with "::lineSize" bytes per row, and
458 setting "::data" expects data aligned with this value, unless
459 "lineSize" is set together with "data" to indicate another
460 alignment. See "Data access" for more.
461 mean
463 Returns mean value of pixels. Mean value is "::sum" of pixel
464 values, divided by number of pixels.
465 palette [ @PALETTE ]
467 A color palette, used for representing 1, 4, and 8-bit bitmaps,
468 when an image object is to be visualized. @PALETTE contains
469 individual color component triplets, in BGR format. For example,
470 black-and-white monochrome image may contain palette as
471 "[0,0,0,255,255,255]".
472 See also "colormap".
474 pixel ( X_OFFSET, Y_OFFSET ) PIXEL
476 Provides per-pixel access to the image data when image object is in
477 disabled paint state.
478 Pixel values for grayscale 1- and 4- bit images are treated
480 specifically, such that like 8-bit function, values cover range
481 between 0 and 255. F.ex. pixel values for grayscale 1 bit images
482 are 0 and 255, not 0 and 1.
483 In paint state same as "Prima::Drawable::pixel".
485 preserveType BOOLEAN
487 If 1, reverts the image type to its old value if an implicit
488 conversion was called during "end_paint()".
489 rangeHi
491 Returns maximum pixel value in the image data.
492 rangeLo
494 Returns minimum pixel value in the image data.
495 scaling INT
497 Declares the scaling strategy when image is resized. Strategies
498 "ist::None" through "ist::Box" are very fast scalers, others not
499 so.
500 Can be one of "ist:::XXX" constants:
502 ist::None - image will be either stripped (when downsizing)
504 or padded (when upsizing) with zeros
505 ist::Box - image will be scaled using simple box transform
506 ist::BoxX - columns will behave same as in ist::None,
507 rows will behave same as in ist::Box
508 ist::BoxY - rows will behave same as in ist::None,
509 columns will behave same as in ist::Box
510 ist::AND - when row or columns is to be shrunk, leftover pixels
511 will be AND-end together (for black on white)
512 ( does not work for floating poing pixels )
513 ist::OR - when row or columns is to be shrunk, leftover pixels
514 will be OR-end together (for white on black)
515 ( does not work for floating poing pixels )
516 ist::Triangle - bilinear interpolation
517 ist::Quadratic - 2rd order (quadratic) B-Spline approximation of Gaussian
518 ist::Sinc - sine function
519 ist::Hermite - B-Spline interpolation
520 ist::Cubic - 3rd order (cubic) B-Spline approximation of Gaussian
521 ist::Gaussian - Gaussian transform with gamma=0.5
522 Note: Resampling scaling algorithms (those greater than
524 "ist::Box"), when applied to Icons with 1-bit icon mask, will
525 silently convert the mask in 8-bit and apply the same scaling
526 algorithm to it. This will have great smoothing effect on mask
527 edges if the system supports ARGB layering (see "Layering" ).
528 size WIDTH, HEIGHT
530 Manages dimensions of the image. On set-call, the image data are
531 changed accordingly to the new dimensions, and depending on
532 "::scaling" property, the pixel values are either scaled or
533 truncated.
534 stats ( INDEX ) VALUE
536 Returns one of calculated values, that correspond to INDEX, which
537 is one of the following "is::XXX" constants:
538 is::RangeLo - minimum pixel value
540 is::RangeHi - maximum pixel value
541 is::Mean - mean value
542 is::Variance - variance
543 is::StdDev - standard deviation
544 is::Sum - sum of pixel values
545 is::Sum2 - sum of squares of pixel values
546 The values are re-calculated on request and cached. On set-call
548 VALUE is stored in the cache, and is returned on next get-call.
549 The cached values are discarded every time the image data changes.
550 These values are also accessible via set of alias properties:
552 "::rangeLo", "::rangeHi", "::mean", "::variance", "::stdDev",
553 "::sum", "::sum2".
554 stdDev
556 Returns standard deviation of the image data. Standard deviation
557 is the square root of "::variance".
558 sum Returns sum of pixel values of the image data
560 sum2
562 Returns sum of squares of pixel values of the image data
563 type TYPE
565 Governs the image pixel format type. TYPE is a combination of
566 "im::XXX" constants. The constants are collected in groups:
567 Bit-depth constants provide size of pixel is bits. Their actual
569 value is same as number of bits, so "im::bpp1" value is 1,
570 "im::bpp4" - 4, etc. The valid constants represent bit depths from
571 1 to 128:
572 im::bpp1
574 im::bpp4
575 im::bpp8
576 im::bpp16
577 im::bpp24
578 im::bpp32
579 im::bpp64
580 im::bpp128
581 The following values designate the pixel format category:
583 im::Color
585 im::GrayScale
586 im::RealNumber
587 im::ComplexNumber
588 im::TrigComplexNumber
589 im::SignedInt
590 Value of "im::Color" is 0, whereas other category constants
592 represented by unique bit value, so combination of "im::RealNumber"
593 and "im::ComplexNumber" is possible.
594 There also several mnemonic constants defined:
596 im::Mono - im::bpp1
598 im::BW - im::bpp1 | im::GrayScale
599 im::16 - im::bpp4
600 im::Nibble - im::bpp4
601 im::256 - im::bpp8
602 im::RGB - im::bpp24
603 im::Triple - im::bpp24
604 im::Byte - gray 8-bit unsigned integer
605 im::Short - gray 16-bit unsigned integer
606 im::Long - gray 32-bit unsigned integer
607 im::Float - float
608 im::Double - double
609 im::Complex - dual float
610 im::DComplex - dual double
611 im::TrigComplex - dual float
612 im::TrigDComplex - dual double
613 Bit depths of float- and double- derived pixel formats depend on a
615 platform.
616 The groups can be masked out with the mask values:
618 im::BPP - bit depth constants
620 im::Category - category constants
621 im::FMT - extra format constants
622 The extra formats are the pixel formats, not supported by "::type",
624 but recognized within the combined set-call, like
625 $image-> set(
627 type => im::fmtBGRI,
628 data => 'BGR-BGR-',
629 );
630 The data, supplied with the extra image format specification will
632 be converted to the closest supported format. Currently, the
633 following extra pixel formats are recognized:
634 im::fmtBGR
636 im::fmtRGBI
637 im::fmtIRGB
638 im::fmtBGRI
639 im::fmtIBGR
640 variance
642 Returns variance of pixel values of the image data. Variance is
643 "::sum2", divided by number of pixels minus square of "::sum" of
644 pixel values.
645 width INTEGER
647 Manages the horizontal dimension of the image data. On set-call,
648 the image data are changed accordingly to the new width, and
649 depending on "::scaling" property, the pixel values are either
650 scaled or truncated.
651 Prima::Icon properties
653 autoMasking TYPE
654 Selects whether the mask information should be updated
655 automatically with "::data" change or not. Every "::data" change is
656 mirrored in "::mask", using TYPE, one of "am::XXX" constants:
657 am::None - no mask update performed
659 am::MaskColor - mask update based on ::maskColor property
660 am::MaskIndex - mask update based on ::maskIndex property
661 am::Auto - mask update based on corner pixel values
662 The "::maskColor" color value is used as a transparent color if
664 TYPE is "am::MaskColor". The transparency mask generation
665 algorithm, turned on by "am::Auto" checks corner pixel values,
666 assuming that majority of the corner pixels represents a
667 transparent color. Once such color is found, the mask is generated
668 as in "am::MaskColor" case.
669 "::maskIndex" is the same as "::maskColor", except that it points
671 to a specific color index in the palette.
672 When image "::data" is stretched, "::mask" is stretched
674 accordingly, disregarding the "::autoMasking" value.
675 mask SCALAR
677 Provides access to the transparency bitmap. On get-call, returns
678 all bitmap pixels, aligned to 4-byte boundary in 1-bit format. On
679 set-call, stores the provided transparency data with same
680 alignment.
681 maskColor COLOR
683 When "::autoMasking" set to "am::MaskColor", COLOR is used as a
684 transparency value.
685 maskIndex INDEX
687 When "::autoMasking" set to "am::MaskIndex", INDEXth color in teh
688 current palette is used as a transparency value.
689 maskType INTEGER
691 Is either "im::bpp1" (1) or "im::bpp8" (8). The latter can be used
692 as a layered (argb) source surface to draw with blending effect.
693 Prima::DeviceBitmap properties
695 type INTEGER
696 A read-only property, that can only be set during creation,
697 reflects whether the system bitmap is black-and-white 1-bit
698 ("dbt::Bitmap"), is colored and compatible with widgets
699 ("dbt::Pixmap"), or is colored with alpha channel and compatible
700 with layered widgets ("dbt::Layered").
701 The color depth of a bitmap can be read via "get_bpp()" method;
703 monochrome bitmaps always have bit depth of 1, layered bitmaps have
704 bit depth of 32.
705 Prima::Image methods
707 bar X1, Y1, X2, Y2
708 Outside the paint state uses owen implementation for drawing a
709 rectangular shape. The following properties are respected:
710 "color", "backColor", "rop", "rop2", "fillPattern",
711 "fillPatternOffset", "region". "rop2" accepts either "rop::CopyPut"
712 or "rop::NoOper" values, to produce either opaque or transparent
713 fill pattern application.
714 Inside the paint state is identical to "Drawable::bar".
716 bitmap
718 Returns newly created Prima::DeviceBitmap instance, with the image
719 dimensions and with the bitmap pixel values copied to.
720 clear [X1, Y1, X2, Y2]
722 Same as "Drawable::clear" but can be used also outside of the paint
723 state.
724 clone %properties
726 Creates a copy of the image and applies %properties. An easy way to
727 create a down-sampled copy, for example.
728 codecs
730 Returns array of hashes, each describing the supported image
731 format. If the array is empty, the toolkit was set up so it can not
732 load and save images.
733 See Prima::image-load for details.
735 This method can be called without object instance.
737 dup Returns a duplicate of the object, a newly created Prima::Image,
739 with all information copied to it. Does not preserve graphical
740 properties (color etc).
741 extract X_OFFSET, Y_OFFSET, WIDTH, HEIGHT
743 Returns a newly created image object with WIDTH and HEIGHT
744 dimensions, initialized with pixel data from X_OFFSET and Y_OFFSET
745 in the bitmap.
746 fill_chord, fill_ellipse, fill_sector, flood_fill
748 Same as "Drawable::" functions but can be used also outside of the
749 paint state.
750 get_bpp
752 Returns the bit depth of the pixel format. Same as "::type &
753 im::BPP".
754 get_handle
756 Returns a system handle for an image object.
757 load (FILENAME or FILEGLOB) [ %PARAMETERS ]
759 Loads image from file FILENAME or stream FILEGLOB into an object,
760 and returns the success flag. The semantics of "load()" is
761 extensive, and can be influenced by PARAMETERS hash. "load()" can
762 be called either in a context of an existing object, then a boolean
763 success flag is returned, or in a class context, then a newly
764 created object ( or "undef" ) is returned. If an error occurs, $@
765 variable contains the error description string. These two
766 invocation semantics are equivalent:
767 my $x = Prima::Image-> create();
769 die "$@" unless $x-> load( ... );
770 and
772 my $x = Prima::Image-> load( ... );
774 die "$@" unless $x;
775 See Prima::image-load for details.
777 NB! When loading from streams on win32, mind "binmode".
779 map COLOR
781 Performs iterative mapping of bitmap pixels, setting every pixel to
782 "::color" property with respect to "::rop" type if a pixel equals
783 to COLOR, and to "::backColor" property with respect to "::rop2"
784 type otherwise.
785 "rop::NoOper" type can be used for color masking.
787 Examples:
789 width => 4, height => 1, data => [ 1, 2, 3, 4]
791 color => 10, backColor => 20, rop => rop::CopyPut
792 rop2 => rop::CopyPut
794 input: map(2) output: [ 20, 10, 20, 20 ]
795 rop2 => rop::NoOper
797 input: map(2) output: [ 1, 10, 3, 4 ]
798 mirror VERTICAL
800 Mirrors the image depending on boolean flag VERTICAL
801 premultiply_alpha CONSTANT_OR_IMAGE
803 Applies premultiplication formula to each pixel
804 pixel = pixel * alpha / 256
806 where alpha either is a constant, or a pixel value in an image
808 put_image, put_image_indirect, stretch_image
810 Same as "Drawable::" functions but can be used also outside of the
811 paint state.
812 Extends raster functionality to access alpha channel either using
814 constant alpha values or "Prima::Icon" as sources. See explanation
815 of "rop::" constants in "Raster operations" in Prima::Drawable.
816 resample SRC_LOW, SRC_HIGH, DEST_LOW, DEST_HIGH
818 Performs linear scaling of gray pixel values from range (SRC_LOW -
819 SRC_HIGH) to range (DEST_LOW - DEST_HIGH). Can be used to visualize
820 gray non-8 bit pixel values, by the code:
821 $image-> resample( $image-> rangeLo, $image-> rangeHi, 0, 255);
823 rotate DEGREES
825 Rotates the image. Where the angle is 90, 180, or 270 degrees, fast
826 pixel flipping is used, otherwise fast Paeth rotation is used.
827 Eventual resampling can be controlled by "scaling" property (
828 probably not worth it for functions with support of more than 1
829 pixel).
830 Resulting images can be 1 pixel too wide due to horizontal shearing
832 applied twice, where in worst cases 1 pixel from the original image
833 can take 3 horizontal pixels on the result.
834 save (FILENAME or FILEGLOB), [ %PARAMETERS ]
836 Stores image data into image file FILENAME or stream FILEGLOB, and
837 returns the success flag. The semantics of "save()" is extensive,
838 and can be influenced by PARAMETERS hash. If error occurs, $@
839 variable contains error description string.
840 Note that when saving to a stream, "codecID" must be explicitly
842 given in %PARAMETERS.
843 See Prima::image-load for details.
845 NB! When saving to streams on win32, mind "binmode".
847 shear X, Y
849 Applies shearing to the image. If the shearing is needed only for
850 one axis, set shear factor for the other one to zero.
851 to_region
853 Creates a new Prima::Region object with the image as the data
854 source.
855 transform @MATRIX
857 Applies generic 2D transform matrix to the image, where matrix is 4
858 numbers.
859 Tries first to split matrix into series of shear and scale
861 transforms using LDU decomposition; if an interim image needs to be
862 too large, fails and returns "false".
863 Rotation matrices can be applied too, however, when angles are
865 close to 90 and 270, either interim images become too big, or
866 defects introduced by shearing become too visible. Therefore the
867 method specifically detects for rotation cases, and uses Paeth
868 rotation algorithm instead, which yields better results. Also, if
869 the angle is detected to be 90, 180, or 270 degrees, fast pixel
870 flipping is used.
871 Eventual resampling can be controlled by "scaling" property.
873 ui_scale %OPTIONS
875 Resizes the image with smooth scaling. Understands "zoom" and
876 "scaling" options. The "zoom" default value is the one in
877 "$::application->uiScaling", the "scaling" default value is
878 "ist::Quadratic" .
879 See also: "uiScaling" in Application
881 Prima::Image events
883 "Prima::Image"-specific events occur only from inside load call, to
884 report image loading progress. Not all codecs (currently JPEG,PNG,TIFF
885 only) are able to report the progress to the caller. See "Loading with
886 progress indicator" in Prima::image-load for details,
887 "watch_load_progress" in Prima::ImageViewer and "load" in
888 Prima::Dialog::ImageDialog for suggested use.
889 HeaderReady EXTRAS
891 Called whenever image header is read, and image dimensions and
892 pixel type is changed accordingly to accomodate image data.
893 "EXTRAS" is the hash to be stored later in "{extras}" key on the
895 object.
896 DataReady X, Y, WIDTH, HEIGHT
898 Called whenever image data that cover area designated by
899 X,Y,WIDTH,HEIGHT is acquired. Use "load" option "eventDelay" to
900 limit the rate of "DataReady" event.
901 Prima::Icon methods
903 alpha ALPHA <X1, Y1, X2, Y2>
904 Same as "Drawable::alpha" but can be used also outside of the paint
905 state.
906 combine DATA, MASK
908 Copies information from DATA and MASK images into "::data" and
909 "::mask" property. DATA and MASK are expected to be images of same
910 dimension.
911 create_combined DATA, MASK
913 Same as "combine", but to be called as constructor.
914 image %opt
916 Renders icon graphics on a newly created Prima::Image object
917 instance upon black background. If $opt{background} is given, it
918 is used instead.
919 premultiply_alpha CONSTANT_OR_IMAGE = undef
921 Applies premultiplication formula to each pixel
922 pixel = pixel * alpha / 256
924 where alpha is the corresponding alpha value for each coordinate.
926 Only applicable when "maskType" is <im::bpp8>.
927 split
929 Returns two new Prima::Image objects of same dimension. Pixels in
930 the first is are duplicated from "::data" storage, in the second -
931 from "::mask" storage.
932 ui_scale %OPTIONS
934 Same as "ui_scale" from "Prima::Image", but with few exceptions: It
935 tries to use "ist::Quadratic" only when the system supports ARGB
936 layering. Otherwise, falls back on "ist::Box" scaling algorithm,
937 and also limits the zoom factor to integers (2x, 3x etc) only,
938 because when displayed, the smooth-scaled color plane will not
939 match mask plane downgraded to 0/1 mask, and because box-scaling
940 with non-integer zooms looks ugly.
941 Prima::DeviceBitmap methods
943 dup Returns a duplicate of the object, a newly created
944 Prima::DeviceBitmap, with all information copied to it. Does not
945 preserve graphical properties (color etc).
946 icon
948 Returns a newly created Prima::Icon object instance, with the pixel
949 information copied from the object. If the bitmap is layered,
950 returns icons with maskType set to "im::bpp8".
951 image
953 Returns a newly created Prima::Image object instance, with the
954 pixel information copied from the object.
955 get_handle
957 Returns a system handle for a system bitmap object.
958
960 Dmitry Karasik, <dmitry@karasik.eu.org>.
961
963 Prima, Prima::Drawable, Prima::image-load, Prima::codecs.
964 PDL, PDL::PrimaImage, IPA
966 ImageMagick, Prima::Image::Magick
968perl v5.34.0 2021-07-22 pod::Prima::Image(3)