Here I will write about the advances made from latest stable version 2.9.0 to next major version 3.0.0.
- 2021/09/02: New release 2.9.9 of our G’MIC Image Processing Framework.
From 2.9.9 to 3.0.0 (work in progress)
New features:
- [gmic_qt-300] New filter Patterns / Triangular Pattern, generates random patterns based on random subdivisions of triangles.
- [gmic_qt-300] New filter Deformations / Breaks. Somehow similar to filter Deformations / Crease but with more sharp and triangular shapes (request from David Revoy).
-
[math-core-300] New functions
da_size(),da_insert(),da_push()andda_remove()to natively manage dynamic arrays in math expressions.
Improvements / Modifications:
- [gmic-qt] Recode filter Colors / Chromatic Aberrations from scratch, adding lot of new controls to render more different types of chromatic aberrations.
-
[stdlib-300] Recode command
delaunayfrom scratch, allowing to output the delaunay triangulation as two sets of points/faces coordinates. -
[stdlib-300]_ Recode command
minimal_pathto make it faster. -
[stdlib-300]_ Recode command
uniform_distributionto get more uniform results for odd number of points. -
[core-300] Improve processing of input filenames with extensions
arw,cr2,nefand.dng. When read, they are now passed todcrawas a preprocessing step to get a 16-bits/channel input image. -
[core-300] Command
distanceindijkstra + pathmode : path is now returned as the second channel of the distance image, rather than a second output image.
Bugfixes:
-
[math-core-300] Fix bug when function
polygon()is called with invalid arguments (s.a.nancoordinates). -
[math-core-300] Fix code to correctly parse conditions as
var%==sth.
From 2.9.8 to 2.9.9
New features:
- [gmic-qt-299] New filter Artistic / Paint With Brush, simulates the conversion of an input photograph to a painting.
-
[core-299] Explicit specialization of
+commandis now possible when writing custom commands. Useful to avoid image copy when a custom command inserts new images on the list without having to modify the input selection (e.g.psnrormselike commands). -
[stdlib-299] New command
input_csvreads the content of a.csvfile (usual text file format for storing tabular data).
Improvements / Modifications:
-
[core-299] Commands
plotandgraphnow accept formulas that output a vector-valued function. -
[core-299] Builtin commands that insert new images on the list can be called indifferently with a starting
+sign or not (e.g.+pass,+shared,+input). -
[core-299] Command
histogramnow always requires a valid argument ((noarg)not accepted). -
[stdlib-299] Commands
mseandpsnrhave been recoded as custom commands (rather than builtin).
Bug fixes:
-
[core-299] Fix input expression
(list_of_values:c)which was not unrolled along thecaxis.
From 2.9.7 to 2.9.8:
Release date: 2021/07/16
New features:
-
[gmic-qt-298] The API of G’MIC-Qt has been revamped and allows more flexibility for the host software. The stand-alone version
gmic_qthas many new options to control it from the command line, e.g. to allow batch processing. -
[stdlib-298] Started the implementation of a machine learning library (new commands whose names start with
nn_*). Hopefully, this will enable the development of useful ML-based image processing filters in G’MIC in the future. -
[stdlib-298] New command
meigenreturns an approximation of themlargest eigenvalues of a (large) square matrix, using the Arnoldi iteration method. New math functionmeig()does the same, inside the math parser (included inmath_lib). -
[stdlib-298] New command
arg2imgsplits specified list of arguments and return each as a new image (as a null-terminated string). -
[math-core-298] New function
fill()fills the content of a variable (particularly vector-valued) from a given expression, as inV = vector10(); fill(V,k,k^2 + k + 2);(3-arguments version), orfill(V,u(-1,1))(2-arguments version). -
[math-core-298] New function
repeat()iterates over the content of a variable, as inV = vector10(); repeat(10,k,print(k^2));(3-arguments version), orrepeat(10,print(u))(2-arguments version). -
[math-code-298] New function
set(expr,'variable_name')sets the specified variable name with the value of expressionexpr(seen as a scalar value or a string). -
[math-core-298] New functions
deg2rad()andrad2deg()in math parser, to convert angles in degree from/to radians. -
[math-core-298] New postfix operator
°converts degrees to radians, as incos(45°), that is equivalent tocos(45*pi/180). -
[math-core-298] New predefined constant value
nwhich stands for the maximum number of threads a mathematical expression may use for its evaluation. -
[core-298] When loading a
.pngor.tifffile, the bit depth of the input image (8,16or32) is now returned to the status. -
[core-298] Added new ‘string prepend operator’ that can be used to update variables, as in
str=foo str..=before_that is equivalent tostr=before_foo.
Improvements / Modifications:
-
[stdlib-298] Command
apply_curve: recoded to manage any kind of value ranges. -
[math-core-298] Function
store()is now able to store an input vector as an image of any size (it just pads with zeros in case the requested size is larger than the vector’s size). -
[math-core-298] Improved JIT compiler to avoid unnecessary copy of memory blocks when initializing variables. As a result, complex math expressions requires less memory to be evaluated (and less bytecode instructions, so it’s probably slightly faster).
-
[math-core-298] Function
rot(): angle argument must be specified in radians now (use new postfix operator°to specify angles in degrees). -
[math_core-298] Function
_(comment)(that does nothing) can now be followed by a value, without requiring a;separator. -
[core-298]
webpimages with alpha-channels now load as 4-channels images into G’MIC (was keeping only RGB channels before). -
[core-298] Optimized the storage of global variables. This greatly improves speed of commands that use a lot of global variables (e.g.
parse_gui). -
[core-298] Optimized the resizing of images (with command
resize) when moving average interpolation is used (typically for image downsizing). It now uses a multi-threaded approach. -
[core-298] Command
textnow accepts an ‘empty’ string as a first argument. -
[core-298] Commands
convolveandcorrelatehave been mostly recoded, leading to visible optimizations, particularly for large kernels (\approx +50% speed gain). -
[web-298] Binary packages for Ubuntu 21.04 ‘Hirsute’ are now available.
Bug fixes:
-
[core-298] When forcing file format in input/output expressions, s.a.
ext:filename,extwas not accepting numbers. This is fixed. -
[core-298] Storage of image names in
.gmzformat could be wrong. Now fixed. -
[core-298] Fix commands
convolveandcorrelatewhenchannel_mode==0and whenchannel_mode==2with small kernels (<=5). -
[math-core-298] Fixed bug about thread number
tthat could sometimes take value0for two different threads. -
[stdlib-298] Fixed
morph_rbf: old version ofresizewas used instead ofresize_as_image. -
[stdlib-298] Fixed
help: when two commands had the same name, the help displayed for that command was not corresponding to the command with the higher priority.
From 2.9.6 to 2.9.7
New features:
- [gmic-qt-297] New filter Light & Shadows / Guided Light Rays renders light rays, guided by a shape, drawn on a top transparent layer by the user.
- [gmic-qt-297] New filter Light & Shadows / Equalize Light is a filter that proposes different ways to equalize light changes in an image.
- [gmic-qt-297] New filter Rendering / Sine Curve, a very versatile filter to render 2D/3D sinusoidal curves, with a lot of tunable parameters.
Improvements / Modifications:
- [gmic-qt-297] Filter Artistic / Stylize has 6 new predefined styles, provided by Christine Garner, which simulates different drawing styles (pencils,pastel,charcoal).
- [gmic-qt-297] Work from Jan Tojnar allow to compile the G’MIC-Qt plug-in for the latest development version v.2.99 of GIMP.
- [gmic-qt-297] For Windows users : plug-in now works with GIMP 2.10.24. The new GIMP release introduced upgraded dll that were causing issues with our plug-in binary.
-
[gmic-qt-297] For Windows users: provided archives now include an additional
styles/folder that avoid the G’MIC-Qt plug-in to look like a Windows95 application. -
[core-297] Added input option that sets the resolution of the rendering when loading a
.pdffile as a bitmap image. -
[core-297] Added function to load
.heicand.aviffiles (based on the use oflibheif). This is not enabled by default. Thanks to Lukas Rusak for this nice contribution. -
[core-297] Remove the default dependency of the project on the OpenCV library, for the CLI tool
gmic. OpenCV is a monster library that requires a lot of external libraries (add more than 150 library dependencies to the executablegmic, on my Ubuntu), and which is used only by commandcameraand for saving/loading video files (considering there is however a fallback method whenffmpegis installed). Not worth it. -
[core-297] Force autostart command
cli_startnot to take arguments. -
[core-297] Windows: Improve detection of external tools that could be used by G’MIC to improve I/O (e.g.
ffmpeg). -
[core-297] When saving a video file from odd-sized images: frames are automatically extended with one row and/or column to make sure they have an even number of rows/columns, avoiding error when saving the video file.
-
[stdlib-297] Re-implementation of the G’MIC Markdown engine, from scratch (faster, and now have a syntax closer to “usual” markdown). See what syntax this Markdown engine accepts here: G'MIC - GREYC's Magic for Image Computing: A Full-Featured Open-Source Framework for Image Processing - G'MIC Markdown
-
[stdlib-297] Help displayed on terminal now takes into account the number of columns of the terminal, to expand as much as possible.
-
[web-297] Added “latest dev” links in download page, to let users download the binaries corresponding to the latest development version of G’MIC. We hope we’ll get more feedback and bug report from users, to improve software development.
- [web-297] Stop releasing packages for Debian8 and Windows/GIMP2.8 from the G’MIC download web page.
Bug fixes:
-
[core-297] Fix rendering bug when rendering out-of-screen triangles with z-buffer. Had an impact on the rendering of 3d objects (command
object3d). -
[math-core-297] Fixed potential wrong characters returned by function
get()when retrieving a G’MIC variable as a string. -
[stdlib-297] As usual, a lot of small bugfixes in the G’MIC standard library.
From 2.9.5 to 2.9.6
Bug fixes:
-
[gmic-qt-296] Fix bug that prevented users getting filter updates if their account names have accentuated characters in it (wrong handling of UTF-8 string that describe
%APPDATA%).
From 2.9.4 to 2.9.5
New features:
-
[gmic-qt-295] G’MIC-Qt plug-in has now a 8bf interface, thanks to @PDN_GMIC. It can be run now in Photoshop / PaintShop Pro / Photoline / Affinity Photo and theoretically any software that supports 8bf plug-ins.
-
[gmic-qt-295] New button “Copy G’MIC Command to Clipboard” added, to make it easier getting the G’MIC command line corresponding to the applied filter.
- [gmic-qt-295] New filter Artistic / Ghost creates a ghot effect from an input image.
- [gmic-qt-295] New filter Colors / Tune HSV Colors allows to finely remap HSV basis colors acording to the user’s settings. Use complex CLUT reconstruction internally.
- [gmic-qt-295] New filter “Rendering/Hypotrochoid” renders hypotrochoid curves.
-
[gmic-qt-295] New filter "About/What’s New? " shows recent modifications of the plug-in filters (new or removed).
-
[core-295] New command
output_objthat saves a 3D object into a Wavefront.objfile (something that Blender can import easily).
-
[core-295] Input/output of
.raw,.cimgand.cimgzdata files now acceptboolas a possible type. In that case, the image data are stored using a bitwise format. -
[math-core-295] New function
addr(expr)returns the internal pointer address to the specified expressionexpr.
Improvements / Modifications:
-
[core-295] Command interpreter has been slightly optimized:
- Debug info are not included by default anymore when reading a command file;
- Checking the validity of the interpreter state is now done only in
debugmode. - Static analyzer determines if item substitution can be skipped for some items.
- Added more memory slots to store variables, leading to less string comparisons when accessing a variable value, in case many variables are defined.
-
[core-295] Native commands that define default arguments now accepts
,as an argument (just as with custom commands). -
[core-295] Command
split3dhas a different argumentfull_splitwhich allows to enable splitting of 3D object primitives as single images in the resulting list. -
[core-295] File options, if any specified, are now passed to custom commands
input_extandoutput_ext. -
[core-295] New pre-defined variables
$_os,$_flags,$_pixeltypegive additional information on G’MIC interpreter and its running environment. -
[stdlib-295] Commands
replace_strandreplace_seqslightly optimized.
Bug fixes:
-
[stdlib-295] Command
display: Fix “shaking aspect ratio” bug happening when visualizing small images. -
[math-core-295] Fixed user assignment of reserved variable
in. -
[math-core-295] Fixed wrong behavior of expression
v[ind] = value, whenvwas a pre-defined variable (asia,ic,in, …). -
[core-295] Command
svd: check that specified selected images are indeed matrices (no depth or multiple channels).
From 2.9.3 to 2.9.4
Improvements / Modifications:
-
[math-core-294] Function
sort()can be now applied only on a subset of the input vector. -
[core-294] Avoid compilation warning in function
CImgList<T>::font(), withg++>=v.10.
Bug fixes:
-
[core-294] Fix invalid memory read that could happen in function
gmic::add_commands(). -
[core-294] Fix invalid memory read that could happen in command
uncommand. -
[core-294] Fix null pointer indirection that could happen in function
gmic::run().
From 2.9.2 to 2.9.3
New features:
-
[gmic-qt-293] New filter Patterns / Random Pattern renders an image containing a random pattern, based on a random complex-valued function generator. Associated command
random_patternhas been added in the G’MIC stdlib as well.
-
[gmic-qt-293] New filter Degradations / Rebuild From Blocs. It subdivides an image into blocs, and replace each bloc by the most similar bloc found in the set of other blocs.
-
[gmic-qt-293] New filter: Degradations / Blur [Multidirectional].
-
[gmic-qt_293] New translation system has been set up to allow translations of all filter widgets, rather than just the interface widgets.
-
[stdlib-293] New command
display_camera(shortcut:dc) opens a simple interactive webcam viewer. -
[math-core-293] New function
expr('formula',_w,_h,_d,_s)outputs a vector of sizew*h*d*swith values generated from the specified formula, as if one were filling an image with dimensions(w,h,d,s). -
[math-core-293] New function
get('varname',_size)returns the value of the specified variable, as a vector ofsizevalues, or as a scalar (ifsizeis zero, or not specified). -
[math-core-293] New function
arg0()which is basically the same asarg()but with the argument index starting from0rather than1. -
[stdlib-293] New command
parse_cliis a parser of.gmicfiles that extracts the description of the#@clicommands and outputs them in various formats. -
[stdlib-293] New documentation generator implemented. Reference documentation is more flexible and can be now written with a light Markdown-like syntax (
gmd, for G’MIC Markdown). New commands to convertgmdto html or ascii outputs have been set up. -
[web-293] Web pages of the project have been improved. New command reference pages have been set up.
Improvements / Modifications:
-
[gmic-qt-293] Default choice for confirmation dialog appearing when suppressing a Fave filter is now set to
Yes. -
[core-293] External calls to ImageMagick’s
convertare better managed on Windows, when filename or path contain spaces or UTF-8 characters. -
[core-293] Command
text: improvement of character spacing rules. -
[core-293] Command
resize: Argument list as[image],_interpolation,_boundary_conditions,_ax,_ay,_az,_acis not permitted anymore, because it can be ambiguous compared to[image],_height,_depth,_spectrum,_interpolation,_boundary_conditions,_ax,_ay,_az,_ac. New commandresize_as_image(eq. tori) replaces the use of the deprecated argument list. -
[core-293] Revert back on the behavior of commands
discardandsplitwhen invoked with a list of values as the argument (i.e no specified axis). They don’t try to preserve the geometry of the input images anymore (this was creating annoying discontinuities in the behavior of these commands). -
[math-core-293] Function
get()has now an additional option to retrieve the variable value as a string rather than as a value or a set of values. -
[math-core-293] Complex functions
cabs(), cconj(), carg(), ccos(), csin(), ctan(), cexp(), clog(), ccosh(), csinh()andctanh()now accepts scalar-valued arguments (and still returns complex numbers, i.e.vector2). -
[stdlib-293] Command
convolve_fftnow accepts an optionalboundary_conditionargument. -
[compiling-293] Compilation environment on Windows has been rebuilt from scratch. It now uses a recent version of the
g++compiler (and 3rd-party libs), and Link-Time Optimization is now enabled when compiling the G’MIC binaries.
Bugfix:
-
Command
+eand+warn: Avoid first undesired line-feed character put onstdout. -
Command
convolve_ffthas been recoded and is more precise and fast now. -
Fixed small bug in scope management that made some specific sequences of flow commands not working as expected.
From 2.9.1 to 2.9.2
New features:
- [gmic-qt-292] New plug-in filter Artistic / Doodle creates doodles from input photographs.
-
[core-292] New native command
mproj(stands formatrix projection) computes the projection of a matrix into a prescribed dictionary (defined as another matrix), with the choice of various projection methods : ortho-projection, matching pursuit, and orthogonal matching pursuit. Useful stuff for sparse dictionary learning. Math evaluator gets an equivalentmproj()function as well. -
[math-core-292] New pre-defined variables
inandin#indreturn the L2-norm of respectively current image and image#ind, viewed as vectors or matrices (in that case, the Frobenius norm). Corresponding functionin(_#ind)does the same for images with dynamic content (e.g. images under modification). -
[stdlib-292] New commands
xyz2jzazbz,rgb2jzazbz,jzazbz2xyzandjzazbz2rgbconvert RGB or XYZ colors to the Jzazbz colorspace (and vice-versa). Code for these commands has been borrowed from this page by Alan Gibson (@snibgo).
Improvements / Modifications:
-
[core-292] Remove support for deprecated double-hyphen prefix (
--command). Use+commandinstead. -
[core-292] A few internal functions for matrix calculus have been better parallelized, when applied on large matrices (multiplication, norm, dot product, …).
-
[core-292] Command
invertaccepts an optional argument to choose between different solver for matrix inversion (can be0=SVDor1=LU). SVD solver is slower for large matrices, but less numerically instable. Math evaluator functionsinvert()andpseudoinvert()also gets this optional argument. -
[core-292] Commands
fillandevalbetter manage auto- or forced parallelization when image has a few columns. -
[core-292] Commands
solve,invert,svd,trisolveandblur: use double precision for computation. -
[core-292] Command
object3d: Add modification in a 3D object format to define a perfect sphere primitive from a pair(center,point), rather than specifying it only with its diameter(point1,point2). -
[core-292] Command
+echois now implemented as a native command (faster). -
[core-292] Command
echobetter manages messages starting withCRcharacter. -
[core-292] Commands
discardandsplit: When possible, preserve principal axis of processed images. -
[core-292] Command
label: Change output behavior in order to manage multi-channels input images, and output always a single-channel label image even when a multi-spectral input image is given. -
[core-292] Command
label: New argument to choose between L1 and L2-norms for tolerance comparisons. -
[math-core-292] Function
transp(),inv()andpseudoinv()have been renamed respectively astranspose(),invert()andpseudoinvert()to be more coherent with the name of the corresponding pipeline commands. -
[stdlib-292] Remove useless command
unrepeat(eq. todiscardwithout arguments). -
[stdlib-292] Add illuminant
Eas a possible choice for color conversion commands involving the XYZ colorspace. Set it as default illuminant when no explicit illuminant is specified. -
[install-292] Ubuntu/Debian packages provided on the G’MIC website now add an entry to the Desktop menu with the stand-alone version of G’MIC-Qt and ZArt.
Bugfix:
-
[core-292] Math evaluator: Fix possibly wrong pointer in call to
delete. -
[core-292] Command
solve: Fix compilation issue when Lapack support is enabled. -
[stdlib-292] Filter Apply External CLUT: Implement
Strengthcontrol. -
[stdlib-292] Fix command
colormap, for some degenerated cases. -
[gmic-qt_292] Fix possible canvas shift when applying a filter inside a selection.
From 2.9.0 to 2.9.1
New features:
- [gimp-qt-291] New filters Various / Import RGB-565 File and Various / Export RGB-565 File allows to import/export data from raw RGB-565 files.
-
[gimp-qt-291] New API feature allowing to hide unused items or comboboxes for hosts that don’t implement all the plug-in features.
-
[core-291] New command
remove_named(shortcutrmn) removes images with specified names from the list (equivalent torm[${"-nmd name"}]). -
[core-291] Interactive 3D viewer (command
display3d, a.k.ad3d) has been recoded from scratch as a custom command.
-
[core-291] New substituting expression
{*,f}returns0or1whether the display window is in fullscreen mode or not. -
[math-core-291] New function
swap(a,b), swaps the values of given argumentsaandb. -
[math-core-291] New functions
vargkth(),vargmin(),vargmax(),vargminabs(),vargmaxabs(),vavg(),vkth(),vmin(),vmax(),vminabs(),vmaxabs(),vmed(),vprod(),vstd(),vsum()andvvar()are similar to the functions whose names do not start withv, but works with vector-valued arguments (and return a vector-valued result). -
[math_core-291] New function
string()creates a string from a list of arguments (similar toecho()but outputs a vector). -
[stdlib]-291] New command
rbfcomputes interpolated 1D/2D/3D functions from a set of sparse keypoints.
-
[stdlib-291] New command
percentileapplies a “percentile filter” to selected image (mix between averaging and dilate/erode/median).
-
[stdlib-291] New commands
input_565andoutput_565allows to read/write images files in raw RGB565 format. -
[stdlib-291] New command
is_macosreturns1or0whether G’MIC is run on MacOS or not.
Improvements:
-
[core-291] Commands
solveandinvertnow runs in parallel when possible (also apply for functionssolve()andinv()in the math parser). -
[core-291] Special input string
(value1{,|;|/|^}value2{,|;|/|^}...[:{x|y|z|c|,|;|/|^}])has now an optional:axisoption to unroll the image values along a specified axis. -
[core-291] Slightly improved the memory footprint of the G’MIC interpreter. A few memory allocations are avoided when not necessary.
-
[core-291] Variable font size are now used in interactive viewer when displaying a list of image (command
display), in order to make it readable on HDPI screens. -
[core-291] Optimize commands
isoline3dandisosurface3d, by lowering the number of grid points evaluations. -
[core-291] Better support UTF-8 paths and filenames on Windows.
-
[core-291] Command
uncommandnow has a shortcut nameum. -
[math_core-291] Add a pre-optimization step that detects common simple expressions for speeding up evaluation for these cases.
-
[math_core-291] Change encoding of
unsigned intvalues in functionsf2ui()andui2f(), so that no subnormal float numbers are used for the encoding. This avoids issues for commands dealing with large 3D objects, when aggressive optimization flags are used for compilinggmic(which is the case by default). -
[math_core-291] Improve syntax for accessing a vector component, where the left side of the
[]operator is complex. For instance,I[#0,off][2]is now recognized (needed parentheses before, i.e.(I[#0,off])[2]). -
[stdlib-291] Command
display2d: font with adaptive size is now displayed in order to deal with hidpi screens. -
[stdlib-291] Command
display2d: hexadecimal display of color values is now enabled for 1,2,3,4-channel images (not just 3 as before). -
[stdlib-291] Command
sort_listhas been optimized when used with thentype of sorting.
Bugfix:
-
[core-291] Fix a few (critical) OpenMP issues with clang++ compiler.
-
[core-291] Fix possible buffer overflow happening when reading some malformed .bmp files.
-
[core-291] Substituting expressions s.a.
${"command"}now checks that the providedcommanddoes not change the number of images in the list, and throws an error otherwise (was possibly leading to segfaults, e.g.sp lena +b 10 d ${"max u 10"}). -
[core-291] Fix
argmaxcommand andargmax()function in math parser (regressions introduced in 2.9.0). -
[math_core-291] Fix modulo function when called with
infornanvalues. -
[math_core-291] Fix function
isint()when passingNaNorinfvalues. -
[stdlib-291] Fix filter Artistic / Cutout (regression of 2.9.0).
