chlumsky / msdf-atlas-gen Goto Github PK
View Code? Open in Web Editor NEWMSDF font atlas generator
License: MIT License
MSDF font atlas generator
License: MIT License
Make is so you can specify padding between glyphs in the atlas
Pull request that adds this feature:
#80
Hello, Chlumsky, I am trying to render such letters as 'g', comma, or apostrophe right. How can I use the info about planeBounds and advance to do it? I've looked at freetype but it has different notations
On an m1 macbook, I cloned the repository recursively, and ran make
The following error is printed
msdfgen/ext/import-font.cpp:6:10: fatal error: 'ft2build.h' file not found
#include <ft2build.h>
^~~~~~~~~~~~
1 error generated.
Using the cmake GUI, it reports to me:
Could NOT find Freetype (missing: FREETYPE_LIBRARY FREETYPE_INCLUDE_DIRS)
I have tried installing freetype via brew install freetype
, and modifying the Makefile to point freetype like so:
-I /opt/homebrew/include/freetype2
This instead produces the error:
ld: library not found for -lfreetype
clang: error: linker command failed with exit code 1 (use -v to see invocation)
I believe macOS should already include the freetype library anyway.
I have successfully built this library on other macOS machines, so I'm not sure if I'm doing something wrong or there is something unique about my setup that is causing this issue.
Hello,
I built the program with sanitizers enabled (-fsanitize=address,undefined). Running the program with charset file containing cyrillic letters
['А', 'Я']
trips ubsan
; ./msdf-atlas-gen -imageout Cantarell-Regular.png -json Cantarell-Regular.json -font Cantarell-Regular.ttf -charset a.txt
Neither atlas size nor glyph size selected, using default...
msdf-atlas-gen/utf8.cpp:22:32: runtime error: left shift of negative value -48
Failed to load character set specification.
which corresponds to
msdf-atlas-gen/msdf-atlas-gen/utf8.cpp
Line 22 in cad221a
Casting to unsigned char before shifting appears to fix the issue.
Hi,
I noticed that msdfgen includes, such as this one are like <msdfgen.h>
and not <msdfgen/msdfgen.h>
.
If I am not mistaken, the msdfgen
cmake does nto provide access to this kind of include (inferred from this line but I am no really good with cmake so I might have missed something). It seems to work here because of how git submodules work and the manual -I msdfgen
in the makefile.
Would you be interested in a PR to change that ? I actually need it to integrate this repo to conan package system, but it can be done externally so this modification is not necessary on my side, but I figured you might want to have it here too.
When generating the atlas with some fonts, I need to set inverseYAxis
to true
, but I cannot find a way to do this without removing const
from GlyphGeometry
's getShape()
function. My intuition tells me, that there is a good reason for const here, and it shouldn't just be removed. Another option would be to const_cast the shape. This also feels weird, but has the advantage that nothing would have to be changed in this repository.
To clarify, here how I do it:
[...]
// Storage for glyph geometry and their coordinates in the atlas
std::vector <msdf_atlas::GlyphGeometry> glyphs;
[...]
for(msdf_atlas::GlyphGeometry & glyph : glyphs){
glyph.getShape().inverseYAxis = true; // only works if we remove const from GlyphGeometry::getShape()
// or we do a const_cast
msdfgen::Shape & shape = const_cast <msdfgen::Shape &>(glyph.getShape());
shape.inverseYAxis = true;
glyph.edgeColoring(&msdfgen::edgeColoringInkTrap,
3.0, // max corner angle
0); // seed
}
[...]
Is there a better way? If not, what about either an additional non-const function for getShape, or an additional inverseYAxis parameter when loading the glyphs in FontGeometry::loadCharset
and GlyphGeometry::load
?
I assume I am not the only one running into this, but if you think const_casting is fine, feel free to simply close this issue. Otherwise I'm happy to send you a PR, or, if there is a better way to do this I would greatly appreciate a pointer in the right direction.
I have created a MSDF
msdf-atlas-gen-master bastianzuhlke$ ./bin/msdf-atlas-gen -pxrange 0 -font OpenSans-Regular-V2.ttf -charset charset.txt -size 36 -format bin -imageout OpenSans-Regular-V2.raw -json OpenSans-Regular-V2.json -pots -type msdf
When I render the word Lorem I get a dark grey line left from the "L". I can see that in the atlas left to the "L" is the "M".
So I suppose that the shader is fetching texels belonging to the M.
"L"
"atlasBounds": {
"left": 73.5,
"bottom": 95.5,
"right": 90.5,
"top": 123.5
}
"M"
"atlasBounds": {
"left": 44.5,
"bottom": 95.5,
"right": 72.5,
"top": 123.5
}
No MipMap is used.
TEXTURE_MAG_FILTER -> Linear
TEXTURE_MIN_FILTER -> Linear
The atlas is 256x256. To calculate the uvs I am diving the left / bottom / right / top with 256.
In general such border lines appear here and then.
If i am using an atlas generated by msdf-bmfont-xml which offers texturePadding I do not have these artefacts. Unfortunately msdf-bmfont-xml has some flaws why I can not use it.
Any idea to avoid this issue ?
If I understand correctly, the current padding parameter is only between glyph and glyph. However, it doesn't seem to be helpful for applying some special effects. The correct way should be to apply padding when generating glyph's sdf, which means that the padding area should also contain sdf data.
The json contains and empty array for kerning.
Command: ./msdf-atlas-gen -font Montserrat-SemiBold.ttf -json Montserrat.json -imageout Montserrat.json.png -type mtsdf
(The double extension on -imageout
is just how the engine handles fonts)
Font used: Montserrat SemiBold (ttf
)
Platform: Fedora Linux x86_64
Thanks for this tool.
Would you be interested in packaging it for conan ? I'm using conan as my package manager for C++, and it would be nice to have access to msdf-atlas-gen there, for easy integration in build systems. Also conan-center provides prebuilt binaries of packages for a lot of different configurations which is handy.
msdfgen is already packaged for example.
I think this might need a small rework of the build script, for example a move to CMake.
Currently this project does not have a license, which means that it cannot be redistributed in any way. Can you please include a license file with the project to make the terms explicit?
I expect that this is a simple omission, since msdfgen itself is already licensed under MIT.
I've noticed that all the fonts I've tested with so far have a value of 0.5 added to each entry in the atlasBounds structure, for example on this 1-character 16px atlas I used to test with:
"atlasBounds": { "left": 0.5, "bottom": 1.5, "right": 14.5, "top": 15.5 }
Is it safe to just truncate these values? In all your example code and others example code, the 0.5 is just ignored, so I am curious as to why it's added in the first place?
EDIT: Ah is it to force the GPU to chose the correct pixel when calculating UVs?
Great tool by the way, and really excellent paper as well.
I'm trying to integrate this awesome tool into the game engine I'm making. I found that in the output JSON file, there is an attribute of "atlas"
listed as "yOrigin": "bottom"
, and it seems like this is hardcoded in the original C++ code. It would be better if the tool could receive the Y origin direction as a command-line parameter since the graphics API I'm working with has the top side of the texture as the Y origin and this gives users enough flexibility.
I've been trying to get this to build for ubuntu 20.04 and am failing.
I've
git clone [email protected]:Chlumsky/msdf-atlas-gen
cd msdf-atlas-gen
git submodule init
git submodule update
sudo vcpkg install
mkdir build
cd build
cmake ..
and hit a bunch of errors. I've tried to chase down any missing needed libraries and sudo apt-get install
'd them, I've installed vcpkg
on my system and set /opt/vcpkg
as VCPKG_ROOT
in my .bashrc
, I've run vcpkg install
in the msdf-atlas-gen
directory: what am I supposed to be doing that I'm not?
My latest error is that it can't find tinyxml2
, even though I see it in the vcpkg_installed
directory, and I've sudo apt-get install libtinyxml2-dev
'd.
Currently, the charset parser is implemented in Charset::load
and can only work with a FILE *
handle. It would be much better if it was possible to parse the data from other sources, such as a string in memory. The function could be reimplemented in the following way to allow for various implementations of reading input and outputting the parsed values.
template <int (*READ)(void *, int, void *), void (*ADD)(unsigned, void *)>
bool parseCharset(void *userData, bool disableCharLiterals = false);
Hello there,
Can we integrate msdf-atlas-gen on the mobile apps and generate msdf type atlas on mobile for drawing text in real-time?
Thank You for Your hard work.
Loving the whole system by the way, been using this for years.
I have been using a CharSet.txt consisting of basically 0x0000 -> 0x10ffff.
I realised though that this misses out glyph index 0, which is ".notdef".
So I thought, ok, I should generate using GlyphSet.txt. However, although this seems to work, in that it outputs the glyphs in the range, I noticed the entries in the JSON now don't have the "unicode" field.
So how can I get it to output all glyphs in a font (surely a common requirement?) but include their unicodes, or say, zero, or some undefined value if they don't have a unicode value (such as .notdef amongst some others)?
The sample code for dynamic atlas doesn't work with 1.2.2 - could you do a newer release with that functionality?
i'm using mtsdf and drawing outlines around my text, but the outlines from adjacent characters is bleeding in very slightly around the borders of some of my characters. if there were a way to set the glyphs to not be packed so tight (ie "leave an additional 2px border around each character") I would expect that could solve it!
I am trying to generate a sdf using a icon font created in iconmoon.
I get the above message and it does not generate the atlas.
The command I used was
msdf-atlas-gen -type sdf -font .\icons.ttf -json icons_font.json -format png -imageout icons_font.png -charset charset.txt -dimensions 256 256
and the charset.txt uses the text you would use in the css to display the icon.
"\e900","\e901","\e902","\e903"
Any assistance in helping me generate the sdf from this icon font would be greatly appreciated.
Hey, I'm a massive fan of your work.
I am building a web map engine and I originally used a loop + blinn approach where I prep the vertices with webworkers and create SDFs via the GPU, but the complexity was high, I wasn't supporting all glyph types well, and the cost on the webworkers was too much considering I needed to prep lots of other data for the GPU.
Long story short I switched to MSDF, created a NodeJS wrapper of your work and manage the glyphs in my own way to create some stunning results:
Comparatively, a high enough resolution MSDF has really sharp edges and asian glyphs look absolutely gorgeous. I like to use roughly 22px size per glyph with 6px range for the best results (I like large ranges for nice large stroke effects around the glyph if necessary).
Here's what 16px with 4px range looks like (all Roboto medium+regular fonts)
I noticed the lower I go, obvious artifacts start showing up.
Just curious if there is anything I can do to get the glyphs smaller without sacrificing the quality of the glyphs? If you zoom in on this last picture you will notice it starts getting really confused on where the edges really exist.
Since the browser fetches chunks of glyphs at a time in this scenario, obviously the smaller the better.
Any advice would be appreciated.
EDIT - I just noticed you did some work in May, do you think this could have a substantial effect on the results?
Hi!
Sorry for bad English skill.
I want to pack glyph atlas into multiple textures as texture 2D array.
So I try to add a new member "int textureIndex" into class "GlyphGeometry", and I also try to add the parameter "textureIndex" to "TightAtlasPacker::pack()".
I desire it finally output the attribute "textureIndex" of glyphs into JSON file and load it to render text using texture 2D array.
But "config.emSize" and "config.pxRange" seem to have depended on value of "atlasPacker.getScale()" and "atlasPacker.getPixelRange()" in "main.cpp".
I don't know what should I do if it has multiple value of "atlasPacker"(that means we have multiple value of "atlasPacker.getScale()" ) after I added a loop to generate multiple bitmaps.
And I am not sure I am on the right track or not.
Does anyone can provide me some advice?
Thanks.
Hi,
When exporting to json all the characters are included even if charset file supplied. I don't think this should be the case.
Thanks for this tool, its great!
James
A few things came up for me when trying to use this package and I was hoping someone could lend a hand. I'm using the JSON output and am confused about the planeBounds
and atlasBounds
fields. I think atlas bounds are the pixel space coordinates in the texture for a given glyph and I can calculate uvs using atlasBounds divided by atlas dimensions. Is this correct?
Secondly I have no idea what planeBounds
is for, I was looking for a description in the docs but didn't see anything. Could you give me a quick explaination?
Lastly, I'm trying to figure out the relationship between emRange
, pixelRange
, and size in ems. What I'd like to do is generate a font at a given point size but after tinkering a while I didn't figure out how to do this, could you please point me in the right direction?
Thanks! By the way I'm really impressed with the atlas generation performance, Unity's TextMeshPro takes up to 30 seconds to build an atlas but msdf-atlas-gen is around 100ms for the same font.
first of all, i want to thank you for making this great technique for free! love it so far!
thanks man!
anyway, im using json as an output :
{ "atlas": { "type": "msdf", "distanceRange": 2, "size": 48, "width": 64, "height": 64, "yOrigin": "bottom" }, "name": "raleway_regular", "metrics": { "emSize": 1, "lineHeight": 1.1739999999999999, "ascender": 0.94000000000000006, "descender": -0.23400000000000001, "underlineY": -0.10000000000000001, "underlineThickness": 0.050000000000000003 }, "glyphs": [ { "unicode": 65, "advance": 0.67900000000000005, "planeBounds": { "left": -0.013666666666666617, "bottom": -0.030416666666666609, "right": 0.69466666666666677, "top": 0.74041666666666672 }, "atlasBounds": { "left": 0.5, "bottom": 26.5, "right": 34.5, "top": 63.5 } }, { "unicode": 66, "advance": 0.66800000000000004, "planeBounds": { "left": 0.064833333333333368, "bottom": -0.030416666666666609, "right": 0.64816666666666678, "top": 0.74041666666666672 }, "atlasBounds": { "left": 35.5, "bottom": 26.5, "right": 63.5, "top": 63.5 } } ], "kerning": [ { "unicode1": 66, "unicode2": 65, "advance": -0.014 } ] }
i couldnt find the width and height information for each of Glyphs?
also i tried to parse the information using freetype in my project.
im using 48 as pixel size. i got this results
why does it produce difference results? what am i doing wrong?
im using this command line
msdf-atlas-gen.exe -font raleway_regular.ttf -fontname raleway_regular -type msdf -format png -size 48 -pxrange 2 -charset charset.txt -imageout test.png -json font.json -csv font.csv
Hi,
I have exported the font data(CSV format) and texture into external files and using the code below to build vertex buffers but was not able to calculate offset_x and offset_y. Here is the code, please let me know if anyone finds any other mistake in the code.
typedef struct Glyph
{
unsigned int code;
float advance;
float pleft; // plane left
float pright; // plane right
float ptop; // plane top
float pbottom; //plane left
float aleft; // atlasleft
float aright; // atlasright
float atop; //atlastop
float abottom; //atlasbottom
float kerning; // Not supported right now
} Glyph;
// font->gdata is std::vector<Glyph> where are the glyph info is stored
void add_text_msdf(FTGL *font, VertexBuffer *vb, const char* text, float x, float y)
{
// Number of glyphs in text
size_t num_glyph = 0;
// Position to draw
int xpos = x;
int ypos = y;
while (*text) {
if (*text >= 32 && *text < 128) {
// Calculate kerning if supported
/*
float kerning = 0.0f;
if( num_glyph > 0) {
kerning = font->gdata[*text-32-1].kerning - font->gdata[*text-32].kerning;
}
xpos += kerning;
*/
// Width and Height of Glyph;
float width = font->gdata[*text-32].aright - font->gdata[*text-32].aleft;
float height = font->gdata[*text-32].abottom - font->gdata[*text-32].atop;
// Quad Coordinates
float x0 = xpos + ?; // How to calculate offset_x
float y0 = ypos + ?; // How to calculate offset_y
float x1 = x0 + width ;
float y1 = y0 - height;
// UV Coordinates
float s0 = font->gdata[*text-32].aleft / 512.0; // atlas width
float t0 = font->gdata[*text-32].atop / 512.0; // atlas height
float s1 = width / 512.0; // atlas width
float t1 = height / 512.0; // atlas height
GLuint ioff = num_glyph * 4;
GLuint indices[] = {0+ioff, 1+ioff, 3+ioff, // first triangle
1+ioff, 2+ioff, 3+ioff}; // Second triangle
GLfloat vertices[] = {
x0, y0, 0, s0, t0, // top left
x0, y1, 0, s0, t1, // top right
x1, y1, 0, s1, t1, // bottom right
x1, y0, 0, s1, t0 }; // bottom left
vb->push_back(vertices, 20, indices, 6);
xpos += font->gdata[*text-32].advance * 20.0; // Multiply by font size, am I correct?
num_glyph++;
}
++text;
}
}
//
Hello,
I would be interested in having multiple fonts in a single atlas texture. I can probably contribute towards that, but before starting, I would like to know if there are already plans or even maybe work in that direction?
Thank you, kind regards.
Hello! I'm interested in using this library for a project, and so I decided I'd clone and test it out to get a feel for its output etc., but I ran into an error; I ran the following commands (approximately):
$ git clone --recurse-submodules -j8 https://github.com/Chlumsky/msdf-atlas-gen/
$ cd msdf-atlas-gen
$ make
Here's the output:
❯ make
mkdir -p bin
g++ -I /usr/local/include/freetype2 -I /usr/include/freetype2 -I artery-font-format -I msdfgen/include -I msdfgen -D MSDFGEN_USE_CPP11 -D MSDF_ATLAS_STANDALONE -std=c++11 -pthread -O2 -o bin/msdf-atlas-gen msdfgen/core/*.cpp msdfgen/lib/*.cpp msdfgen/ext/*.cpp msdf-atlas-gen/*.cpp -lfreetype
msdf-atlas-gen/TightAtlasPacker.cpp:72:54: warning: multiple unsequenced modifications to 'maxScale' [-Wunsequenced]
while (maxScale < 1e+32 && TRY_PACK(maxScale = 2*minScale))
^
msdf-atlas-gen/TightAtlasPacker.cpp:69:116: note: expanded from macro 'TRY_PACK'
#define TRY_PACK(scale) (lastResult = !tryPack(glyphs, count, DimensionsConstraint(), width, height, padding, (scale), unitRange+pxRange/(scale), miterLimit))
^ ~~~~~
msdf-atlas-gen/TightAtlasPacker.cpp:75:55: warning: multiple unsequenced modifications to 'minScale' [-Wunsequenced]
while (minScale > 1e-32 && !TRY_PACK(minScale = .5*maxScale))
^
msdf-atlas-gen/TightAtlasPacker.cpp:69:116: note: expanded from macro 'TRY_PACK'
#define TRY_PACK(scale) (lastResult = !tryPack(glyphs, count, DimensionsConstraint(), width, height, padding, (scale), unitRange+pxRange/(scale), miterLimit))
^ ~~~~~
2 warnings generated.
In file included from msdf-atlas-gen/main.cpp:20:
In file included from msdf-atlas-gen/msdf-atlas-gen.h:29:
In file included from msdf-atlas-gen/BitmapAtlasStorage.h:33:
msdf-atlas-gen/BitmapAtlasStorage.hpp:51:37: error: non-member function cannot have '&&' qualifier
return (msdfgen::Bitmap<T, N>() &&) bitmap;
^~
1 error generated.
make: *** [all] Error 1
Any idea what's causing that error? Here's the output of g++ --version
in case it's helpful:
❯ g++ --version
Configured with: --prefix=/Applications/Xcode.app/Contents/Developer/usr --with-gxx-include-dir=/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk/usr/include/c++/4.2.1
Apple clang version 12.0.5 (clang-1205.0.22.9)
Target: arm64-apple-darwin20.3.0
Thread model: posix
InstalledDir: /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin
As you can see I'm using Apple's g++
binary on an Apple M1, but I don't know if that's the issue or if it's something else?
Thanks!
When I run this command:
$ .\msdf-atlas-gen.exe -font Dangrek-Regular.ttf -size 128 -type sdf -imageout dangrek.png -json dangrek.json
Loaded geometry of 95 out of 95 glyphs.
Atlas dimensions: 668 x 668
Atlas image file saved.
Glyph layout and metadata written into JSON file.
Looks to sharp to me compared to the images you see in the documentation:
When I change the output type to msdf, this is what I get:
So not sure if I'm doing something wrong, or that's just the way it's supposed to be.
I'm trying to link this to my project, and I seem to only have one compiling error at the moment, and it's '&' requires I-value
at ImmediateAtlasGenerator.hpp at line 39 on the most recent source code download.
Is this simply a linking error? Just as an FYI, I use C++ 20 on the msvc build compiler.
Why is the csv ouput lacking information and not documented at all?
Subj, usually texture packer apps accept padding param between symbols so shader is easier to code, simple clamp()
is not enough if I draw want to draw the letter with spacing, for example when i want to scale it
Thanks for awesome lib btw!
I like CSV for keeping the glyph data as it's much more space-efficient than json, however this means the font data (atlas and metrics fields) gets discarded. It'd be nice to be able to output both a CSV with the glyph data, and a JSON file with the metrics. Or maybe two csv files?
Would appreciate support for SVG file shapes. This is supported in msdfgen, so why not here too?
Is there anywhere I can go to help make sense of the output layout data?
for example, I have this output regarding the character 'A':
{"unicode":65,"advance":0.63300000000000001,
"planeBounds":{"left":0.0023802681992336704,"bottom":-0.032253256704980887,"right":0.63011973180076619,"top":0.77125325670498079},
"atlasBounds":{"left":0.5,"bottom":43.5,"right":25.5,"top":75.5}
},
unicode
is self explanatory,
advance
is, I assume, the proportional-to-font-size distance to advance the cursor for the next character
planeBounds
I assume are the bounds of the glyph on a [-1,1] plane (for both X and Y)? However, if that were the case, looking at my output image I would expect the bounds for 'A' to be something along the lines of "left":-1,"bottom":-0.7,"right":0.1,"top":-0.5
, which it isn't even close to...
atlasBounds
I assume would have been similar to planeBounds
, but in pixel coordinates?
Any help and I'd happily contribute a blurb to your README!
Thanks for your work!
Hi! I am using this amazing program to generate font atlases for my game.
However, I've been struggling to get the text properly aligned.
The system expects to get 2 coordinates, the top left and bottom right positions, with the top left being the closest one to [0, 0] and the bottom right one being the one closest to [screenWidth, screenHeight]. In other words, the screen uses a Y-down coordinate system.
Do you have any tips and/or tricks on how to get it working?
So far I've mangaged to cook up the following with the hope of at least getting something working, however it doesn't feel like the base line is straight, and certain characters are heavily displaced(e.g comma)
FontCharacter meta = Meta.Characters[glyph];
Vector2 uvStart = new Vector2(meta.AtlasBounds.Left, meta.AtlasBounds.Top) / new Vector2(Meta.AtlasMetrics.Width, Meta.AtlasMetrics.Height);
Vector2 uvEnd = new Vector2(meta.AtlasBounds.Right, meta.AtlasBounds.Bottom) / new Vector2(Meta.AtlasMetrics.Width, Meta.AtlasMetrics.Height);
Vector2 screenStart = new Vector2(meta.PlaneBounds.Left, Meta.Metrics.Ascender + meta.PlaneBounds.Bottom) * size;
Vector2 screenEnd = new Vector2(meta.PlaneBounds.Right, Meta.Metrics.Ascender + meta.PlaneBounds.Top) * size;
drawCallback(uvStart, uvEnd, screenStart, screenEnd, offset);
offset.X += meta.Advance * size;
Hello there,
I would like to suggest a new feature to atlas generation.
It seems to be pretty common to pack a font across several textures whenever the glyphs can't fit into the desired texture size.
This is especially useful in contexts like mobile gaming and huge character sets where glyphs may overflow a single 2048x2048 atlas.
Cheers! Thanks for the relevant work in distance fields generation!
I'd like to add multiple fonts into a single atlas and export the metrics to json. I'm then going to write a tool to convert this to C code. I need the font name in there to be able to identify the font.
Thanks :)
Same font, same command line, but now the font draws bigger in my application and also seems misaligned on the vertical axis.
Top image is how it was, bottom image is how it is.
Command line is:
msdf-atlas-gen.exe -font FontAwesome.otf -fontname FontAwesome -and -font Aileron-Regular.otf -fontname Aileron-Regular -and -font Aileron-SemiBold.otf -fontname Aileron-SemiBold -charset CharSet.txt -yorigin top -minsize 64 -pxrange 16 -nokerning -imageout FontAtlas.png -json FontAtlas.json > FontAtlas.log 2>&1
Is it expected to be this different between versions?
Hi,
Export fnt (txt) or json with bmfont compatible format.
Is it doable?
Thanks in advance.
success
is never set to true in saveImageText
, so you always receive an error
The program crashes immediately after I run it (I tried all versions).
There is no error displayed.
Hi,
I'm using your wonderful tool to create my fonts, but since the margin between the glyphs is only 1, some glyphs at some font sizes are getting artifacts from the other glyphs. Can you please add a parameter to define an extra margin around the glyphs?
Kind regards
guich
Hi again, I'm sure this isn't directly related because of your library, but I noticed it only happens when trying to use it in Release build. So I compiled the game from source and included it in my project. I can run the example on Debug and it compiles completely fine and runs exactly how I expected, which is great. My issue comes in with compiling to Release (yes I am using the Release Library compiled files from both msdf-atlas-gen and msdfgen). I get this weird error:
I have two projects that you can see in this screenshot, I have cosmic-engine-engine which is the backend engine which does use msdf-atlas-gen, and I have cosmic-engine-editor, which includes both the libraries and the engine. I can compile and run Debug but I can't with Release (because of this error). Now I know it's something having to do with the Release build of the library as if I were to comment out all the code relating to msdf-atlas-gen/msdfgen, both can compile and run, but with the example code, only Debug can run, Release has compile errors. I didn't change anything about the solution for msdf-atlas-gen except I believe I did change the runtime library used so that my engine could even use the libraries. I'll keep looking into it, maybe I have some mismatch going on, and I did remove the one preprocessor definition that you explained in my last issue.
I'll hopefully be able to solve this before you even reply, maybe I just accidentally changed something, but so far everything seems to be unchanged from the source code except for the runtime library, which again, needed to be changed for the sake of being able to link it.
And no, I only have vs2022 installed with it's own compiler, and I opened up your code via the same vs2022 version, as well as set the toolset and c++ version to the same of my engine/editor.
Edit*: Doesn't happen on the engine project version on Release build, however the engine project is just a static library and I've been told some linking errors may not occur in this type of build process.
Hey, I've been using MSDF for a while now and really like it, thanks for sharing your great projects with the world!
I'm working on a UI package for Unity and since I'm unhappy with Unity's fonts I wanted to ship with MSDF instead. Are there any restrictions that prevent me from distributing my package that bundles msdf and the generator within it to customers?
The atlas generator loads a subset of glyphs from a TTF or OTF font file
that's nice, but I'm not picky and just want access to all glyphs. is there a way to do this?
already tried charset.txt with [0, 1114111]
in it but the resulting image looked completely broken
In hardmask
and softmask
atlas -type
s, there are no special considerations to ensure that the glyph quads can be placed exactly onto the pixel grid. Instead, the plane bounds dictate that each glyph can be shifted by a different fractional pixel amount. This should be resolved.
Hi Viktor, awesome work on these libs and the idea behind them!
I've been working with it over the last couple days and can't seem to get any kerning results, the list in e.g. a converted json output is always empty. I dug in and it seems the FT_Get_Kerning()
call in msdfgen always yields zero kerning. I tried updating FT to their latest version as well, to no success. Any clues why this is the case? Wasn't able to root-cause it.
Cheers and all the best!
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.