glText
glText copied to clipboard
vallentin
Custom fonts?
Assuming the user would handle loading image data. Then it would definitely be possible to implement custom fonts in the form of bitmap fonts. The default (only) font, is actually a bitmap font baked into the header.
Assuming the user would handle loading image data. Then it would definitely be possible to implement custom fonts in the form of bitmap fonts. The default (only) font, is actually a bitmap font baked into the header.
instead could you point me to where the font is baked in to the header?
The baked in font data, is packed into _gltFontGlyphRects which stores the rectangle information of each glyph, and _gltFontGlyphData which stores the pixels of each font in relation to the rectangle bounds. By packed I mean it quite literally, as bit packing is used minimize the amount of data. For instance, the pixels are only 2-bit, i.e. 00 is transparent, 01 is white, 10 is black.
You can see all the logic for unpacking the data, along with creating the font in the subsequent _gltCreateText2DFontTexture() function. However, if we were to introduce a new function, e.g. gltCreateFont(image_data, glyph_rects), then the logic needed to implement that function is essentially already found in _gltCreateText2DFontTexture() if you ignore all the unpacking.
If you're able to and want to give a go at implementing it, then I'd happily mentor.
how would i go about creating one of these bitmap fonts?
Essentially, you'd be able to just draw a font in any image editing application. However, you wouldn't be able to load it (easily), until that functionality is implemented of course.
Then when you have the image data, you'd essentially need to define a list of glyphs (_GLTglyph). Which include the char it represents, the bounds on the image, the bounds translated into UVs and a boolean drawable representing if it's a drawable glyph, i.e. a TAB is a valid glyph, but in the baked font it draws nothing of course.
https://github.com/vallentin/glText/blob/8200fa70e32acec0a3cd777d404f41ee0c203ca4/gltext.h#L173-L183
The baked in font data, is packed into
_gltFontGlyphRectswhich stores the rectangle information of each glyph, and_gltFontGlyphDatawhich stores the pixels of each font in relation to the rectangle bounds. By packed I mean it quite literally, as bit packing is used minimize the amount of data. For instance, the pixels are only 2-bit, i.e.00is transparent,01is white,10is black.You can see all the logic for unpacking the data, along with creating the font in the subsequent
_gltCreateText2DFontTexture()function. However, if we were to introduce a new function, e.g.gltCreateFont(image_data, glyph_rects), then the logic needed to implement that function is essentially already found in_gltCreateText2DFontTexture()if you ignore all the unpacking.If you're able to and want to give a go at implementing it, then I'd happily mentor.
how would i go about creating one of these bitmap fonts?
Essentially, you'd be able to just draw a font in any image editing application. However, you wouldn't be able to load it (easily), until that functionality is implemented of course.
Then when you have the image data, you'd essentially need to define a list of glyphs (
_GLTglyph). Which include the char it represents, the bounds on the image, the bounds translated into UVs and a booleandrawablerepresenting if it's a drawable glyph, i.e. a TAB is a valid glyph, but in the baked font it draws nothing of course.https://github.com/vallentin/glText/blob/8200fa70e32acec0a3cd777d404f41ee0c203ca4/gltext.h#L173-L183
as one more question in the struct _GLTglyph does x and y represent pixels offset or some other form of offset? same with w and h
as one more question in the struct _GLTglyph does x and y represent pixels offset or some other form of offset? same with w and h
Yes, the x and y is the pixel position in the image data, and the w(idth) and h(eight) is the size in pixels.
So knowing the glyph rect and the image size, then calculating UVs, is of course as follows:
u1 = x / imageWidth
v1 = y / imageHeight
u2 = u1 + (w / imageWidth)
v2 = v1 + (h / imageHeight)
the only hard part seems to be loading the image data isn't it
If we take _gltCreateText2DFontTexture() as a reference again. Then the new e.g. gltCreateFont() function would start at line 1161, as that and forward is essentially the logic it needs. As gltCreateFont() would need to receive image_data and glyphs and thus all the previous logic can be extracted immediately from those.
how could I say convert a TTF to a bitmap one?
There's many tools for that, if you just search "TTF to bitmap font". I've previously heard about Bitmap Font Generator, but I don't recall if I've used it.
Hmm, doesn't seem so, or at least not in a straight forward manner. I just looked here on GitHub, and I found fontbm, which mentions Linux. But once again, I haven't used it, I literally just found it.
Would definitely be cool to be able to specify multiple baked fonts.
Something like
gltUseFont(text, blob, blob_size, blob_glyph_count)
