diff options
Diffstat (limited to 'README')
-rw-r--r-- | README | 69 |
1 files changed, 35 insertions, 34 deletions
@@ -2,34 +2,35 @@ NAME libglitter - Subpixel-rendering library DESCRIPTION - libglitter is a C library for subpixel-rendered text from - an greyscale-antialiased text image. libglitter is designed - to be used in conjunction with other font libraries: it can - neither rasterise glyphs nor does it know about the montor's - subpixel layout or rendering configurations. + libglitter is a C library for subpixel-rendered text from an + greyscale-antialiased text image. libglitter is designed to be + used in conjunction with other font libraries: it can neither + rasterise glyphs nor does it know about the montor's subpixel + layout or rendering configurations. - To use libglitter you first need to output's pixel density, - then you apply hinting to the text you want to render using - this information. The next step is to get the output's + To use libglitter you first apply hinting to the text so that + the glyph outlines aligns with the output's pixel-grid as + closely as possible. The next step is to get the output's subpixel arrangement and scaling factor, then assuming that - the output's subpixel arrangement i subpixel-rendering - compatible and that the native resolution is used, you - multiply the output's DPI:s so that you get the number of - subpixels per inch instead of number of pixels per inch - (some subpixels may have be counted multiple times depending - on the subpixel arrangement) and you create a greyscale - raster of the text for these new DPIs. After this you - create an uninitialised colour raster for text and the - output's pixel density, and split it into one raster per - colour channel using libglitter_split_uint64_raster(3) - or libglitter_split_uint32_raster(3), alternatively you - create one raster for each colour channel directly. Then you - use libglitter_compose_double(3), libglitter_compose_float(3), + the output's subpixel arrangement is subpixel-rendering + compatible and that its native resolution is used, you + rasterise the text using greyscale-antialiasing into a raster + sized according to the output's horizontal and vertical + subpixel densities (rather than pixel densities as normally + done with greyscale-antialiasing; some subpixels may have be + counted multiple times depending on the subpixel arrangement). + After this you create an uninitialised colour raster for text + and the output's pixel density, and split it into one raster + per colour channel using libglitter_split_uint64_raster(3) or + libglitter_split_uint32_raster(3), this is when libglitter is + first used in this process; alternatively you create one + raster for each colour channel directly. Then you use + libglitter_compose_double(3), libglitter_compose_float(3), libglitter_compose_uint64(3), libglitter_compose_uint32(3), libglitter_compose_uint16(3), or libglitter_compose_uint8(3) - to create the subpixel-antialiased image of the text; you - may have to first call libglitter_reorder_rasters(3) to - put the rasters in the expected order. + to create the subpixel-antialiased image of the text; you may + have to first call libglitter_reorder_rasters(3) to put the + rasters in the expected order. An optional next step is to use lessen the intensity of the subpixel-antialiasing with libglitter_desaturate_double(3), @@ -43,17 +44,17 @@ DESCRIPTION and libglitter_colour_space_convert_rasters_double(3) or libglitter_get_colour_space_conversion_matrix_float(3) and libglitter_colour_space_convert_rasters_float(3) to - convert a colour space the application can output in. - It is however out of the scope of libglitter to get the - output's colour space and the conversion matrix to any - other colour space than sRGB or CIE XYZ. + convert a colour space the application can output in. It is + however out of the scope of libglitter to get the output's + colour space and the conversion matrix to any other colour + space than sRGB or CIE XYZ. - The finally step of the rendering process is out of scope - for libglitter, but is to ensure that all floating-point - values (if floating-point rasters are used) are withing - [0, 1] and convert the rasters, which only contain - ink-on intensities, into the desired colour's with the - output's transfer function applied. + The finally step of the rendering process is out of scope for + libglitter, but is to ensure that all floating-point values + (if floating-point rasters are used) are withing [0, 1] and + convert the rasters, which only contain ink-on intensities, + into the desired colour's with the output's transfer function + applied. The application may also desire to call libglitter_enable_- acceleration(3) at the beginning of its execution to enable |