본문으로 건너뛰기

Font Generation - fontgen

feature idstatusdescription
fontgendraftgenerate custom fonts with embedded emoji/glyphs

Overview​

The fontgen feature generates custom font files (TTF/OTF) with embedded emoji images and custom glyphs, providing an alternative to image placeholders.

Expanded Scope​

While the initial intent of fontgen was to enable runtime dynamic generation of PNG-emoji fonts, its scope is broader and designed as a general-purpose font generation and manipulation system. This includes embedding arbitrary images and glyphs into fonts, supporting a wide range of use cases beyond emoji.

Example Use Cases​

  • User-uploaded emoji sets: Allow users to upload their own emoji collections and generate fonts embedding these images for consistent rendering across platforms.
  • Cross-platform Apple Color Emoji support: Generate fonts that embed Apple Color Emoji images (sbix or CBDT/CBLC tables) to enable consistent emoji rendering on non-Apple platforms.
  • User-defined custom fonts: Enable users to create fonts with custom glyphs or icons, embedding bitmap or vector images directly into font files.
  • Design and export workflows: Integrate with design tools like Glyphs app or similar, allowing font designers to export fonts with embedded images or color glyphs for downstream use.

Goals​

  • Loading and exporting fonts: Support loading existing fonts and exporting new or modified fonts with embedded images.
  • Partial loading and streaming of PNG glyphs: Enable partial font loading and streaming of bitmap glyphs to optimize memory and performance.
  • Embedding images in fonts: Support embedding images using standard tables such as sbix (Apple Color Emoji) and CBDT/CBLC (Google's color bitmap glyphs).
  • Exporting TTF/OTF: Generate standard TTF or OTF font files that include embedded images and custom glyph data, suitable for use in text rendering engines.

Technical Details​

  • Runtime font baking: The system can generate font shards dynamically at runtime, composing embedded images and glyphs into font tables.
  • Supported tables: Full support for color bitmap tables like sbix and CBDT/CBLC to embed PNG or other bitmap formats.
  • Partial glyph updates: Allows updating or streaming individual glyph images without rebuilding the entire font file.
  • Integration with SkParagraph/Skia: Designed to work seamlessly with Skia's text rendering stack, including SkParagraph, enabling proper text layout, cursor positioning, and font fallback semantics.

Pros and Cons Compared to Placeholder Approach​

  • Pros:
    • Provides true font fallback semantics and text engine integration.
    • Enables proper cursor positioning and text selection.
    • Supports complex text shaping and layout with embedded images.
  • Cons:
    • More complex to implement and maintain.
    • Font file sizes can increase due to embedded images.
    • Potentially slower font loading compared to simple placeholders.

Notes for Grida​

This feature is currently exploratory within Grida. It is undecided whether fontgen will become the primary method for emoji and glyph rendering, but it provides a powerful and flexible capability for general-purpose font generation and embedding images within fonts. It complements the existing placeholder-based approaches by offering an alternative path with richer integration and semantic fidelity.

See Also​