Skip to main content

Text rendering

Overview​

The text stack is optional and sits beside the point renderer:

  1. GlyphAtlas — CPU-rasterized bitmap atlas; simple to reason about; rebuilds on DPR changes.
  2. MsdfGlyphAtlas — precomputed MSDF atlas + JSON metrics; better minification; fixed atlas cost.
  3. TextLayer — owns labels, map matrix, cull rect, declutter options; feeds TextRenderer.
  4. TextRenderer — GPU pipeline for glyph instances and frame uniforms.

Import paths​

import {
  TextLayer,
  createTextLayerForPointsRenderer,
  GlyphAtlas,
  tryLoadMsdfAtlasFromUrls,
} from "stratum-map";

Workflow​

  1. Build TextLabel[] (anchor in map CRS, strings, priorities, optional halos).
  2. textLayer.setLabels(labels) and update setMapToClipMatrix each frame.
  3. textLayer.rebuildInstancesIfDirty(viewState) then encode the text pass after points (or interleaved per your compositing rules).

See Text layer API and Text renderer API.

Runnable: GPU text labels demo (OpenLayers overlay + TextLayer).

MSDF assets​

Ship atlas.png + atlas.json (versioned schema). Use tryLoadMsdfAtlasFromUrls or construct MsdfGlyphAtlas from existing GPU textures in advanced integrations.