[barrulus]

Description

Ocean bands and waves repeatedly rasterize complex SVG masks during map movement, causing substantial panning lag in presets such as Atlas. Cache these layers as viewport-sized images so nearby pans can reuse the rendered artwork.

The cache refreshes for zoom, viewport coverage, coastline edits and rendering-quality changes. Original vectors remain available as a fallback and are restored in save/export clones, so .map files and SVG exports retain their original paths and masks. No dependencies or save-format changes.

Refs #1966. This addresses ocean embellishments; Frostbite's separate settlement-icon filter cost remains.

Performance

Average animation-frame interval during the same 30-step pan (lower is better):

Preset	Original vectors	Cached
Ink	66.1 ms	22.2 ms
Ancient	105.4 ms	25.4 ms
Cinderwood	160.0 ms	30.5 ms
Atlas	113.8 ms	29.6 ms
Frostbite	163.5 ms	123.0 ms
Gloom	53.0 ms	27.9 ms

Measured with Chromium 155 / SwiftShader, seed 1966, 1600×900 viewport, political layers, Balance quality, DPR 1 and 3× zoom. Camera transforms were asserted before and after each drag. These software-rendering measurements show the relative improvement, not expected timings on a particular GPU.

Validation

• Build, TypeScript and repository-wide Biome checks passed.
• 59 focused tests passed across ocean caching, ocean rendering, coastal waves and exports.
• Screenshots matched pixel-for-pixel within the map comparison area for all six presets above.
• Browser checks covered DPR 2 and zoom changes, coastline invalidation, disabling effects, .map save/reload, SVG export and the PNG export pipeline.
• Unit tests cover cache reuse, vector restoration, stale async completion, failure fallback and blob URL release.
• Manually reviewed in the local test server.

Raster allocations are capped at approximately 16 megapixels per ocean layer. Very large or high-DPI views may use reduced on-screen resolution; an existing image can briefly scale while its replacement loads. Native vector saves and exports are unaffected.

[netlify]

✅ Deploy Preview for afmg ready!

Name	Link
🔨 Latest commit	59cba15
🔍 Latest deploy log	https://app.netlify.com/projects/afmg/deploys/6ad1e289018552527bc2a78e
😎 Deploy Preview	https://deploy-preview-1967--afmg.netlify.app
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.

---

To edit notification comments on pull requests, go to your Netlify project configuration.

[Azgaar]

@barrulus, the solution is working fine, but I'm not sure I like how isolated it is. We may want to raster-render any layer (like we tried with reliefs), not just ocean bands. And this should be optional under performance settings.

So I would prefer a generic 'rasterization' select in performance settings. If it's yes, then we do this caching (rasterization can be more clear name) for what we support. The rasterization module can be much more universal and work on pretty much any layer, not just on ocean bands.

[Azgaar]

This would also close #1352 if we add Relief. But we need to think how to make the solution clean.

@barrulus, let's me know if you would rather merge as it's now and leave a major rework for later.

[barrulus]

This would also close #1352 if we add Relief. But we need to think how to make the solution clean.

@barrulus, let's me know if you would rather merge as it's now and leave a major rework for later.

I think merge as is now, using the same mechanism to overall improve all relief work is a great idea, but it is a larger piece of work, this is a short win that will benefit people today

[Azgaar]

Ok, let's get back to it later. We have a Draft open for Raster layers.