UvReport

Remarks

Five of these are MeshTailor's and the rest are what a renderer needs. docs/plan/42 § Part 4. The paper's metric set is the best-specified one in the literature — and BoundaryJaggedness in particular is what separates "low distortion" from "an artist would accept this", which nothing else in the field writes down.

⚠ SeamLengthNormalized divides by the square root of the area, not by the area. The paper defines it over area, which is not dimensionless: halve a model's scale and the figure doubles. Both are reported so the number compares across models and the published one is still recoverable.

⚠ PackingEfficiency and EffectiveEfficiency are both here because the first one flatters. Raw used-area-over-atlas-area rewards a packer that leaves no room to bleed into, and the gap between the two is exactly what a margin setting costs.

Fields and properties (13)

• public int ChartCount

  The headline. Fewer is better at equal distortion.

• public float Compactness

  4πA/P² per island, averaged. Round islands pack and filter better than tendrils.

• public float Convexity

  A/A(hull) per island, averaged — the other half of shape quality.

• public float SeamLength

  Total seam length, in world units.

• public float SeamLengthNormalized

  Seam length over the square root of surface area, so it is scale-free.

• public float BoundaryJaggedness

  Discrete curvature over resampled boundary loops.

• public UvDistortion Distortion

  The four measures.

• public float PackingEfficiency

  Island area over atlas area, before margin.

• public float EffectiveEfficiency

  The same after margin — what the atlas actually costs.

• public UvTexelDensity TexelDensity

  Achieved density across every chart.

• public IReadOnlyList<UvStageTiming> Stages

  Time and element counts, per stage.

• public IReadOnlyList<string> Warnings

  What the caller should know: a chart that had to be split, a rescale, a spill into a second tile.

• public bool IsInjective

  Whether the mapping is invertible everywhere, which is the one field that is pass or fail.

Methods (1)

• public UvReport(int ChartCount, float Compactness, float Convexity, float SeamLength, float SeamLengthNormalized, float BoundaryJaggedness, UvDistortion Distortion, float PackingEfficiency, float EffectiveEfficiency, UvTexelDensity TexelDensity, IReadOnlyList<UvStageTiming> Stages, IReadOnlyList<string> Warnings)

  Everything measured about an unwrap, which is the answer to "is this good?".

Used by (32)

• ChartOutcomeVixen.Geometry.Uv
• CharterVixen.Geometry.Uv
• FlattenOutcomeVixen.Geometry.Uv
• FlattenerVixen.Geometry.Uv
• PackOutcomeVixen.Geometry.Uv
• PackerVixen.Geometry.Uv
• UvAllTestsVixen.Geometry.Uv.Tests
• UvChartDeterminismTestsVixen.Geometry.Uv.Tests
• UvChartInvarianceTestsVixen.Geometry.Uv.Tests
• UvChartQualityTestsVixen.Geometry.Uv.Tests
• UvChartTestsVixen.Geometry.Uv.Tests
• UvDensityTestsVixen.Geometry.Uv.Tests
• UvFacetedGroupTestsVixen.Geometry.Uv.Tests
• UvFlattenBijectivityTestsVixen.Geometry.Uv.Tests
• UvFlattenCotangentTestsVixen.Geometry.Uv.Tests
• UvFlattenDeterminismTestsVixen.Geometry.Uv.Tests
• UvFlattenDistortionTestsVixen.Geometry.Uv.Tests
• UvFlattenInvarianceTestsVixen.Geometry.Uv.Tests
• UvFlattenTestsVixen.Geometry.Uv.Tests
• UvFlattenTopologyTestsVixen.Geometry.Uv.Tests
• UvPackDeterminismTestsVixen.Geometry.Uv.Tests
• UvPackEfficiencyTestsVixen.Geometry.Uv.Tests
• UvPackPropertyTestsVixen.Geometry.Uv.Tests
• UvPackingTestsVixen.Geometry.Uv.Tests
• UvStackingTestsVixen.Geometry.Uv.Tests
• UvUdimTestsVixen.Geometry.Uv.Tests
• UvUnwrapVixen.Geometry.Uv
• UvUnwrapDeterminismTestsVixen.Geometry.Uv.Tests
• UvUnwrapPipelinePropertyTestsVixen.Geometry.Uv.Tests
• BlockoutUvPanelVixen.Editor.Blockout
• GeometryRunnerVixen.Cli
• ModelRetopologyVixen.Editor.Assets