gltfpack: Warn when position quantization error is significant for any mesh

When float quantization is not used, we use the same quantization
settings for all meshes. This simplifies the flow as skins and materials
may need to change based on the quantization grid, and in some cases
eliminates micro gaps between meshes, but may result in reduced
precision compared to using per-mesh grids.

For now we detect this case and suggest increasing the number of bits
used for positions or switching to floating-point quantization.

Author: zeux

gltf/stream.cpp

@@ -25,6 +25,11 @@ struct Bounds
 		return min.f[0] <= max.f[0] && min.f[1] <= max.f[1] && min.f[2] <= max.f[2] && min.f[3] <= max.f[3];
 	}
 
+	float getExtent() const
+	{
+		return std::max(max.f[0] - min.f[0], std::max(max.f[1] - min.f[1], max.f[2] - min.f[2]));
+	}
+
 	void merge(const Bounds& other)
 	{
 		for (int k = 0; k < 4; ++k)
@@ -108,7 +113,20 @@ QuantizationPosition prepareQuantizationPosition(const std::vector<Mesh>& meshes
 		result.offset[0] = b.min.f[0];
 		result.offset[1] = b.min.f[1];
 		result.offset[2] = b.min.f[2];
-		result.scale = std::max(b.max.f[0] - b.min.f[0], std::max(b.max.f[1] - b.min.f[1], b.max.f[2] - b.min.f[2]));
+		result.scale = b.getExtent();
+	}
+
+	if (b.isValid() && settings.quantize && !settings.pos_float)
+	{
+		float error = result.scale * 0.5f / (1 << (result.bits - 1));
+		float max_rel_error = 0;
+
+		for (size_t i = 0; i < meshes.size(); ++i)
+			if (bounds[i].isValid())
+				max_rel_error = std::max(max_rel_error, error / bounds[i].getExtent());
+
+		if (max_rel_error > 5e-2f)
+			fprintf(stderr, "Warning: position data has significant error (%.0f%%); consider using floating-point quantization (-vpf) or more bits (-vp N)\n", max_rel_error * 100);
 	}
 
 	result.node_scale = result.scale / float((1 << result.bits) - 1) * (result.normalized ? 65535.f : 1.f);