Differences

This shows you the differences between two versions of the page.

--- 0.8.3:models [2022/04/25 11:14] – [Supported file formats] admin
+++ 0.8.3:models [2022/04/25 11:26] (current) – [Instancing by TModelDef] admin
@@ Line 316: / Line 316: @@
 ===== Instancing / Model duplication =====
+==== Instanced Rendering ====
+Since 0.8.3.2265 instanced rendering of the same mesh for x-times was implemented.
+Use this method to render for example grass or trees. It is not recommended to use it for animated charaters,
+because they would be rendered the same (with the same animation frame).
+You can simply setup instanced rendering in your TGorillaModel or TGorillaMesh object.
+<file pascal>
+/// we create 4 instances in a row
+GorillaModel1.AddInstance(TMatrix3D.CreateTranslation(Point3D(-10, 0, 0)));
+GorillaModel1.AddInstance(TMatrix3D.CreateTranslation(Point3D(-5, 0, 0)));
+GorillaModel1.AddInstance(TMatrix3D.CreateTranslation(Point3D(5, 0, 0)));
+GorillaModel1.AddInstance(TMatrix3D.CreateTranslation(Point3D(10, 0, 0)));
+/// afterwards we can modify a specific instance
+var LTransf : TMatrix3D;
+LTransf := TMatrix3D.CreateScaling(Point3D(2, 2, 2));
+LTransf := LTransf * TMatrix3D.CreateRotationY(DegToRad(90));
+LTransf := LTransf * TMatrix3D.CreateScaling(Point3D(2, 2, 2));
+GorillaModel1.Instances[2] := LTransf;
+</file>
+==== Instancing by template (TModelDef) ====
 Since 0.8.3.1966+ we've refactored the way to instanciate models. The previous method was incomplete in handling animated models.
 In the following example we show how to load a complex animated model with multiple meshes and animations.
-We will load up a template, which will be used for duplication. The template itself will not be rendered in our example, but of course that's possible too.
+We will load up a template (TModelDef), which will be used for duplication. The template (TModelDef) itself will not be rendered in our example, but of course that's possible too.
 Each duplicated model will run a different animation to show instanciating is still fast, but allows to handle models differently.