Revision as of 15:22, 13 January 2012

Overview

This is a Sims 3:RCOL chunk for Body Geometry.

Format

DWORD 'GEOM'
DWORD Version			// 5
DWORD tgi offset		// see Sims 3:Key table
DWORD tgi size			// see Sims 3:Key table
DWORD EmbeddedID		// if no embedded material, will be 0
                                // otherwise, if the "morphskin" or "morphskincloth" shaders are used in the material, it will be the FNV32 hash of "SimSkin",
                                // and if the "morpheye" shader is used, it will be the FNV32 hash of "SimEyes"
--if (EmbeddedID != 0):
	DWORD ChunkSize
	--insert MTNF chunk	// follows the same format of MTNF from Sims 3:0x01D0E75D, starting at the 'MTNF' tag

DWORD mergeGroup
DWORD sortOrder                 // often zero

DWORD NumVerts			// Number of vertices
DWORD FCount			// Number of vertex elements
--repetition FCount:
	--insert Vertex Format	// see below
--repetition NumVerts:
	--insert Vertex Data	// see below

DWORD ItemCount                 // will usually be 1 (hardcoded in pipeline exporter), but sometimes other numbers
--repetition ItemCount:
	BYTE BytesPerFacePoint

DWORD NumFacePoints		// means faces*3
--repetition NumFacePoints:
	WORD[3]			// standard face index (three unsigned words per face) for 2-byte FaceFmt

DWORD index 			// to a Skin Controller(0x00AE6C67)

DWORD count1
--repetition count1:
	DWORD bonehasharray	// 32-bit hash of used bone names.

--insert TGI Block List	// see Sims 3:Key table

Vertex Format

DWORD DataType
DWORD SubType
BYTE BytesPerElement

DataType

SubType

You calculate the offset of each element from the sum of the previous BytesPerElement

Based on contributions from karybdis and atavera

Vertex Data

Reading Vertex Data Blocks

For each vertex, you need to loop through the VertexFormat as defined, and read the information in order.

The order can be completely arbitary, and varied in size.

Here is some example code in c# to do so:

for (int i = 0; i < numVerts; i++)
{
	for (int j = 0; j < vertexFormats.Count; j++)
	{
		float x = 0;
		float y = 0;
		float z = 0;
		vertexFormat vf = (vertexFormat)vertexFormats[j];
		switch (vf.dataType)
		{
			case 1:
				x = reader.ReadSingle();
				z = reader.ReadSingle();
				y = reader.ReadSingle();
				sb.AppendLine("  XYZ: " + x.ToString() + " " + y.ToString() + " " + z.ToString() );
				break;
			case 2:
				x = reader.ReadSingle();
				z = reader.ReadSingle();
				y = reader.ReadSingle();
				sb.AppendLine("  Normal: " + x.ToString() + " " + y.ToString() + " " + z.ToString() );
				break;
			case 3:
				float u = reader.ReadSingle();
				float v = reader.ReadSingle();
				sb.AppendLine("  UV: " + u.ToString() + " " + v.ToString() );
				break;
			case 4:
				sb.AppendLine("  Bone: " + reader.ReadUInt32().ToString() );
				break;
			case 5:
				float w1 = reader.ReadSingle();
				float w2 = reader.ReadSingle();
				float w3 = reader.ReadSingle();
				float w4 = reader.ReadSingle();
				sb.AppendLine("  Weights: " + w1.ToString() + " " + w2.ToString() + " "
					+ w3.ToString() + " " + w4.ToString() );
				break;
			case 6:
				x = reader.ReadSingle();
				z = reader.ReadSingle();
				y = reader.ReadSingle();
				sb.AppendLine("  Tangent Normal: " + x.ToString() + " " + y.ToString() + " " + z.ToString() );
				break;
			case 7:
				// Note, not splitting this up yet just reading an int
				sb.AppendLine("  TagVal: " + reader.ReadUInt32().ToString() );
				break;
			case 10:
				sb.AppendLine("  VertexID: " + reader.ReadUInt32().ToString() );
				break;
		}
	}
}

Note that basically we just look through the VertexFormat, and read the data in order. There is no fixed "vertex format" version, it's all controlled via the vertexformat data.