Overview
The Subchunk class represents a 16×16×16 subdivision of a chunk, responsible for generating optimized triangle meshes with smooth lighting and ambient occlusion.
class Subchunk {
public:
Chunk* parent;
World* world;
glm::ivec3 subchunk_position;
glm::ivec3 local_position;
glm::vec3 position;
std::vector<uint32_t> mesh;
std::vector<uint32_t> translucent_mesh;
Subchunk(Chunk* p, glm::ivec3 pos);
void update_mesh();
};
src/chunk/subchunk.h, src/chunk/subchunk.cpp
Constructor
Subchunk(Chunk* p, glm::ivec3 pos)
Creates a new subchunk within a parent chunk.
Parameters:
p - Parent chunk pointerpos - Subchunk position (0-7 in each axis within the chunk)
Example:
Subchunk* sub = new Subchunk(chunk, glm::ivec3(0, 2, 0));
Methods
update_mesh()
Rebuilds the mesh for this subchunk by iterating through all blocks and generating visible faces with smooth lighting and AO.
Process:
- Clear existing mesh data
- Iterate through all 4096 blocks (16×16×16)
- Check each of 6 faces for visibility
- Generate vertices with packed lighting data
- Separate opaque and translucent geometry
Performance: This is a CPU-intensive operation. The engine limits how many subchunks can rebuild per frame via Options::CHUNK_UPDATES.
Private Helper Methods
These methods are used internally during mesh generation:
smooth()
float smooth(float a, float b, float c, float d)
ao_val()
float ao_val(bool s1, bool s2, bool c)
s1, s2 - Two side neighborsc - Corner neighbor
Returns: AO factor (0-3 neighbors blocking → darker)
get_face_ao()
std::array<float, 4> get_face_ao(bool s1, bool s2, bool s3, bool s4,
bool s5, bool s6, bool s7, bool s8)
get_smooth_face_light()
std::array<float, 4> get_smooth_face_light(float light, float l1, float l2,
float l3, float l4, float l5,
float l6, float l7, float l8)
get_neighbour_voxels()
std::array<glm::ivec3, 8> get_neighbour_voxels(glm::ivec3 npos, int face)
npos - Position of the neighbor blockface - Face index (0-5)
Returns: 8 neighbor block positions
get_light() / get_skylight()
std::array<float, 4> get_light(int block, int face, glm::ivec3 pos, glm::ivec3 npos)
std::array<float, 4> get_skylight(int block, int face, glm::ivec3 pos, glm::ivec3 npos)
get_shading()
std::array<float, 4> get_shading(int block, BlockType& bt, int face, glm::ivec3 npos)
add_face()
void add_face(int face, glm::ivec3 pos, glm::ivec3 lpos, int block,
BlockType& bt, glm::ivec3 npos)
face - Face index (0-5)pos - Block world positionlpos - Local position within subchunkblock - Block type IDbt - BlockType referencenpos - Neighbor block position
can_render_face()
bool can_render_face(BlockType& bt, int block_number, glm::ivec3 position)
- Always render if neighbor is air or transparent
- For glass blocks, render only if neighbor is NOT glass
- Skip face if neighbor is the same opaque block
Returns: true if face should be rendered
Data Members
mesh / translucent_mesh
std::vector<uint32_t> mesh;
std::vector<uint32_t> translucent_mesh;
Position Fields
subchunk_position - Position within parent chunk (0-7 per axis)local_position - Position within chunk (in blocks)position - Absolute world position (in blocks)
Integration
Subchunks are managed by their parent Chunk:
// Chunk schedules subchunk mesh updates
chunk->update_subchunk_meshes();
// Chunk processes the queue
chunk->process_chunk_updates(); // Rebuilds up to CHUNK_UPDATES subchunks
// Chunk merges subchunk meshes
chunk->update_mesh(); // Combines all subchunk meshes into one VBO
- Mesh generation is expensive: Each subchunk can have up to 6 × 16³ = 24,576 potential faces
- Face culling: Most faces are culled (only exposed faces render)
- Smooth lighting: 9-sample interpolation per vertex adds overhead
- Update batching:
Options::CHUNK_UPDATES limits rebuilds per frame
Subchunk mesh updates are deferred and batched. After modifying blocks, call chunk->update_at_position() to queue affected subchunks rather than rebuilding immediately.
See Also