basic fully working octree renderer

This commit is contained in:
2024-09-10 13:19:45 -04:00
parent 53c4d067e2
commit 3417250544
9 changed files with 572 additions and 314 deletions

498
Cargo.lock generated

File diff suppressed because it is too large Load Diff

View File

@@ -10,14 +10,13 @@ rustflags = ["-Z", "threads=14"]
[dependencies]
bytemuck = {version="1.14.0", features=["derive"]}
nalgebra = {version="0.32.5", features=["bytemuck"]}
nalgebra = {version="0.33.0", features=["bytemuck"]}
ndarray = "0.15.6"
pollster = "0.3"
rand = "0.8.5"
simba = "0.8.1"
simdnoise = "3.1.6"
smaa = "0.14.0"
wgpu = "0.20.1"
wgpu = "22.0.0"
bevy_ecs = "0.13.2"
bevy_derive = "0.13.2"
winit = {version="0.30.1", features=["serde"]}

View File

@@ -55,7 +55,7 @@ impl Client<'_> {
// camera orientation
let old_camera = state.camera;
if self.grabbed_cursor {
let delta = input.mouse_delta * 0.003;
let delta = input.mouse_delta * 0.003 / state.camera.scale;
if delta.x != 0.0 {
state.camera.orientation = Rotation3::from_axis_angle(&state.camera.up(), delta.x)
* state.camera.orientation;

View File

@@ -83,6 +83,12 @@ impl Input {
self.mouse_just_released.insert(button);
}
},
WindowEvent::MouseWheel { device_id, delta, phase } => {
self.scroll_delta = match delta {
MouseScrollDelta::LineDelta(_, v) => v,
MouseScrollDelta::PixelDelta(v) => (v.y / 2.0) as f32,
};
}
_ => (),
}
}

View File

@@ -6,7 +6,6 @@ pub use command::*;
use super::camera::Camera;
use crate::client::rsc::CLEAR_COLOR;
use nalgebra::Vector2;
use smaa::{SmaaMode, SmaaTarget};
use util::Texture;
use voxel::VoxelPipeline;
use winit::dpi::PhysicalSize;
@@ -20,7 +19,6 @@ pub struct Renderer<'a> {
config: wgpu::SurfaceConfiguration,
staging_belt: wgpu::util::StagingBelt,
voxel_pipeline: VoxelPipeline,
smaa_target: SmaaTarget,
camera: Camera,
depth_texture: Texture,
}
@@ -43,6 +41,7 @@ impl<'a> Renderer<'a> {
label: None,
required_features: wgpu::Features::empty(),
required_limits: wgpu::Limits::default(),
memory_hints: wgpu::MemoryHints::default(),
},
None, // Trace path
))
@@ -79,15 +78,6 @@ impl<'a> Renderer<'a> {
// doesn't affect performance much and depends on "normal" zoom
let staging_belt = wgpu::util::StagingBelt::new(4096 * 4);
let smaa_target = SmaaTarget::new(
&device,
&queue,
size.width,
size.height,
surface_format,
SmaaMode::Smaa1X,
);
let depth_texture = Texture::create_depth_texture(&device, &config, "depth_texture");
Self {
@@ -100,7 +90,6 @@ impl<'a> Renderer<'a> {
device,
config,
queue,
smaa_target,
depth_texture,
}
}
@@ -117,14 +106,10 @@ impl<'a> Renderer<'a> {
let view = output
.texture
.create_view(&wgpu::TextureViewDescriptor::default());
let smaa_frame = self
.smaa_target
.start_frame(&self.device, &self.queue, &view);
{
let render_pass = &mut encoder.begin_render_pass(&wgpu::RenderPassDescriptor {
let mut render_pass = encoder.begin_render_pass(&wgpu::RenderPassDescriptor {
label: Some("Render Pass"),
color_attachments: &[Some(wgpu::RenderPassColorAttachment {
view: &smaa_frame,
view: &view,
resolve_target: None,
ops: wgpu::Operations {
load: wgpu::LoadOp::Clear(CLEAR_COLOR),
@@ -142,9 +127,8 @@ impl<'a> Renderer<'a> {
timestamp_writes: None,
occlusion_query_set: None,
});
self.voxel_pipeline.draw(render_pass);
}
smaa_frame.resolve();
self.voxel_pipeline.draw(&mut render_pass);
drop(render_pass);
self.staging_belt.finish();
self.queue.submit(std::iter::once(encoder.finish()));
@@ -157,8 +141,6 @@ impl<'a> Renderer<'a> {
self.config.width = size.width;
self.config.height = size.height;
self.surface.configure(&self.device, &self.config);
self.smaa_target
.resize(&self.device, size.width, size.height);
self.depth_texture =
Texture::create_depth_texture(&self.device, &self.config, "depth_texture");

View File

@@ -5,6 +5,7 @@ mod light;
mod view;
pub use color::*;
use wgpu::include_wgsl;
use super::super::UpdateGridTransform;
use crate::{client::{
@@ -35,10 +36,7 @@ pub struct VoxelPipeline {
impl VoxelPipeline {
pub fn new(device: &wgpu::Device, config: &wgpu::SurfaceConfiguration) -> Self {
// shaders
let shader = device.create_shader_module(wgpu::ShaderModuleDescriptor {
label: Some("Tile Shader"),
source: wgpu::ShaderSource::Wgsl(include_str!("shader.wgsl").into()),
});
let shader = device.create_shader_module(include_wgsl!("shader.wgsl"));
let view = Uniform::init(device, "view", 0);
let voxels = Storage::init(device, "voxels", 1);
@@ -123,6 +121,7 @@ impl VoxelPipeline {
alpha_to_coverage_enabled: true,
},
multiview: None,
cache: None,
});
Self {

View File

@@ -57,7 +57,7 @@ fn fs_main(
let aspect = win_dim.y / win_dim.x;
let pixel_pos = vec3<f32>(
(in.clip_position.xy / win_dim - vec2<f32>(0.5)) * vec2<f32>(2.0, -2.0 * aspect),
1.0
view.zoom
);
// move to position in world
@@ -127,6 +127,7 @@ fn trace_full(pos_view: vec4<f32>, dir_view: vec4<f32>) -> vec4<f32> {
t = select(select(t_i.z, t_i.y, hit.y), t_i.x, hit.x);
next_normal = select(select(normals[2], normals[1], hit.y), normals[0], hit.x);
}
// voxel position relative to low_corner
var vox_pos = clamp(vec3<i32>(pos), vec3<i32>(0), dim_i - vec3<i32>(1));
@@ -144,11 +145,10 @@ fn trace_full(pos_view: vec4<f32>, dir_view: vec4<f32>) -> vec4<f32> {
var data_start = 1u;
var i = 0u;
var axis = 0;
var hits = 0;
for (var safety = 0; safety < 1000; safety += 1) {
let node = voxels[group.offset + i];
if node >= LEAF_BIT {
hits += 1;
// leaf
let vcolor = get_color(node & LEAF_MASK);
if vcolor.a > 0.0 {
let diffuse = max(dot(global_lights[0].dir, next_normal) + 0.1, 0.0);
@@ -156,6 +156,7 @@ fn trace_full(pos_view: vec4<f32>, dir_view: vec4<f32>) -> vec4<f32> {
let lighting = max(diffuse, ambient);
let new_color = min(vcolor.xyz * lighting, vec3<f32>(1.0));
color += vec4<f32>(new_color.xyz * vcolor.a, vcolor.a) * (1.0 - color.a);
color = vec4<f32>(f32(safety) / 1000.0, 0.0, 0.0, 1.0);
if color.a > .999 {
return color;
}
@@ -171,14 +172,10 @@ fn trace_full(pos_view: vec4<f32>, dir_view: vec4<f32>) -> vec4<f32> {
pos = pos_start + t * dir;
let old = vox_pos[axis];
vox_pos = vec3<i32>(pos) - low_corner;
vox_pos[axis] += select(0, dir_i[axis], vox_pos[axis] == 0 || vox_pos[axis] == side_len - 1);
// if hits == 1 {
// // var axis_c = vec3<f32>(0.0);
// // axis_c[axis] = 1.0;
// // return vec4<f32>(axis_c, 1.0);
// return vec4<f32>(vec3<f32>(vox_pos), 1.0);
// }
vox_pos = clamp(vox_pos, vec3<i32>(0), side_len - vec3<i32>(1));
vox_pos[axis] += dir_i[axis];
} else if inside3i(vox_pos, vec3<i32>(0), vec3<i32>(side_len - 1)) {
// node
let node_pos = data_start + node;
side_len /= 2;
let vcorner = vox_pos / side_len;
@@ -194,7 +191,8 @@ fn trace_full(pos_view: vec4<f32>, dir_view: vec4<f32>) -> vec4<f32> {
// idrk what to put here tbh but this prolly works; don't zoom out if max
if side_len == 256 {
return color;
let a = f32(safety) / 1000.0;
return vec4<f32>(0.0, 0.0, a, 1.0);
}
// get parent info and reset "pointers" to parent

View File

@@ -0,0 +1,282 @@
// Vertex shader
struct GlobalLight {
dir: vec3<f32>,
};
struct VertexOutput {
@builtin(position) clip_position: vec4<f32>,
};
struct View {
transform: mat4x4<f32>,
width: u32,
height: u32,
zoom: f32,
};
struct VoxelGroup {
transform: mat4x4<f32>,
transform_inv: mat4x4<f32>,
dimensions: vec3<u32>,
offset: u32,
};
@group(0) @binding(0)
var<uniform> view: View;
@group(0) @binding(1)
var<storage, read> voxels: array<u32>;
@group(0) @binding(2)
var<storage, read> voxel_groups: array<VoxelGroup>;
@group(0) @binding(3)
var<storage, read> global_lights: array<GlobalLight>;
@vertex
fn vs_main(
@builtin(vertex_index) vi: u32,
@builtin(instance_index) ii: u32,
) -> VertexOutput {
var out: VertexOutput;
var pos = vec2<f32>(
f32(vi % 2u) * 2.0 - 1.0,
f32(vi / 2u) * 2.0 - 1.0,
) ;
out.clip_position = vec4<f32>(pos.x, pos.y, 0.0, 1.0);
return out;
}
// Fragment shader
@fragment
fn fs_main(
in: VertexOutput,
) -> @location(0) vec4<f32> {
// get position of the pixel; eye at origin, pixel on plane z = 1
let win_dim = vec2<f32>(f32(view.width), f32(view.height));
let aspect = win_dim.y / win_dim.x;
let pixel_pos = vec3<f32>(
(in.clip_position.xy / win_dim - vec2<f32>(0.5)) * vec2<f32>(2.0, -2.0 * aspect),
view.zoom
);
// move to position in world
let pos = view.transform * vec4<f32>(pixel_pos, 1.0);
let dir = view.transform * vec4<f32>(normalize(pixel_pos), 0.0);
var color = trace_full(pos, dir);
let light_mult = clamp((-dot(dir.xyz, global_lights[0].dir) - 0.99) * 200.0, 0.0, 1.0);
let sky_color = light_mult * vec3<f32>(1.0, 1.0, 1.0);
color += vec4<f32>(sky_color * (1.0 - color.a), 1.0 - color.a);
color.a = 1.0;
return color;
}
const ZERO3F = vec3<f32>(0.0);
const ZERO2F = vec2<f32>(0.0);
const DEPTH = 16u;
const FULL_ALPHA = 0.9999;
fn trace_full(pos_view: vec4<f32>, dir_view: vec4<f32>) -> vec4<f32> {
let gi = 0;
let group = voxel_groups[gi];
if group.dimensions.x == 0 {
return vec4<f32>(0.0);
}
let dim_f = vec3<f32>(group.dimensions);
let dim_i = vec3<i32>(group.dimensions);
// transform so that group is at 0,0
let pos_start = (group.transform_inv * pos_view).xyz;
let dir = (group.transform_inv * dir_view).xyz;
let dir_if = sign(dir);
// calculate normals
var normals = mat3x3<f32>(
(group.transform * vec4<f32>(dir_if.x, 0.0, 0.0, 0.0)).xyz,
(group.transform * vec4<f32>(0.0, dir_if.y, 0.0, 0.0)).xyz,
(group.transform * vec4<f32>(0.0, 0.0, dir_if.z, 0.0)).xyz,
);
var next_normal = vec3<f32>(0.0, 0.0, 0.0);
// find where ray intersects with group
let plane_point = (vec3<f32>(1.0) - dir_if) / 2.0 * dim_f;
var pos = pos_start;
var t = 0.0;
if outside3f(pos, ZERO3F, dim_f) {
// time of intersection; x = td + p, solve for t
let t_i = (plane_point - pos) / dir;
// points of intersection
let px = pos + t_i.x * dir;
let py = pos + t_i.y * dir;
let pz = pos + t_i.z * dir;
// check if point is in bounds
let hit = vec3<bool>(
inside2f(px.yz, ZERO2F, dim_f.yz),
inside2f(py.xz, ZERO2F, dim_f.xz),
inside2f(pz.xy, ZERO2F, dim_f.xy),
) && (t_i > ZERO3F);
if !any(hit) {
return vec4<f32>(0.0);
}
pos = select(select(pz, py, hit.y), px, hit.x);
t = select(select(t_i.z, t_i.y, hit.y), t_i.x, hit.x);
next_normal = select(select(normals[2], normals[1], hit.y), normals[0], hit.x);
}
// voxel position relative to low_corner
var vox_pos = clamp(vec3<i32>(pos), vec3<i32>(0), dim_i - vec3<i32>(1));
let dir_i = vec3<i32>(dir_if);
let dir_u = ((dir_i + vec3<i32>(1)) / 2);
let dir_bits = u32(dir_u.x * 4 + dir_u.y * 2 + dir_u.z);
// time to move 1 unit using dir
let inc_t = abs(1.0 / dir);
var side_len = 256;
// "unsigned" minimum cube coords of current tree
var low_corner = vec3<i32>(0);
// time of next 1 unit plane hit in each direction
var color = vec4<f32>(0.0);
var data_start = 1u;
var i = 0u;
var axis = 0;
var hits = 0;
for (var safety = 0; safety < 1000; safety += 1) {
let node = voxels[group.offset + i];
if node >= LEAF_BIT {
// leaf
hits += 1;
let vcolor = get_color(node & LEAF_MASK);
if vcolor.a > 0.0 {
let diffuse = max(dot(global_lights[0].dir, next_normal) + 0.1, 0.0);
let ambient = 0.2;
let lighting = max(diffuse, ambient);
let new_color = min(vcolor.xyz * lighting, vec3<f32>(1.0));
color += vec4<f32>(new_color.xyz * vcolor.a, vcolor.a) * (1.0 - color.a);
if color.a > .999 {
return color;
}
}
// move to next face of cube
let half_len = f32(side_len) / 2.0;
let corner = vec3<f32>(low_corner) + vec3<f32>(half_len) + dir_if * half_len;
let next_t = inc_t * abs(corner - pos_start);
axis = select(select(2, 1, next_t.y < next_t.z), 0, next_t.x < next_t.y && next_t.x < next_t.z);
t = next_t[axis];
next_normal = normals[axis];
pos = pos_start + t * dir;
let old = vox_pos[axis];
vox_pos = vec3<i32>(pos) - low_corner;
vox_pos = clamp(vox_pos, vec3<i32>(0), side_len - vec3<i32>(1));
vox_pos[axis] += dir_i[axis];
} else if inside3i(vox_pos, vec3<i32>(0), vec3<i32>(side_len - 1)) {
// node
let node_pos = data_start + node;
side_len /= 2;
let vcorner = vox_pos / side_len;
vox_pos -= vcorner * side_len;
let j = u32(vcorner.x * 4 + vcorner.y * 2 + vcorner.z);
i = node_pos + j;
data_start = node_pos + 9;
low_corner += vec3<i32>(dir_to_vec(j)) * i32(side_len);
continue;
}
// idrk what to put here tbh but this prolly works; don't zoom out if max
if side_len == 256 {
return color;
}
// get parent info and reset "pointers" to parent
let parent_info_i = data_start - 1;
let parent_info = voxels[group.offset + parent_info_i];
let parent_root = parent_info_i - (parent_info >> 3);
let parent_loc = parent_info & 7;
let loc = 8 - (data_start - 1 - i);
// let test = (parent_root + 9 + voxels[group.offset + parent_root + parent_loc] + loc) == i;
i = parent_root + parent_loc;
data_start = parent_root + 9;
// adjust corner back to parent
let low_corner_adj = vec3<i32>(dir_to_vec(loc)) * i32(side_len);
low_corner -= low_corner_adj;
// update vox pos to be relative to parent
vox_pos += low_corner_adj;
side_len *= 2;
// return vec4<f32>(vec3<f32>(dir_to_vec(parent_loc)) * f32(loc) / 8.0, 1.0);
// return vec4<f32>(vec3<f32>(f32(test)), 1.0);
}
return vec4<f32>(1.0, 0.0, 1.0, 1.0);
}
const LEAF_BIT = 1u << 31u;
const LEAF_MASK = ~LEAF_BIT;
// there's no way this is efficient, mod is faster for all I know
fn dir_to_vec(bits: u32) -> vec3<u32> {
return vec3<u32>(extractBits(bits, 2u, 1u), extractBits(bits, 1u, 1u), extractBits(bits, 0u, 1u));
}
fn get_voxel(offset: u32, pos_: vec3<u32>) -> u32 {
var data_start = 1u;
var i = 0u;
var pos = pos_;
var side_len: u32 = 256;
var safety = 0;
while voxels[offset + i] < LEAF_BIT {
let node_pos = data_start + voxels[offset + i];
side_len /= 2u;
let corner = pos / side_len;
pos -= corner * side_len;
let j = corner.x * 4 + corner.y * 2 + corner.z;
i = node_pos + j;
data_start = node_pos + 8;
if safety == 10 {
return 10u;
}
safety += 1;
}
return voxels[offset + i] & LEAF_MASK;
}
fn get_color(id: u32) -> vec4<f32> {
switch id {
case 0u: {
return vec4<f32>(0.0);
}
case 1u: {
return vec4<f32>(0.5, 0.5, 0.5, 1.0);
}
case 2u: {
return vec4<f32>(0.5, 1.0, 0.5, 1.0);
}
case 3u: {
return vec4<f32>(0.5, 0.5, 1.0, 0.5);
}
default: {
return vec4<f32>(1.0, 0.0, 0.0, 1.0);
}
}
}
fn outside3f(v: vec3<f32>, low: vec3<f32>, high: vec3<f32>) -> bool {
return any(v < low) || any(v > high);
}
fn inside2f(v: vec2<f32>, low: vec2<f32>, high: vec2<f32>) -> bool {
return all(v >= low) && all(v <= high);
}
fn inside3i(v: vec3<i32>, low: vec3<i32>, high: vec3<i32>) -> bool {
return all(v >= low) && all(v <= high);
}

View File

@@ -155,6 +155,8 @@ impl OctTree {
pub fn raw(&self) -> &[OctNode] {
&self.data
}
pub fn mesh(&self) {
}
}
pub struct OctTreeIter<'a> {