Today we are porting a ray tracer from Peter Shirley’s mini book, into a Swift playground. Since I am not going to describe what Ray Tracing is and how it works, I am inviting you to go ahead and read this book as it’s free for Kindle subscribers. If you are not a subscriber, just buy the book like I did. Paying $2.99 is absolutely worth your every penny if you are interested in this topic.
The first thing we would want to do is create a data structure to hold pixel information. In the playground, create a new file named pixel.swift under the Sources folder. Next, let’s write the Pixelstruct. It’s just a simple struct with one variable for each of the RGBA channels. We initialize the alpha channel to 255 which means completely opaque on a [0 - 255] scale:
public struct Pixel {
var r: UInt8
var g: UInt8
var b: UInt8
var a: UInt8
init(red: UInt8, green: UInt8, blue: UInt8) {
r = red
g = green
b = blue
a = 255
}
}
Next, we need to create an array to hold all the pixels on the screen. To compute the color for each pixel we just set Red to be 0 for all pixels, while Green goes from 0 (no Green at all) in the left corner of the screen, to 255 (full Green) in the right corner of the screen. Similarly, the Blue color goes from 0 at the top of the screen, to 255 at the bottom of the screen.
public func makePixelSet(width: Int, _ height: Int) -> ([Pixel], Int, Int) {
var pixel = Pixel(red: 0, green: 0, blue: 0)
var pixels = [Pixel](count: width * height, repeatedValue: pixel)
for i in 0..<width {
for j in 0..<height {
pixel = Pixel(red: 0, green: UInt8(Double(i * 255 / width)), blue: UInt8(Double(j * 255 / height)))
pixels[i + j * width] = pixel
}
}
return (pixels, width, height)
}
Finally, we need a way to create a drawable image from pixels. The Core Image framework provides the CGImageCreate() method, which takes in several parameters necessary for image rendering:
public func imageFromPixels(pixels: ([Pixel], width: Int, height: Int)) -> CIImage {
let bitsPerComponent = 8
let bitsPerPixel = 32
let rgbColorSpace = CGColorSpaceCreateDeviceRGB()
let bitmapInfo = CGBitmapInfo(rawValue: CGImageAlphaInfo.PremultipliedLast.rawValue) // alpha is last
let providerRef = CGDataProviderCreateWithCFData(NSData(bytes: pixels.0, length: pixels.0.count * sizeof(Pixel)))
let image = CGImageCreate(pixels.1, pixels.2, bitsPerComponent, bitsPerPixel, pixels.1 * sizeof(Pixel), rgbColorSpace, bitmapInfo, providerRef, nil, true, CGColorRenderingIntent.RenderingIntentDefault)
return CIImage(CGImage: image!)
}
Next, in the main playground page create the set of pixels for a window with given width and height, and then create the rendered image using this set:
let width = 800
let height = 400
var pixelSet = makePixelSet(width, height)
var image = imageFromPixels(pixelSet)
image
You should see the following image:
Ok, you might be wondering, but where does ray tracing come into play? Let’s deal with that next. Under the Sources folder again, let’s create a convenience class named vec3.swift where we will have a few math helper methods:
struct vec3 {
var x = 0.0
var y = 0.0
var z = 0.0
}
func * (left: Double, right: vec3) -> vec3 {
return vec3(x: left * right.x, y: left * right.y, z: left * right.z)
}
func + (left: vec3, right: vec3) -> vec3 {
return vec3(x: left.x + right.x, y: left.y + right.y, z: left.z + right.z)
}
func - (left: vec3, right: vec3) -> vec3 {
return vec3(x: left.x - right.x, y: left.y - right.y, z: left.z - right.z)
}
func dot (left: vec3, _ right: vec3) -> Double {
return left.x * right.x + left.y * right.y + left.z * right.z
}
func unit_vector(v: vec3) -> vec3 {
let length : Double = sqrt(dot(v, v))
return vec3(x: v.x/length, y: v.y/length, z: v.z/length)
}
Next, we need to create a ray struct. It has as members an origin, a direction and a method that computes the ray tracing equation for any given parameter:
struct ray {
var origin: vec3
var direction: vec3
func point_at_parameter(t: Double) -> vec3 {
return origin + t * direction
}
}
Then we need to compute the color based on whether the ray hits a sphere which we create in the center of the screen:
func color(r: ray) -> vec3 {
let minusZ = vec3(x: 0, y: 0, z: -1.0)
var t = hit_sphere(minusZ, 0.5, r)
if t > 0.0 {
let norm = unit_vector(r.point_at_parameter(t) - minusZ)
return 0.5 * vec3(x: norm.x + 1.0, y: norm.y + 1.0, z: norm.z + 1.0)
}
let unit_direction = unit_vector(r.direction)
t = 0.5 * (unit_direction.y + 1.0)
return (1.0 - t) * vec3(x: 1.0, y: 1.0, z: 1.0) + t * vec3(x: 0.5, y: 0.7, z: 1.0)
}
You noticed that we used another method, called hit_sphere() to determine whether we hit the sphere with the ray, or whether we miss it:
func hit_sphere(center: vec3, _ radius: Double, _ r: ray) -> Double {
let oc = r.origin - center
let a = dot(r.direction, r.direction)
let b = 2.0 * dot(oc, r.direction)
let c = dot(oc, oc) - radius * radius
let discriminant = b * b - 4 * a * c
if discriminant < 0 {
return -1.0
} else {
return (-b - sqrt(discriminant)) / (2.0 * a)
}
}
Back in the pixel.swift file, change the makePixelSet(:) to include creating a ray for each pixel and compute its color before adding the pixel to the set:
public func makePixelSet(width: Int, _ height: Int) -> ([Pixel], Int, Int) {
var pixel = Pixel(red: 0, green: 0, blue: 0)
var pixels = [Pixel](count: width * height, repeatedValue: pixel)
let lower_left_corner = vec3(x: -2.0, y: 1.0, z: -1.0)
let horizontal = vec3(x: 4.0, y: 0, z: 0)
let vertical = vec3(x: 0, y: -2.0, z: 0)
let origin = vec3()
for i in 0..<width {
for j in 0..<height {
let u = Double(i) / Double(width)
let v = Double(j) / Double(height)
let r = ray(origin: origin, direction: lower_left_corner + u * horizontal + v * vertical)
let col = color(r)
pixel = Pixel(red: UInt8(col.x * 255), green: UInt8(col.y * 255), blue: UInt8(col.z * 255))
pixels[i + j * width] = pixel
}
}
return (pixels, width, height)
}
In the main playground page, see the generated new image:
Stay tuned for a part 2 of this article, where we will calculate lights and shades for a more realistic image rendering. The source code is posted on Github as usual.
Until next time!
Everything about the Metal API 