[Book2] Incorrect marble pattern at the final Perlin noise section

[MaximusPrimeForever]

Hi, I'm implementing this code in Rust instead of CPP so I very much hope that this isn't due to language differences.

The issue

Anyways, in the Perlin Noise chapter at the marble pattern section the book proposes this modification to get the marble texture:

color value(double u, double v, const point3& p) const override {
    auto s = scale * p;
    return color(1,1,1) * 0.5 * (1 + sin(s.z() + 10*noise.turb(s)));  // << this line
}

Here is my equivalent Rust code:

fn value(&self, _: f64, _: f64, point: &Point3) -> Color {
    let scaled_point = self.scale * (*point);
    Color::new(1.0, 1.0, 1.0) * 0.5 * (1.0 + (scaled_point.z() + 10.0 * self.noise.turbulence(scaled_point, 7)).sin())
}

And this is the image produced:

This is doesn't really look like marble, and does not resemble the following example from the book:

Potential solution

I've played around with the code, and noticed that if I pass the raw point (unscaled) to the turbulence function like so:

fn value(&self, _: f64, _: f64, point: &Point3) -> Color {
    let scaled_point = self.scale * (*point);
    Color::new(1.0, 1.0, 1.0) * 0.5 * (1.0 + (scaled_point.z() + 10.0 * self.noise.turbulence(*point, 7)).sin()) // << *point instead of scaled_point
}

I get the following image:

Which appears much closer to the intended result.
I cannot say I understand the math behind this chapter very well, so I don't know if this is the correct solution, but wanted to share anyway.

Hope this is helpful!

[gau-nernst]

It seems like your solution matches the code in v3: https://raytracing.github.io/v3/books/RayTracingTheNextWeek.html#perlinnoise/adjustingthephase

Scaling is only applied to the phase p.z()

[MaximusPrimeForever]

It seems like your solution matches the code in v3: https://raytracing.github.io/v3/books/RayTracingTheNextWeek.html#perlinnoise/adjustingthephase

Scaling is only applied to the phase p.z()

Looks like it's fixed for future releases then :)

[hollasch]

No, @gau-nernst is saying that the old version matches your solution, and the new version in development has the defect you point out.

At this point, your comment and solution look correct to me.

In order to properly test this, I'll need to be in the middle of the book 2 progression work that we have scheduled when the book 2 alpha is getting ready for release. I'll revisit then and likely apply your fix.

[hollasch]

Wait, no — I was looking at perlin.h when I should have been looking at texture.h. Testing now.

[hollasch]

Yup, looks good. Nice catch; thank you!

[hollasch]

Fixed.