Offset / Scale in NoiseTool Texture Preview

[Spannerworx]

Hello, I have a question around offsets and scale, because the noise I generate myself via code looks slightly different to how it does in the NoiseTool Texture Preview.

Does the NoiseTool do anything special with the inputs/outputs?

For example, I want to use the DistanceToPoint module. It looks fine in the NoiseTool Texture Preview, and is centred perfectly:

However if I generate it myself, it looks to be zoomed in significantly, and is also not automatically centred – which I had assumed was what this module did. Its the sprite on the right of this screenshot:

The main lines of code, and the tool encode string I used:

FastNoise fnCircularValue = FastNoise.FromEncodedNodeTree("BQABAAAAAAAAAAAAAAAAAAAAAAAAAA==");
minMax = fnCircularValue.GenUniformGrid2D(noiseData, 0, 0, 512, 512, frequency, seed);

Thanks in advance.

[Auburn]

The preview in the noise tool defaults to 0,0 being in the centre, to replicate this in your code you'd want to do GenUniformGrid2D(noiseData, -256, -256, 512, 512, frequency, seed)

[Spannerworx]

Ah, yes - that does the trick, thank you.

Do you know why my zoom level / scale is so off?

[Auburn]

Are you setting the frequency value to the same as preview window, that will affect the scale

[Spannerworx]

Yup, same value: 0.022

    void Start()
    {
        FastNoise fnCircularValue = FastNoise.FromEncodedNodeTree("BQABAAAAAAAAAAAAAAAAAAAAAAAAAA==");

        int seed = 2;
        int size = 512;
        int start = -size / 2;
        float frequency = 0.022f;
        float[] noiseData;
        Texture2D texture = null;
        OutputMinMax minMax;

        noiseData = new float[size * size];
        minMax = fnCircularValue.GenUniformGrid2D(noiseData, start, start, size, size, frequency, seed);
        SetPixels(noiseData, minMax, size, ref texture);
        ApplyTextureToSprite(transform.Find("DistToPoint").gameObject, texture);
    }

    private void SetPixels(float[] noiseOut, OutputMinMax minMax, int texSize, ref Texture2D tex)
    {
        float scale = 255.0f / (minMax.max - minMax.min);

        Color32[] pixels = new Color32[texSize * texSize];
        Color32 color = Color.white;
        for (int i = 0; i < texSize * texSize; ++i)
        {
            float noise = noiseOut[i];
            noise = Mathf.Round((noise - minMax.min) * scale);
            noise = Mathf.Clamp(noise, 0, 255);
            color.r = color.g = color.b = (byte)noise;
            pixels[i] = color;
        }

        tex = new Texture2D(texSize, texSize);
        tex.SetPixels32(pixels);
        tex.Apply();
    }

    private void ApplyTextureToSprite(GameObject GO, Texture2D tex)
    {
        // Create a Rect representing the entire texture
        Rect rect = new Rect(0, 0, tex.width, tex.height);

        // Create a new sprite from the texture
        Sprite newSprite = Sprite.Create(tex, rect, new Vector2(0.5f, 0.5f));

        SpriteRenderer spriteRenderer = GO.GetComponent<SpriteRenderer>();
        spriteRenderer.sprite = newSprite;
    }

[Spannerworx]

@Auburn is the issue/difference likely to be down to something my side, do you think?

I am using the latest version from the downloads page: v0.10.0-alpha

I am using the DistanceToPoint module to capture a circular area that I can then scale it and subtract from other inputs, to achieve this:

[Auburn]

The code for that preview is here, https://github.com/Auburn/FastNoise2/blob/master/NoiseTool/NoiseTexture.cpp#L330
So you can compare it to yours

[Spannerworx]

Not sure where I'm going wrong.

In case it helps, using exact same code as I posted above, with the same frequency in tool and code, I get what I expect from nearly all noise.
For example this is Sine Wave and Checker board, tool on the left and Unity on the right:

Below are Distance To Point (scale/zoom is wrong) and Cellular Distance (comes out too dark):

Hope that helps.

[Auburn]

Both cellular distance have a min output of 0, checkerboard and sine wave have a min of -1. The noise tool shows -1 as black and 1 as white, so it's expected that the minimum is grey. Are you scaling your output colour based on the generated min/max?

[Auburn]

Both cellular distance have a min output of 0, checkerboard and sine wave have a min of -1. The noise tool shows -1 as black and 1 as white, so it's expected that the minimum is grey.
Are you scaling your output colour based on the generated min/max? That would explain both your examples

[Spannerworx]

Yes you are spot on. My scaling is working against me and not giving me what I was expecting.
In the cases where the min/max output is 0 to 1, its almost simply just a case of noise * 255 to get the right intensity.
Thanks for your help and patience Jordan.