What is this mipmap artifact?

[Thaina]

There was a weird white line at the edge of my sphere UV when I import texture with mipmap

It disappear when I disable mipmap

What is this?

[Unique-Player]

My guess is the position of your UV, but since this is sphere i guess it is something with your texture.
Is your texture filled by color only on UV islands or do you have blurred edges?
Look at the image, left one does not have anyting beside on the UV islands, right one has background color to hide seems.

[bgolus]

It’s bilinear filtering and mip mapping with out adequate padding on UV seams. You usually want to keep a good 4~8 pixels of padding around the edges of UV seams with that edge’s color.

Try taking your texture in Photoshop, scale it down to 25%, then scale it back up to the original size using bilinear upscaling. Use that new texture in your modelling program to get a sense of what mipmapping will do.

[Thaina]

I have try tinkering with UV padding but no it’s use. Turn out it was this problem

Strange line artifact when doing %1 on uv coordinates

ok so i have a shader and it is necessary to do a modulus 1 on an uv coordinate. but for some weird reason i get this artifact at the 0 point when doing this. im working of the surface shader template. running unity 2018.2.2f1 on windows 10 64 bit [image] for demonstration this is done in a less complex shader to isolate the issue. code for shader is here: Shader "Custom/bug" { Properties { _MainTex ("Albedo (RGB)", 2D) = "white" {} [Toggle(ENABLE_BUG)] _Bug ("Bug", Float) = …

Just figure out it actually have tex2D sampler LOD for this. I need to control LOD by myself with my self generated UV

[bgolus]

I prefer using tex2Dgrad as I find it’s easier to use, once you understand how to use it at least.

The other potentially even easier option is this:
http://vcg.isti.cnr.it/~tarini/no-seams/

For your case, I’m guessing you’re calculating equilateral projection for the UVs in the fragment shader. You’d need to calculate at least the uv.x twice, like this:

float x_atan = atan(x, z) / (2.0 * UNITY_PI);
float xA = x_atan + 0.5;
float xB = frac(x_atan + 1.0) - 0.5;
float2 uv = float2(
  fwidth(xA) < fwidth(xB) ? xA : xB,
  acos(-y) / UNITY_PI
  );
fixed4 col = tex2D(_MainTex, uv);

[Thaina]

@bgolus I think it was a problem of texture 2D lookup node of shadergraph itself so only calculating UV don’t solve this. I even try to clamp the UV to 0.01-0.99 and the artifact still occur

[bgolus]

@Thaina

I think it was a problem of texture 2D lookup node of shadergraph itself

Shader Graph is just generating normal HLSL, there’s nothing special or unique about its Sample Texture 2D node vs the tex2D() function.

@Thaina

I even try to clamp the UV to 0.01-0.99 and the artifact still occur

Because if the issues is the one in that forum post you linked to, clamping won’t solve the main problem.

@bgolus

The problem comes from how GPUs determine what mipmap level to use for a texture. The short version is the GPU looks to see how much the UV changes from one pixel to the one next to it, and from that it figures out the best mip level to sample the texture at. These are per pixel derivatives. Since you’re using a modulo on the UVs, on one pixel the UV x value might be 0.99, and the next 0.01. So between those two pixels as far as the GPU knows the entire texture has to be shown, so it drops down to the lowest available mip level.

As for why it’s not a constant line, derivatives are calculated for each 2x2 set of pixels. These are called pixel quads. If the modulo happens between two of these quads, the derivatives are don’t see the sudden break in the UVs. If the modulo happens within a quad, then you get the above.

In the case of the post the above quote is in response to, that person was taking a continuous UV and using a modulo on it. This introduces a discontinuity within the UVs that the derivatives see as a sudden, very high rate of change. In your case, if you’re generating UVs in the fragment shader using atan(), that function produces a value range from -π to +π, which you would then scale to a 0.0 to 1.0 range. You’re not using a frac() or a modulo, but there’s still an inherent discontinuity in the value as it wraps around causing the same problem. The trick in the code above is within the 2x2 pixel quads where the sudden discontinuity appears instead of using the normal UV that goes from from 1.0 to 0.0 it uses a continuous UV with a range from 0.5 to 1.5, so for that edge it’s a continuous 1.0 to 1.01.

Here’s the Shader Graph equivalent layout:

And here it is with the original “broken” setup:

(Note: The texture needs to be set to Wrap Mode: Repeat, and not Clamp for the above to work properly)