are vertex positions in a strip interpreted as percent of the texture's native dimension? or am I doing something wrong here

[wayfinder]

Caveat: I'm working with Phaser, so this is pixi v2, but the general principles shouldn't have changed too much.

 

I'm creating a Strip with a vertex array like so:

        var vertices = new PIXI.Float32Array([0, 0, 100, 0, 0, 100, 100, 100]);        var uvs = new PIXI.Float32Array([0, 0, 1, 0, 0, 1, 1, 1]);        var colors = new PIXI.Float32Array([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]);        var indices = new PIXI.Uint16Array([0, 1, 2, 3]);                this.strip = new Phaser.Strip(this.game, 0, 0, 'temp/rockfill512_2', vertices, uvs, colors, indices);

(Phaser.Strip is not much more than a wrapper for PIXI.Strip, it simply passes the arrays on)

 

Now, what I would expect to happen from this is a rectangle at 0, 0 with a width and height of 100px each, and textured with one full copy of the texture, scaled down to fit the size of the rectangle, like so:

 

What actually happens is a rectangle the size of the texture's native dimension (512x512):

 

Playing around with the vertex values, it quickly becomes apparent that the strip isn't interpreting the vertex values as raw pixels, i.e. placing them at (x, y), but instead placing them at (x * texture.width / 100, y * texture.height / 100), i.e. interpreting them as percentages of the native dimensions of the texture. 

 

Is this expected behaviour? It seems exceedingly weird to me...

 

Here's the vertex shader code I'm using:

    this.vertexSrc  = [        'attribute vec2 aVertexPosition;',        'attribute vec2 aTextureCoord;',        'attribute vec4 aVertexColor;',        'uniform mat3 translationMatrix;',        'uniform vec2 projectionVector;',        'uniform vec2 offsetVector;',        'uniform vec3 tint;',        'varying vec2 vTextureCoord;',        'varying vec4 vColor;',        'void main(void) {',        '   vec3 v = translationMatrix * vec3(aVertexPosition, 1.0);',        '   v -= offsetVector.xyx;',        '   gl_Position = vec4(v.x / projectionVector.x - 1.0, v.y / -projectionVector.y + 1.0, 0.0, 1.0);',        '   vTextureCoord = aTextureCoord;',        '   vColor = aVertexColor * vec4(tint, 1.0);',        '}'    ];

If I multiply all numbers in my vertex Array by 100/texture.width (resp. 100/texture.height for y values) before rendering, the result looks as expected. But why?

 

Can you help me understand what's going on?

 

 

[wayfinder]

edit: thought i had something but no.

[xerver]

It is "percentages of the native dimensions" because the vertices are normalized, that is on purpose. You should be using normalized verts.

[wayfinder]

But where does the factor 100 come from? That's what I don't get... I would expect normalized values to be something I'd generate with (x/texture.width, y/texture.height), like the UVs. So with normalized values if I wanted to place a vertex at 140, -32 and my baseTexture had the dimensions 512 x 1024, I'd have to set the vertex value to (140/512, -32/1024). My current situation though is that I'd have to set the vertex value to (140 * 100/512, -32 * 100/1024), and I don't see why...

[wayfinder]

Won't let that deter me though

 

 

[xerver]

Hmm, I think your calculations are off somewhere. The rope/mesh classes don't do anything with a 100 magic number. There also isn't a magic number in the example:

 

http://pixijs.github.io/examples/index.html?s=demos&f=strip-demo.js&title=Strip

[wayfinder]

 

There's nothing using "100" or "0.01" in my code (or Phaser's), do you have an idea where would I even start to look? None of the other kinds of objects (sprites, tilesprites, graphics) in my project behave like this (but of course, I'm not setting their vertices directly)...

[wayfinder]

Okay, I did some more testing, and it appears that this has something to do with the initial texture—when I set a new texture on the strip, things go back to as I initially expected them to be, ie vertex values are interpreted as pixels and not normalized, and no factor of 100. Exceedingly weird.