CODEWITHSUNDEEP
javascriptopengl-esglslwebgl
Iter Ator
Iter Ator

Reputation: 9299

How to detect float precisions in WebGL?

lowp, mediump and highp means not the same precision on each device. How is it possible to detect the actual values of them (the bit length of the float at each precision)?

I mean:

- lowp:   16bit float
- mediump 32bit float
- highp   64bit float (double)

Upvotes: 1

Views: 787

Answers (1)

user128511
user128511

Reputation:

you can call gl.getShaderPrecisionFormat(shaderType, precisionType)

shaderType is either gl.VERTEX_SHADER or gl.FRAGMENT_SHADER.

precisionType is one of gl.LOW_FLOAT, gl.GL_MEDIUM_FLOAT, gl.HIGH_FLOAT, gl.LOW_INT, gl.MEDIUM_INT, gl.GL_HIGH_INT.

It returns an object that looks like this

{
    rangeMin,  // log 2 of the minimum representable magnitude
    rangeMax,  // log 2 of the maximum representable magnitude
    precision, // log 2 of the precision
};

If a high precision floating-point format is not supported for fragment shaders, calling gl.getShaderPrecisionFormat with arguments gl.FRAGMENT_SHADER and gl.HIGH_FLOAT will return 0 for both range and precision.

docs are here which link to the OpenGL ES 2.0 docs here

Upvotes: 5

Related Questions