My best guess concerning this is that the 33 bits are for the mantissa.
Then, the 13 bits are for the exponent. So that gives 46 bits. Then,
there's a sign bit for the mantissa and a sign bit for the exponent, thus
yielding 2 extra bits for a total of 48 bits. So, I guess it is a 48 bit
Okay. In all honesty, I'm not 100% sure what the 33 and 13 actually mean.
Especially the latter one, the precision. The docs say:
precision
Specifies a pointer to the location in which the log2 of the precision of
the format is returned.
So, if that number is the log2 of the precision, the p
Those number look correct. High Precision support is NOT required to be
implemented in the fragment shader and if your implementation does not
support high precision you will get 0s from the glGetShaderPrecisionFormat
query.
On Wednesday, January 30, 2013 8:32:43 PM UTC, bob wrote:
>
> Oh, ok.
Oh, ok. Thanks.
It works if I do this:
GLES20.glGetShaderPrecisionFormat(GLES20.GL_FRAGMENT_SHADER,
GLES20.GL_MEDIUM_FLOAT, range, 0, precision, 0);
range = [33, 33]
precision = [13]
On Wednesday, January 30, 2013 2:27:26 PM UTC-6, RichardC wrote:
>
> 0 means precision format is not support
0 means precision format is not supported
*"If a high precision floating-point format is not supported for fragment
shaders, calling glGetShaderPrecisionFormat with argumentsGL_FRAGMENT_SHADER
and GL_HIGH_FLOAT will return 0 for both range and precision. Support for
a high precision floating-po
5 matches
Mail list logo
