OES_texture_half_float
OES_texture_float

GL_OES_texture_half_float, GL_OES_texture_float

Benj Lipchak, Apple (lipchak 'at' apple.com)

Copyright (c) 2005-2013 The Khronos Group Inc. Copyright terms at
    http://www.khronos.org/registry/speccopyright.html

Khronos-approved extension specifications are updated in response to
issues and bugs prioritized by the Khronos OpenGL ES Working Group. For
extensions which have been promoted to a core Specification, fixes will
first appear in the latest version of that core Specification, and will
eventually be backported to the extension document. This policy is
described in more detail at
    https://www.khronos.org/registry/OpenGL/docs/update_policy.php

Please refer to the ARB_texture_float extension.

Ratified by the Khronos BOP, July 22, 2005.

Last Modified Date:  November 9, 2011

OpenGL ES Extension #36

This extension is written against the OpenGL ES 2.0 Specification.

OES_texture_3D affects the definition of this extension.

These extensions add texture formats with 16- (aka half float) and 32-bit
floating-point components.  The 32-bit floating-point components
are in the standard IEEE float format.  The 16-bit floating-point
components have 1 sign bit, 5 exponent bits, and 10 mantissa bits.
Floating-point components are clamped to the limits of the range
representable by their format.

The OES_texture_half_float extension string indicates that the
implementation supports 16-bit floating pt texture formats.

The OES_texture_float extension string indicates that the 
implementation supports 32-bit floating pt texture formats.

Both these extensions only require NEAREST magnification filter and 
NEAREST, and NEAREST_MIPMAP_NEAREST minification filters to be supported.

1.  What should we do if magnification filter for a texture with half-float
    or float channels is set to LINEAR.

    RESOLUTION: The texture will be marked as incomplete.  
    Only the NEAREST filter is supported.  
    
    The cost of doing a LINEAR filter for these texture formats can be
    quite prohibitive.  There was a discussion on having the shader 
    generate code to do LINEAR filter by making individual texture calls with a 
    NEAREST filter but again the computational and memory b/w costs decided 
    against mandating this approach.  The decision was that this extension
    would only enable NEAREST magnification filter.  Support for LINEAR
    magnification filter would be done through a separate extension.

2.  What should we do if minification filter is set to LINEAR or 
    LINEAR_MIPMAP_NEAREST, NEAREST_MIPMAP_LINEAR and LINEAR_MIPMAP_LINEAR.

    RESOLUTION: The texture will be marked as incomplete.  Only the NEAREST 
    and NEAREST_MIPMAP_NEAREST minification filters are supported.  

    This was decided for the same reasons given in issue #1.  The decision
    was that this extension would only enable NEAREST and NEAREST_MIPMAP_NEAREST
    minification filters, and the remaining OpenGL ES minification filters
    would be supported through a separate extension.
    
3.  Should CopyTexImage2D, CopyTexSubImage{2D|3D} be supported for textures
    with half-float and float channels?
    
    RESOLUTION: No.

Accepted by the <type> parameter of TexImage2D,  TexSubImage2D, 
TexImage3D, and TexSubImage3D

    HALF_FLOAT_OES                0x8D61
    FLOAT                         0x1406

Add a new section called 16-Bit Floating-Point Numbers

"A 16-bit floating-point number has a 1-bit sign (S), a 5-bit
exponent (E), and a 10-bit mantissa (M).  The value of a 16-bit
floating-point number is determined by the following:

    (-1)^S * 0.0,                        if E == 0 and M == 0,
    (-1)^S * 2^-14 * (M / 2^10),         if E == 0 and M != 0,
    (-1)^S * 2^(E-15) * (1 + M/2^10),    if 0 < E < 31,
    (-1)^S * INF,                        if E == 31 and M == 0, or
    NaN,                                 if E == 31 and M != 0,

where

    S = floor((N mod 65536) / 32768),
    E = floor((N mod 32768) / 1024), and
    M = N mod 1024.

Implementations are also allowed to use any of the following
alternative encodings:

    (-1)^S * 0.0,                        if E == 0 and M != 0,
    (-1)^S * 2^(E-15) * (1 + M/2^10),    if E == 31 and M == 0, or
    (-1)^S * 2^(E-15) * (1 + M/2^10),    if E == 31 and M != 0,

Any representable 16-bit floating-point value is legal as input
to a GL command that accepts 16-bit floating-point data.  The
result of providing a value that is not a floating-point number
(such as infinity or NaN) to such a command is unspecified, but
must not lead to GL interruption or termination.  Providing a
denormalized number or negative zero to GL must yield predictable
results."

Add to Table 2.2, p. 12:

            Minimum
   GL Type  Bit Width  Description
   -------  ---------  -----------------------------------
   half     16         Half-precision floating-point value
                       encoded in an unsigned scalar

Add to Table 3.2, p. 62:

    type Parameter  Corresponding  Special
    Token Name      GL Data Type   Interpretation
    --------------  -------------  --------------
    HALF_FLOAT_OES  half           No
                       
Modify first sentence of "Unpacking", p. 62: 

"Data are taken from client memory as a sequence of one of the GL data
types listed in Table 3.2. These elements are..."
                       
Add to Table 3.4, p. 63:

    Format           Type            Bytes per Pixel
    ---------        --------------  ---------------
    RGBA             FLOAT           16
    RGB              FLOAT           12
    LUMINANCE_ALPHA  FLOAT           8
    LUMINANCE        FLOAT           4
    ALPHA            FLOAT           4
    RGBA             HALF_FLOAT_OES  8
    RGB              HALF_FLOAT_OES  6
    LUMINANCE_ALPHA  HALF_FLOAT_OES  4
    LUMINANCE        HALF_FLOAT_OES  2
    ALPHA            HALF_FLOAT_OES  2

Modify fifth paragraph of 3.7.1 Texture Image Specification, p. 67:

"The selected groups are processed as described in section 3.6.2, stopping
after final expansion to RGBA. If the internal format of the texture is 
fixed-point, components are clamped to [0,1]. Otherwise, values are not 
modified."

If OES_texture_3D is not supported, ignore references to TexImage3D and 
TexSubImage3D.

04/29/2005    0.1    Original draft.
06/29/2005    0.2    Added issues on why only NEAREST and 
                     NEAREST_MIPMAP_NEAREST filters are required.
04/21/2006    0.3    Added TexSubImage2D and TexSubImage3D as
                     functions that take the new tokens.
06/14/2006    0.4    CopyTexImage2D, CopyTexSubImag{2D|3D} are
                     not supported.
07/26/2011    0.5    Fix several omissions discovered while writing
                     EXT_color_buffer_half_float.
11/09/2011    0.6    Fix missing FLOAT entries in Table 3.4,
                     add interaction with OES_texture_3D.