IMG_texture_compression_pvrtc2

Name

IMG_texture_compression_pvrtc2

Name Strings

GL_IMG_texture_compression_pvrtc2

Notice

Copyright Imagination Technologies Limited, 2011.

Contact

Ben Bowman, Imagination Technologies (benji 'dot' bowman 'at'
imgtec 'dot' com)
  

Status

Complete

Version

1.0, 19th December 2012

Number

OpenGL ES Extension #140

Dependencies

This extension is written against the OpenGL ES 2.0 Specification.
OpenGL ES 2.0 with OES_texture_npot is required.
  

Overview

This extension provides additional texture compression functionality
specific to Imagination Technologies PowerVR Texture compression format
(called PVRTC2) subject to all the requirements and limitations
described by the OpenGL ES 2.0 specification.

This extension supports 4 and 2 bit per pixel texture compression
formats. Because the compression of PVRTC2 is CPU intensive,
it is not appropriate to carry out compression on the target
platform. Therefore this extension only supports the loading of
compressed texture data.
  

IP Status

Imagination Technologies Proprietary

Issues

1) If this extension does not support driver compression of data,
   how is data compressed?

   Resolution: Textures should be compressed using the
   PVRTexTool available from PowerVR Developer Technology
   (devtech 'at' imgtec 'dot' com)

2) Is sub-texturing supported?

   Resolution: Yes, at block boundaries. This is 4x4 texels for
   the 4bpp format and 8x4 for the 2bpp format. Note it is up to
   the user to ensure the compressor tool is used in the mode which
   removes block edge artefacts if subdata is going to be used for
   eg. a texture atlas.

3) Are non-power of two textures supported?

   Resolution: Yes.

4) How is the imageSize argument calculated for the CompressedTexImage2D
   and CompressedTexSubImage2D functions.

   Resolution: For PVRTC2 4BPP format the imageSize is calculated as:
      ceil(width/4.0) * ceil(height/4.0) * 8.0
   For PVRTC2 2BPP format the imageSize is calculated as:
      ceil(width/8.0) * ceil(height/4.0) * 8.0

5) Note some early 1.9 SGX drivers will return INVALID_VALUE if the width
   or height is not a multiple of the block size.
  

New Procedures and Functions

None.

New Tokens

Accepted by the <internalformat> parameter of CompressedTexImage2D
and the <format> parameter of CompressedTexSubImage2D:

    COMPRESSED_RGBA_PVRTC_2BPPV2_IMG                  0x9137
    COMPRESSED_RGBA_PVRTC_4BPPV2_IMG                  0x9138
  

Additions to Chapter 2 of the OpenGL ES 2.0 Specification (OpenGL Operation)

None.

Additions to Chapter 3 of the OpenGL ES 2.0 Specification (Rasterization)

Add Table 3.8.1:  Specific Compressed Internal Formats

    Compressed Internal Format         Base Internal Format
    ==========================         ====================
    COMPRESSED_RGBA_PVRTC_2BPPV2_IMG         RGBA
    COMPRESSED_RGBA_PVRTC_4BPPV2_IMG         RGBA


Modify Section 3.7.3, Compressed Texture Images

Add to Section 3.7.3, Compressed Texture Images (adding to the end of
the CompressedTexImage section)

If <internalformat> is COMPRESSED_RGBA_PVRTC_2BPPV2_IMG or
COMPRESSED_RGBA_PVRTC_4BPPV2_IMG, the compressed texture is stored using
one of the PVRTC2 compressed texture image formats.  The PVRTC2
texture compression algorithm supports only 2D images without borders.
CompressedTexImage2DARB will produce an INVALID_OPERATION if <border> is
non-zero.

Add to Section 3.7.3, Compressed Texture Images (adding to the end of
the CompressedTexSubImage section)

If the internal format of the texture image being modified is
COMPRESSED_RGBA_PVRTC_2BPPV2_IMG or COMPRESSED_RGBA_PVRTC_4BPPV2_IMG, the
texture is stored using one of the PVRTC2 compressed texture image
formats.  CompressedTexSubImage2D results in an INVALID_OPERATION error
if internal format is COMPRESSED_RGBA_PVRTC_2BPPV2_IMG and one of the
following conditions occurs:

    * <xoffset> is not a multiple of eight.
    * <yoffset> is not a multiple of four.
    * <width> is not a multiple of eight, except when the sum of <width>
      and <xoffset> is equal to TEXTURE_WIDTH.
    * <height> is not a multiple of four, except when the sum of <height>
      and <yoffset> is equal to TEXTURE_HEIGHT.

or if internal format is COMPRESSED_RGBA_PVRTC_4BPPV2_IMG and one of the
following conditions occurs:

    * <xoffset> is not a multiple of four.
    * <yoffset> is not a multiple of four.
    * <width> is not a multiple of four, except when the sum of <width>
      and <xoffset> is equal to TEXTURE_WIDTH.
    * <height> is not a multiple of four, except when the sum of <height>
      and <yoffset> is equal to TEXTURE_HEIGHT.
  

Additions to Chapter 4 of the OpenGL ES 2.0 Specification (Per-Fragment Operations and the Frame Buffer)

None.

Additions to Chapter 5 of the OpenGL ES 2.0 Specification (Special Functions)

None.

Additions to Chapter 6 of the OpenGL ES 2.0 Specification (State and State Requests)

The queries for NUM_COMPRESSED_TEXTURE_FORMATS and
COMPRESSED_TEXTURE_FORMATS include COMPRESSED_RGBA_PVRTC_2BPPV2_IMG
and COMPRESSED_RGBA_PVRTC_4BPPV2_IMG.
  

Errors

INVALID_OPERATION is generated by CompressedTexImage2D if
<internalformat> is COMPRESSED_RGBA_PVRTC_2BPPV2_IMG or
COMPRESSED_RGBA_PVRTC_4BPPV2_IMG, and <border> is not equal to zero.

INVALID_OPERATION is generated by CompressedTexSubImage2D if
INTERNAL_FORMAT is COMPRESSED_RGBA_PVRTC_2BPPV2_IMG and
any of the following apply:

    * <xoffset> is not a multiple of eight.
    * <yoffset> is not a multiple of four.
    * <width> is not a multiple of eight, except when the sum of <width>
      and <xoffset> is equal to TEXTURE_WIDTH.
    * <height> is not a multiple of four, except when the sum of <height>
      and <yoffset> is equal to TEXTURE_HEIGHT.
    * <format> does not match the internal format of the texture image
      being modified.

INVALID_OPERATION is generated by CompressedTexSubImage2D if
INTERNAL_FORMAT is COMPRESSED_RGBA_PVRTC_4BPPV2_IMG and
any of the following apply:

    * <xoffset> is not a multiple of four.
    * <yoffset> is not a multiple of four.
    * <width> is not a multiple of four, except when the sum of <width>
      and <xoffset> is equal to TEXTURE_WIDTH.
    * <height> is not a multiple of four, except when the sum of <height>
      and <yoffset> is equal to TEXTURE_HEIGHT.
    * <format> does not match the internal format of the texture image
      being modified.
  

New State

None.

Revision History

9,  19/12/2012  tjh:  Updated error conditions for subtexturing at the
                        edge of an NPOT texture.
8,  08/08/2012  bcb:  Final tidy up
7,  24/11/2011  bcb:  Added NPOT back.
6,  03/08/2011  bcb:  Added enumerants + further issues
5,  03/08/2011  bcb:  Update from DevTech feedback.
4,  03/08/2011  bcb:  Update issues list from GeorgK feedback.
3,  02/08/2011  bcb:  Update issues list from GrahamC feedback.
2,  01/07/2011  bcb:  Remove NPOT support.
1,  30/06/2011  bcb:  Initial revision.
  