What is the ETC2 compressed texture block size?

Question

What is the size of the texture block GL_COMPRESSED_RGB8_ETC2 and GL_COMPRESSED_RGBA8_ETC2_EAC ?

I used ((w+3)/4)*((h+3)/4)* 8 for GL_ETC1_RGB8_OES , but can't find anything about ETC2 (Khronos documentation is not very clear).

Answer 1

The OpenGL ES 3.0 specification includes the instruction: A texture compressed using any of the ETC texture image formats is described as a number of 4 x 4 pixel blocks in the C.1 ETC Compressed Texture Image Formats section.

Answer 2

FWIW ETC2 extends ETC1 by using some otherwise unused bit patterns, as described by Strom et al in "Compressing Textures Using Invalid Combinations ". (See Also Graphic Equipment Slide Presentation ):

Therefore, the block sizes are the same.

Answer 3

Block sizes are described here: http://www.khronos.org/opengles/sdk/docs/man3/html/glCompressedTexImage2D.xhtml

In particular:

• GL_COMPRESSED_RGB8_ETC2 = ceil (width / 4) * ceil (height / 4) * 8
• GL_COMPRESSED_RGBA8_ETC2_EAC = ceil (width / 4) * ceil (height / 4) * 16

Answer 4

  // ETC1 { 4, 4, 8, COMPRESSED_ETC1_RGB8_OES }, // ETC2 / EAC { 4, 4, 8, COMPRESSED_R11_EAC }, { 4, 4, 8, COMPRESSED_SIGNED_R11_EAC }, { 4, 4, 16, COMPRESSED_RG11_EAC }, { 4, 4, 16, COMPRESSED_SIGNED_RG11_EAC }, { 4, 4, 8, COMPRESSED_RGB_ETC2 }, { 4, 4, 8, COMPRESSED_SRGB8_ETC2 }, { 4, 4, 8, COMPRESSED_RGB8_PUNCHTHROUGH_ALPHA1_ETC2 }, { 4, 4, 8, COMPRESSED_SRGB8_PUNCHTHROUGH_ALPHA1_ETC2 }, { 4, 4, 16, COMPRESSED_RGBA8_ETC2_EAC }, { 4, 4, 16, COMPRESSED_SRGB8_ALPHA8_ETC2_EAC }, 

The first two values ​​are the block size (for example, 4x4), the third value is BytesPerBlock. The fourth value is compression mode. This is from the table that I use to handle all compressed formats that exist.

I removed values ​​that are not useful for this answer (compression and decompression function pointers and preferred pixel formats for source / target data, sRGB, etc.).