What the Different Pixel Types Mean

What the Different Pixel Types Mean

Each packed pixel type includes a base type, for example GL_UNSIGNED_BYTE, and a field width (for example, 3_3_2):

The base type, GL_UNSIGNED_BYTE, GL_UNSIGNED_SHORT, or GL_UNSIGNED_INT, determines the type of "container" into which each pixel's color components are packed.
The field widths, 3_3_2, 4_4_4_4, 5_5_5_1, 8_8_8_8, or 10_10_10_2, determine the sizes (in bits) of the fields that contain a pixel's color components. The field widths are matched to the components in the pixel format, in left-to-right order.
For example, if a pixel has the type GL_UNSIGNED_BYTE_3_3_2_EXT and the format GL_RGB, the pixel is contained in an unsigned byte, the red component occupies three bits, the green component occupies three bits, and the blue component occupies two bits.

The fields are packed tightly into their container, with the leftmost field occupying the most-significant bits and the rightmost field occupying the least-significant bits.

Because of this ordering scheme, integer constants (particularly hexadecimal constants) can be used to specify pixel values in a readable and system-independent way. For example, a packed pixel with type GL_UNSIGNED_SHORT_4_4_4_4_EXT, format GL_RGBA, and color components red == 1, green == 2, blue == 3, alpha == 4 has the value 0x1234.

The ordering scheme also allows packed pixel values to be computed with system-independent code. For example, if there are four variables (red, green, blue, alpha) containing the pixel's color component values, a packed pixel of type GL_UNSIGNED_INT_10_10_10_2_EXT and format GL_RGBA can be computed with the following C code:

GLuint pixel, red, green, blue, alpha; 
pixel = (red << 22) | (green << 12) | (blue << 2) | alpha;

While the source code that manipulates packed pixels is identical on both big-endian and little-endian systems, you still need to enable byte swapping when drawing packed pixels that have been written in binary form by a system with different endianness.