Format of E numbers.

E-number (e.g. E4, E7E4) is named like that for simplicity. Numbers are unsigned.

E stands for 'extensible'.

Number consists of fixed-length blocks, each block contains 'E' flag as elder bit in block,

this bit determines whether there's next block. Next block contains elder bits of number.

Example for E4 (each block has 4 data bits):

value = read 5 bits // where 5 is 4 + extension

result = (value & 0x0F) // retrieve 4 lower bits

if value & 0x10: // check extension flag

value = read 5 bits // where 5 is 4 + extension

result = result | ((value & 0x0F) << 4)

if value & 0x10: // check extension flag

...

Similar is fair for E2, E3, E5, E7... each block has equal amount of bits.

Numbers having several E in format name (e.g. E7E4) have appropriate block sizes and last block is cycled.

Example for E7E4 (first block has 7 data bits, all others have 4):

value = read 8 bits // where 8 is 7 + extension

result = (value & 0x7F) // retrieve 7 lower bits

if value & 0x80: // check extension flag

value = read 5 bits // where 5 is 4 + extension

result = result | ((value & 0x0F) << 7)

if value & 0x10: // check extension flag

... read E4 blocks

That's it.