Common hex lookup for pair of characters

[xtonik]

Common hexadecimal lookup for conversion pair of characters from text representation to number is about 15% faster for both byte and char variants.

The gain was probably achieved by reduced number of conversion from / to int / byte and preshifting values for higher digit. As highest bit is used for detection bad characters, preshifted values must stored in array of shorts not to interference with the detection - the array occupies 256 bytes more.

I hope this pull request is not faulty as similar #70.