If a 16-bit value in the range `0...255` is encoded as 1 byte

[KimberlWilliams]

          If a 16-bit value in the range `0...255` is encoded as 1 byte 

and everything else is encoded as 2 bytes.

How does one instead encode the the range -128...127 as 1 byte?
Answer: just add/subtract 128 to/from each value.

I can't decide how to optimize the encoder...
So instead here is an non-optimized version w/delta coding and signed integer support
(that may or may not be correct)
https://gist.github.com/aqrit/ef84d284cebe861d9e57db4129bcafc3

Originally posted by @aqrit in #28 (comment)