[Ballot comment]
Thanks for addressing my earlier comments. However, the changes you made to the IP Address definitions in section 4.2 got things a bit mashed up. For example, the way you have it written now sounds like the port number is only in decimal notation if used with IPv4 addresses, which I know was not your intention. The other thing I noticed was that the reference to RFC 5952 is to section 6 (i.e., just the port number discussion). You don't say specifically what the numeric representation (hex or decimal) of the address itself is. I presume you are simply taking the rules from sections 4 and 5, which use hex for everything except in the case of special addresses where the last 32 bits are dotted decimal.

I suggest the following replacements (which also get rid of the superfluous MUSTs):

Destination IP address:port The IP address of the downstream server
and the port number, separated by a single ':'. IPv4 addresses
are represented in "dotted decimal" notation as per [RFC1166].
IPv6 addresses are represented using the hexadecimal notation
detailed in Section 4 of [RFC5952] (or the special-case mixed
hexadecimal and decimal notation detailed in Section 5 of
[RFC5952]) and enclosed in square brackets ('[' and ']').

Source IP address:port The IP address of the upstream client and the
port number over which the SIP message was received, separated
by a single ':'. IPv4 addresses are represented in "dotted
decimal" notation as per [RFC1166]. IPv6 addresses are
represented using the hexadecimal notation detailed in Section 4
of [RFC5952] (or the special-case mixed hexadecimal and decimal
notation detailed in Section 5 of [RFC5952]) and enclosed in
square brackets ('[' and ']').