NMOS IS-04 requires that the GMID of the PTP reference clock be given with lowercase hex number notation. The corresponding SDP attribute a=ts-refclk has it the other way around, i.e. uppercase hex only. Somewhere a conversion must happen either way, but I don't see where.
Furthermore, nmos-cpp seems to assume that the a=ts-refclk attribute is always in the media section of the SDP file. Indeed that is what ST2110 (stupidly) demands. Other relevant standards don't have this restriction, for example AES67, and indeed the RFC 7273 they all depend on. One would have thought that the normal case is a common reference clock, so that putting it into the global section of the SDP file would be the most prevalent case, if it weren't for ST2110.
So I think nmos-cpp should at least recognise and properly handle either style when receiving an SDP. At the moment, it insists on ST2110 on input, too.
Please tell if you want me to produce a patch.
NMOS IS-04 requires that the GMID of the PTP reference clock be given with lowercase hex number notation. The corresponding SDP attribute
a=ts-refclkhas it the other way around, i.e. uppercase hex only. Somewhere a conversion must happen either way, but I don't see where.Furthermore, nmos-cpp seems to assume that the
a=ts-refclkattribute is always in the media section of the SDP file. Indeed that is what ST2110 (stupidly) demands. Other relevant standards don't have this restriction, for example AES67, and indeed the RFC 7273 they all depend on. One would have thought that the normal case is a common reference clock, so that putting it into the global section of the SDP file would be the most prevalent case, if it weren't for ST2110.So I think nmos-cpp should at least recognise and properly handle either style when receiving an SDP. At the moment, it insists on ST2110 on input, too.
Please tell if you want me to produce a patch.