Lyndel,

The Streamzap protocol is similar to RC5, but isn't the same as RC5. I'm pretty sure there is no way to represent Streamzap signals in the "5000" form of "clean" RC5 signals.

Your learned signal above is clean in the sense that it has no timing glitches or extra pulses or other problems common to dirty learns. The Streamzap protocol does have a toggle bit, which in the learned for will never toggle. I'm not sure how much of a problem that will be.

There are ways to represent signals other than RC5 which include toggle bits. I don't know enough about them to tell you whether or how they might encode Streamzap.

I do have details on the structure of the Streamzap protocol, so my newer version of ccf2efc (in the JP1-KM group) decodes it, and I could make a control file for MakeHex to generate Pronto Hex for it (which still wouldn't do anything for the toggle bit).