The protocol that uses was misnamed "Dreambox" by the person here who first reverse engineered it. At the moment I forget its correct name, but "Dreambox" is the best keyword anyway to search for the many other threads where similar signals are discussed.

Some of those threads discuss why these signals rarely learn correctly, and they link to cleaned up codes for many devices using these signals. Your device might use exactly the same signals as one of those, so you may find good signals by testing a few cleaned up codes found through other threads.

I don't have time today for a more detailed examination of that signal. Maybe tomorrow. But if the correct signal is not a perfect match for one discussed in another thread, I probably can't deduce the correct signal from just one sample. I would need learns from several different original keys.