Today I released the first version of a noise tolerant DCF77 clock in my blog: http://blog.blinkenlight.net/2012/12/01/dcf77-project/. This is just the first version, more are to come. The next version will be significantly more advanced. However this simple implementation already outperforms all commercial DCF77 clocks that I have around.
Today I released part 1 of the real thing. http://blog.blinkenlight.net/2013/01/01/dcf77-project-part-1-get-ready-for-the-real-thing/. It is just the phase lock for a noise tolerant decoder. Have a look at my analysis ;)
Sounds you are making Serious Work (with capitals) of it. I recall that DC77-decoding was one of my first sketches but never got beyond a "standard" implementation.
I bookmarked the 2 books about signal processing!
Here comes "the second decoder" http://blog.blinkenlight.net/2013/02/01/the-second-decoder/. I have to proceed in small steps as noise tolerant decoding is not as simple as just waiting for some successive decoded values to match. The biggest issue is how to test this code as the error correction tends to correct implementation errors as well ;)