scaling issue when using scypy's firwin2 to get coefficients for a filter

I am trying to use SciPy's firwin2 command to get the coefficients that correspond to a filter of my creating. I took white noise in Audacity, applied an Equalization curve, exported said curve out of Audacity, and I am using firwin2 to get the coefficients for said filter. I manage to get coefficients out that give me a correct curve shape but the magnitude of the signal is WAAAAAAY off. It's almost $50$ dB higher than what it should be. Does anyone know when I could be leaving out a scaling factor?

• probably should show your code – endolith Apr 3 '15 at 23:14