I realise it's not exactly a solution, but have you tried recording sound separately (in like Audacity for example) and then importing into Decker? I've found even when the audio editor is working I get better results this way since I can adjust the levels and such in Audacity before converting down into Decker's format