Audio Analyzer with Filter

Here is our scenario: We generate tens of thousands of voicemails on a daily basis. Of these numerous files we get a small percentage of actual "real" human recordings which are the files we really want to hear. The remaining files may be operator intercepts, beeps, ringing, hangup click, tones, etc. We want to have a program developed that will mark the files we want to hear as "Good" and the files we do not want to hear as "Bad". Part of this process will be learned through stored streams as you will discover when reading the documentation. I am attaching a Mockup of the basic design as well as the requirements for the project. Please bid if you feel you have the skills to handle such a unique project.