Origonal Signal

Over all the system was a success. We were able to isolate and eliminate the frequencies that had the feedback.  Notice, in the spectrogram, that the root frequency of the feedback, located at approximately 3000 Hz was almost completely eliminated.

Click here to listen to the  Original Signal   in comparison to the  Final Signal.

Conclusion

When listening to the edited signal you may have noticed that there is still some audible feedback signal.  However this is a flaw in our model not in the system itself.  This is because a real world situation, if you eliminate the root resonance, then the harmonics also disappear.  Another attribute that cannot be captured in the recording is the idea of adjusting the Eq. levels. Our system cuts out the majority of the frequencies containing feedback, thus leaving only the edge frequencies of the feedback.  These will be so attenuated that they will not feed back like the original feedback signal.  Instead they will disappear once the filtering begins. We were unable to sow this because out feedback signal was recorded and not dynamic. However it is easy to envision how those real world attributes would help the system. 

Lastly, all of the processors, hardware and software used in this model are readily available, thus it is simple to imagine the real life implementation of this system.

---

Home             Back to F.S.S.