Weekly Progress Report
2000-03-20
1 Current work
  1. Our first project meting has been held and half the projectplan is written. Everyone’s tasks are clear everyone is anxious to start. We have also had our first discussion with Martin Nilsson.
  2. The work with the DSP has started and proceeds slowly. We have been able to sample from the VCR to the DSP. We have also been able to sample the sound from the DSP to the speaker. This means that it’s possible to make sampling from the VCR to Matlab. We are beginning to feel comfortable with the DSP.
  3. We have stated to look at methods of detecting beeps. One method has been found from Proceedings on Detection the "explicit word beep" in the Jerry Springer Show, Stockholm, Sweden, February 17, 2000. It is the last described written by Jonas Rutström. This method looks reasonable but we have to find a method to detect the difference between beep and silence.
  4. We have also look on different methods to replace the beeps with “best” possibly sound. Many theories have been found but few of them are able to handle beeps of varying lengths. Two things is clear:
 
2 Deviation from projectplan
So far there are no great deviation from the projectplan but it’s possible that we are a little bit ahead schedule.

2.1 Measures taken

We have decided that will continue to work hard in the beginning to have some margin in May.

2.2 Suggestion on solution

No suggestions.
 
3 Preceding work
  1. The projectplan will be written and handed over to Martin Nilsson as planned. The web page will be produced. The next project meeting will be held 2000-03-23 cl.09:00 in the laboratory.
  2. We will try to be more familiar to the DSP and implement useful algorithms from books. Also we will look at the possibility to buffer things in the DSP.
  3. We will look at Martin Nilsson method of detecting beeps and start to adjust the most appropriate method to our problem in Matlab.
  4. Next we will try to focus on a few interesting methods on replacing beeps. Also some method to calculate noise variance for the AR-process. We will also start to build a testing environment in Matlab where we can simulate the algorithms without the DSP.
 
4 General comments
We have some future questions that we have discuss later:
 
5 Appendix
 
[Signaler, Sensorer & System] [KTH]