I've transferred the latest signal data from Hannover to Berkeley, so we can now look at the waterfall plots for the latest cleaning/scoring run (200 birdies, 100K pixels scored). I've found some anomalies, listed below. I'm posting these mostly for Eric's benefit, but interested SETI@home users are encouraged to study them and to find other anomalies; please post these as replies to this message.

Notes:

• Signals with negative IDs are birdie signals; low-order digits are birdie #.
  
• The top 200-300 multiplets are mostly birdie signals.
  
• You can get textual debugging output as follows:
  
  
  * RFI removal: in a waterfall plot, do custom RFI removal, then click "Show stdout"
  * Scoring: in a pixel page, click "Scoring details".
  

  
  These both produce lots of cryptic output, which you can make sense of by looking at C++ code.
  

If you have ideas for improving the tools, let me know.
--------------------------
Anomalies:
1) multiplets with lots of signals (e.g. non-bary) seem to have excessive (i.e. large negative) scores.
2) top-scoring pixels have excessive scores
3) some birdie signals are isolated, e.g. this
4) triplets in top multiplets look like RFI, e.g. this.
5) drifting RFI removal still not working well in cases like this.

Have you considered utilizing optical processing for signal analysis via Bragg cell technology?

Would allow extremely fast analysis.

---

Never engage stupid people at their level, they then have the home court advantage.....

Also a neural network could also help to detect anomalies.

Just discovering this post now, but I think what this could be ideal for a citizen science project in Zooniverse.

With a more clear sets of instructions about how to do the analysis you can get this done in a second.

---