A signal's "chirp" is a measure of how fast its frequency is changing during the (short) period of its detection. Chirp is due to acceleration of the sender and receiver: e.g., planetary orbit and rotation. We apply corrections to cancel out our receiver's motion, but the sender may not be doing this.

A few weeks ago I noticed that many high-scoring multiplets had signals with widely different chirp rates over fairly short periods (like tens of minutes). Such signals are unlikely to be from the same source; planets don't move like that.

Another kind of inconsistency involves both chirp rate and frequency. Suppose you have two signals, a minute apart, both with 1Hz/sec chirp rate. You'd expect the 2nd signal to be 60 Hz higher than the 1st. If it's way off from this, chances are they're unrelated.

I modified the multiplet-finding algorithm to produce multiplets that are consistent (within limits) in chirp rate and frequency. It adds another step, "chirp pruning", to the current algorithm. This works as follows. Group the multiplet's signals by time, in periods of an hour or so. Within each period, find the median chirp rate. Discard signals too far away. Adjust the frequencies of the remaining signals to "de-chirp" them. Find the median frequency, and discard signals too far away.

This algorithm seems to work OK. It was a medium code change because the signals in a multiplet are no longer contiguous in time, and I had to change some data structures to reflect this.

-- D

So this will exclude spaceships changing their course and sending messages about that, right? ;)

---

SETI apps news
We're not gonna fight them. We're gonna transcend them.

Presumably, this would also filter out multi-planet systems, sending messages between their planets? The chirp rates would differ for each planet in the system.

---

Presumably, this would also filter out multi-planet systems, sending messages between their planets? The chirp rates would differ for each planet in the system.

In that case the transmissions would be directed, and we'd only hear one of them.