Friday, January 04, 2008
What I'd be looking for in order to find ETI.
If I had the equipment and programming knowledge, this is what I'd be looking for:
- A patch of noise similar to the rest of the background noise, but a noticible hair stronger in signal than the rest for a given frequency range. A little plateau of sorts in the background noise spectrum.
- This noise bump is likely to exhibit some sign of doppler shift in frequency over time.
- Natural white noise pops and clicks are fairly random in spread, and fairly random in timing. When looking at this particular patch of "louder white noise" - this would be part of the "Aha!" factor: this clump should exhibit an uncanny syncronicity of pops and clicks in the given spread when viewed at fine enough resolution on the time scale. That means large groupings of pulses of signal vs no signal should get all their ducks in a row on a given time beat.
- Do some signal processing to filter out the synchronized pops and clicks (those that exhibit an obvious time beat pattern) from the rest.
- Then look for obvious signs of digital encoding over the spread of pops and clicks.
- Try to figure out if any signal clues are embedded. If it's a "hello" signal, there should be a decryption key built into the data stream somewhere.