Hacker News new | past | comments | ask | show | jobs | submit
Does it have? You can still decode a signal with negative SNR: https://heungno.net/?p=4373
What I got is that SNR is relative to a time period. If you average the received signal over a longer period of time, the noise tends to cancel out but the signal tends to stack up (if the transmitter is still transmitting the same bit). Traditionally you'd use a narrower filter to remove more noise while keeping the signal, but today's (and last decade's and the decade's before that) fast computers can do part of the averaging in software, which leads to seeing a digital signal come in as noise, and then a signal magically appears after processing. This can also be done with CDMA signals provided that you're locked onto the timing of the code, which is something analog electronics can't do.
Ooo, sorry for the tangent, but this might be my chance to get a real answer for this. Can someone who really understands RF explain to me at what distance the concept behind this site breaks down, for a truly advanced technological species?

https://lightyear.fm

If's fun to think that our Sirius tech cousins at the BBQ under a Texas sized parabolic dish aimed at Austin would be jamming Nelly's "hot in here".

Over distance its about fighting the noise in between the source and the receiver while also fading because of the free space loss, think of a flashlight - not a laser. So nelly volume ticks down while the local stations ramp up..

To keep your car jamming you'd build a growing antenna attached to your ford festiva that as you made your way would compensate for this loss by collecting more signal to focus back to a feed horn, a parabolic - like a larger magnifying glass focusing more ant burning heat in the winter versus the summer.

Very roughly it seems it would be the size of Texas when you arrived at the BBQ, assuming you are traveling the speed of light and left in the early aughts.

You wouldn't hear the song until you hit the break because its the frequency over time that pumps the jam.

> If's fun to think that our Sirius tech cousins at the BBQ under a Texas sized parabolic dish aimed at Austin would be jamming Nelly's "hot in here".

OK, but a giant parabolic dish is some parochial 20th century Earth tech.

I was imagining some little guys who create a 100 cubic AU grid of omnidirectional sensors, with a sensor every 1000km, all hooked up the mother of all DSPs. I can visualize that system identifying some pretty faint waves vibing in the noise. Am I wrong in thinking that this system could pickup AM radio really far away, easily... and once they got sick of that, even FM?

Each of the small detectors need to decern what is noise and what is not. They wouldn't know static from a station and having more clueless detectors wouldn't give you more any information in that regard.

An AU cubic grid of detectors would inform you where a signal originates from by comparing free space loss over the area of the coverage. IF you could discern a station from static.

> Each of the small detectors need to decern what is noise and what is not.

In my un-optimized imaginary system, the sensors are very sensitive and dumb, like me. All the difficult work is done by the central DSP-like brain that can identify even the tiniest of waves moving through the grid.

The utility comes from seeing the relative values in the grid... a pattern of tiny changes in some arc, moving through the grid.

Sure, killer triangulation (actually radial measurement?), but also possibly a decent 500 light-year AM tuner?

That's the game, more sensitive receivers receive more noise too. The game is sending something that will not look like the noise. The longer the sequence, the more likely to decode something but the slower the symbol rate (bandwidth)

Imagine your array popped out SHORT-SHORT-SHORT-LONG-LONG-LONG-SHORT-SHORT-SHORT

You just heard a morse code for SOS! The shorts where detected over 100Mhz (FM) +30db for 1 second each and the longs were 3 seconds each on a carrier that sites at +10db. That's amplitude modulation and that looks like intelligence but unless you knew morse code - it wouldn't make any sense.

The further away you get from the source, the more those decibel spikes weaken and will eventually be no different than the noise floor. Your super computer with a billion ears, only hears ~static~.

Try this, imagine instead if there was no free space loss in the electromagnetic field - we'd wouldn't be hearing humming but SCREAMS from all the noise sources from EVERYWHERE as if we were right next to them, forever. It would impossible to decern anything from anywhere. Communication is defined by its distance because signals have differing origins. Sensitivity, or lack there in, is a feature not a failure.

Who's to say a quasar isn't just a lovely time clock for signals encoded in the noise and we haven't figured out what the breakpoint from noise is yet?

{"deleted":true,"id":42478376,"parent":42468817,"time":1734771616,"type":"comment"}
> Who's to say a quasar isn't just a lovely time clock for signals encoded in the noise

Apart from a) quasars are broad-spectrum, not narrow frequencies (at least I assume that is the case), and b) the power required is too large for a civilization to realistically be able to generate. Not to mention that all that power is overkill for intragalactic communication.

But it’s a good sci-fi idea!

In the spectrum plot you can see a signal, and you can see a noise floor that's below the signal. So.... Yes? Does it not? I'm well aware of negative SNR signals like GNSS, but this doesn't look like that.