Cocktail party ideas (2022)
danluu.com
danluu.com
https://news.ycombinator.com/item?id=30185229 - Feb 2022 (110 comments)
"We have frequency modulation, amplitude modulation, phase modulation... But EM waves also have polarization. Can we do polarization modulation as well to transmit more information?"
So, a cocktail party idea.
IMO, this attitude of taking the known facts (in the example, modulation types, EM wave proprerties) and reasoning from them to obtain new conclusions and ideas (polarization modulation) is actually very commendable.
Thhis is how the pioneers first arrived at the ideas, after all, plain and simple. So we need this behavior to move forward.
Often, the only problem with the idea are missing facts: IIRC one of the answers why it didn't work was
"Polarization information gets lost in transmission due to multiple reflections"
So it's not like the layman could have foreseen this. Their theory is at least "self-consistent". And then, the only difference between them and the pioneer is that the pioneer came before, tried it, and found out it didn't work already.
But now consider:
"Big telecom is so dumb. Why don't they just do polarization modulation? Instant 2x internet speed"
The difference that makes this one bad IMO is the thought that, even though we've been doing radio for 100 years, nobody has thought of this idea before, that you came up with after taking EM 101.
It means arrogance, but IMO it also means ignorance: I like to think that having learned one field to some detail has given me appreciation for the hidden details other fields probably have...
P.S.: could not find the EE.SX question link. These are not exact quotes, I edited a little for rhetorical effect.
But let's engage it. I'm no RF expert so just cocktail party spitballing here. I'm not 100% convinced about the multiple reflection thing. I mean, couldn't you just do channel estimation with pilot waves or something like they do for AM? Also makes it work for mobile communications since random user antenna orientation for polarization is analogous to random user distance from transmitter. Also what about fixed links like parabolic dishes and geostationary satellites? Maybe the problem is that it's hard for the transmitter? I mean you can't just have 1 fixed amplifier, that does nothing but amplify, driving 1 fixed antenna if you want to change polarization right? I mean you'd have to have an electronically controlled antenna array or maybe have two amplifiers that ate driven in different ways. Either way this is fundamentally different from the basic transmitter architecture I have in my head where all modulation is done in baseband and then you just upconvert, filter, amplify.