Author |
Topic: Does Interference exist - or ?? (Read 1195 times) |
|
udippel
Newbie
Posts: 30
|
 |
Does Interference exist - or ??
« on: Mar 12th, 2003, 8:14am » |
 Quote  Modify
|
Since we had the Quantum-Paradox in here and the notorious mirror-inversion, I dare to ask your esteemed opinion on this one. (Not really a riddle, though).
There is this chap ex-MIT CS prof who insists that "Interference" doesn't exist. It is only a phenomenon of instrumentation - instruments that cannot distinguish good and bad - so to say.
David Reed - yes that David Reed wants to let us know that radio photons actually pass through each other like the light photons pass through each other when they at any moment of finite length thrust a complete image onto the rear of the camera obscura.
"Interference cannot be defined as a meaningful concept until a receiver tries to separate the signal. It's the processing that gets confused, and the confusion is highly specific to the particular detector," Reed says. "Interference isn't a fact of nature. It's an artifact of particular technologies."
Actually, in the first instance I was tempted to ask this chap to share his crack with me. But then he got me. I gave myself some 15 minutes to contradict one of his statements after the other, but couldn't.
So, it has became a riddle - somehow ! 
If you think that *I* am the guy having a good smoke, here is the complete brubble:
http://www.salon.com/tech/feature/2003/03/12/spectrum/index.html?x
Hope you forgive my somewhat off-topic riddle ... !
Uwe
|
|
 IP Logged |
|
|
|
towr
wu::riddles Moderator
Uberpuzzler
Some people are average, some are just mean.
Gender: 
Posts: 13730
|
|
Re: Does Interference exist - or ??
« Reply #1 on: Mar 12th, 2003, 9:04am » |
Quote Modify
|
I don't have the feeling this is anything new..
He's claiming that radio interference is mainly a problem of the receiving instrument. Which is true, since a receiver that would resonate less with the frequencies near the signal, and more with the signal itself (in other words greater seperation of signal and noise), would allow for narrower bands, and thus more bands, and thus more information flow.
Knowledge about the signal and the noise can help a lot to further seperate them, you don't have to depend on pure physical properties, but also semantic or syntactic properties (after all we make the signal, so we can put error reduction codes and stuff in).
|
| « Last Edit: Mar 12th, 2003, 9:08am by towr » |
IP Logged |
Wikipedia, Google, Mathworld, Integer sequence DB
|
|
|
wolfgang
Newbie
Posts: 12
|
|
Re: Does Interference exist - or ??
« Reply #2 on: Mar 12th, 2003, 9:41am » |
Quote Modify
|
I think it's being oversimplified. Interference exists, but not in the sense that information is destroyed in the resulting chaos, but in the sense that information can't be retrieved and separated as is (without a really major increase in technology). To make the analogy with light, if you want 2 radio stations using the same color, you would have to invent a lens to focus the signals from one radio tower, and keep looking directly at that tower to "see" its signal without interference from the other tower. Except you can't see the tower usually. The signals bounce and bend before reaching your radio. So the problem might be more like looking at one spot on the wall of your living room and figuring out the size, shape, and color of everything in the room based on all the light that has mixed together and bounced off that wall. Or pinpointing every single car whose engine sounds are echoing off the walls of a big city. Our ears can pick out one voice in a crowded room, cut through the interference, but not if there are too many voices with the same pitch in the same general direction.
I've heard that there were once islanders in the Pacific who could look at the wave patterns around their boat and figure out the location of islands hundreds of miles away based on interference patterns. It perhaps can be done, but to just say interference doesn't exist so let's just treat radios as eyes and allow radio stations to use any frequency they want is a bit oversimplifed.
|
|
IP Logged |
|
|
|
Icarus
wu::riddles Moderator
Uberpuzzler
Boldly going where even angels fear to tread.
Gender:
Posts: 4863
|
|
Re: Does Interference exist - or ??
« Reply #3 on: Mar 12th, 2003, 3:47pm » |
Quote Modify
|
I don't have time right now to answer more fully - but it sounds to me like your smoker is mis-interpreting what "inteference" is. By itself, interference does not destroy information. But when information is destroyed by something else, such as the measuring device, interference determines what information survives. That is a very real phenomenon. Proclaiming it to be only the fault of bad measurement completely overlooks that any measuring device is a physical object, so what it measures is always a real effect. It is only in interpreting what that measurement means that we can make mistakes.
|
|
IP Logged |
"Pi goes on and on and on ...
And e is just as cursed.
I wonder: Which is larger
When their digits are reversed? " - Anonymous
|
|
|
Icarus
wu::riddles Moderator
Uberpuzzler
Boldly going where even angels fear to tread.
Gender:
Posts: 4863
|
|
Re: Does Interference exist - or ??
« Reply #4 on: Mar 12th, 2003, 8:35pm » |
Quote Modify
|
Okay - I finally got the chance to read the article and see what he was saying myself. While he may be right about the advantages of changing our current broadcast paradigm, he is himself committing very bad science when he says that interference is only an artifact of technology.
First of all, as towr points out, the meaning of interference he is refering to is the inability to separate distinct radio signals. This is slightly different than the meaning of the word as it is normally used in physics and which I used in my previous post. But even here his argument has mistakes. The Heisenberg Uncertainty Principle places fundamental limits on how closely you can define the frequency of a signal. The smaller the frequency is, the greater the uncertainty, so the more bandwidth it needs to occupy. At the other end of the spectrum, we run into other problems. Gamma rays do not react well with any sort of matter. The higher the frequency the worse the problem. Eventually, you have no way to create, control, or recieve them, because nothing can withstand the interaction. This effectively limits the overall available bandwidth for transmitting. Since each broadcast frequency requires a finite amount of bandwidth to occupy, there is a limit to the number of broadcasting frequencies available that is independant of any particular technology.
Next, when two signals at the same frequency arrive at the same location, they do pass through each other without loss. But you cannot separate the signals by examing the EM field at a single point. To see this, just consider a point-sized measuring device located at a point where the two waves add together to zero. (If you are familiar with wave behavior, you know that such places exist for waves of the same amplitude.) All measurements taken at the point do not detect any EM activitity at all. There is nothing in the behavior of the EM field at the point that allows you to say that any signals exist at all, much less allow you to sort them out. More generally, to separate the two signals you need to receive information over a region of space. How big a region depends on the wavelength. Since bigger is not better with electronics, this limits the minimum frequency we can use. The Heisenberg Uncertainty priciple also comes into play here, limiting how sure our reciever can be of the frequency he recieves.
There is also a lower limit to how short a light/radio burst can be, which in turn limits how much information can be placed on a single carrier wave.
In short, the are fundamental limits on how much information can be sent by radio/light that have nothing to do with any particular technology.
|
|
IP Logged |
"Pi goes on and on and on ...
And e is just as cursed.
I wonder: Which is larger
When their digits are reversed? " - Anonymous
|
|
|
James Fingas
Uberpuzzler
Gender:
Posts: 949
|
|
Re: Does Interference exist - or ??
« Reply #5 on: Mar 26th, 2003, 1:09pm » |
Quote Modify
|
Icarus,
Continuing your train of thought, there is a law similar to the Heisenberg Uncertainty Principle in communication theory. It says that the time duration and frequency range of a signal are inextricably linked. To make a signal shorter in time, you must increase the bandwidth of the signal. Likewise, to figure out exactly which frequency a signal is at, you have to observe it for a long period of time. A signal that is time-limited cannot also be band-limited (it must have frequency components over an infinite frequency range). In reality, there are no time-limited signals and no band-limited signals.
However, the biggest problem with saying there's no such thing as interference is that it ignores Shannon's limit (far and away the most important result in communication theory). He states that for a given signal/noise ratio (ratio of signal power over noise power) and a given bandwidth, there is an absolute maximum amount of data (he calls it entropy) that you can transmit. Even worse, you get sharply diminishing returns if you try to use more and more power to increase your signal/noise ratio while still using the same bandwidth. You also get diminishing returns, then hit a hard limit, if you try to maintain your data throughput by increasing your transmitted bandwidth while keeping your transmitted power low. Limiting the bandwidth of your signal only helps by allowing you to filter out some of the noise (stuff not in your bandwidth range) before you do your processing, reducing the background noise level.
Therefore, in order to send lots of data, you need lots of bandwidth. What I think Reed is trying to say in this article is that allocating huge blocks of bandwidth to these monopolistic users is inefficient. I would agree with him somewhat, and as cell-phone companies are finding out, you can get a whole lot more data throughput by using many localized low-power transmitters than by using large-area transmitters. However, this is something that will take a long time to develop. The problem with unlicensing the whole radio band is how do you keep somebody from broadcasting all over your localized low-power signals?
|
|
IP Logged |
Doc, I'm addicted to advice! What should I do?
|
|
|
Icarus
wu::riddles Moderator
Uberpuzzler
Boldly going where even angels fear to tread.
Gender:
Posts: 4863
|
|
Re: Does Interference exist - or ??
« Reply #6 on: Mar 26th, 2003, 7:51pm » |
Quote Modify
|
on Mar 26th, 2003, 1:09pm, James Fingas wrote:| Heisenberg Uncertainty Principle in communication theory. It says that the time duration and frequency range of a signal are inextricably linked. To make a signal shorter in time, you must increase the bandwidth of the signal. Likewise, to figure out exactly which frequency a signal is at, you have to observe it for a long period of time. A signal that is time-limited cannot also be band-limited (it must have frequency components over an infinite frequency range). |
|
Actually, that IS the Heisenberg Uncertainty Principle, or rather, this result and the H.U.P. are both applications of the exact same basic principle of waves. H.U.P. is what happens when you apply it to probability waves in quantum mechanics.
Quote:| there is an absolute maximum amount of data (he calls it entropy) |
|
 He calls DATA entropy? A basic result of Statistical mechanics is that entropy measures the LACK of information about a system! I don't doubt the result, but I am curious about this phrasing.
The points you make are what I was trying to say, but did not have the appropriate background to say it correctly. Thanks.
Quote:| What I think Reed is trying to say in this article is that allocating huge blocks of bandwidth to these monopolistic users is inefficient. I would agree with him somewhat |
|
He may indeed have valid points about the need to overhaul our broadcast system, but the way he puts it is very bad science, made worse because he is accusing others of the same thing!
|
|
IP Logged |
"Pi goes on and on and on ...
And e is just as cursed.
I wonder: Which is larger
When their digits are reversed? " - Anonymous
|
|
|
towr
wu::riddles Moderator
Uberpuzzler
Some people are average, some are just mean.
Gender:
Posts: 13730
|
|
Re: Does Interference exist - or ??
« Reply #7 on: Mar 27th, 2003, 12:25am » |
Quote Modify
|
In information theory and communication theory entropy and information (data) are very much linked. The minimum avarage bandwith data takes is defined by the entropy of the data, which is a measure of how much information is needed to transmit the data.
So it's like action and reaction, the size is the same, the direction opposite. And since it's size that matters in this case, they're used interchangeably.. (direction always follows from context)
At least that's what I think I know about it. (based on my prior experience ID3 decision trees (which use information theory) and the first course of machine learning this trimester (which mentioned the link with communication theory))
|
|
IP Logged |
Wikipedia, Google, Mathworld, Integer sequence DB
|
|
|
James Fingas
Uberpuzzler
Gender:
Posts: 949
|
|
Re: Does Interference exist - or ??
« Reply #8 on: Mar 27th, 2003, 2:04pm » |
Quote Modify
|
Icarus,
Yes, the information you're trying to transmit through the system is measured as "entropy". Look at it this way: you transmit more information than you need, so that your signal has a certain amount of redundancy. All the redundant information is defined in the communication protocol. The information you are trying to communicate can't be defined ahead of time, so it's random (as far as you are concerned when you're designing the communication channel).
So the amount of entropy in the channel is the amount of information you send that is not defined in the communication protocol, and is therefore fundamentally unknown and random.
That's how I think of it anyways ...
|
|
IP Logged |
Doc, I'm addicted to advice! What should I do?
|
|
|
|
RIDDLES SITE
WRITE MATH!
Home 
Help 
Search 
Members 
Login 
Register
Reply
Notify of replies
Send Topic
Print