[link] A Radio Engineer's Essay on Why SETI Won't Work

[Junghalli]

An interesting essay I found on the web. Especially interesting is his claim that we wouldn't be able to detect digital signals. I've bolded the relevant parts. If this is true it sounds like it might be close to impossible to detect an alien civilization's internal radio chatter even with very powerful future radio telescopes. I don't really have the knowledge to evaluate this claim, unfortunately, any technical people here who'd care to weigh in?

http://www.bidstrup.com/seti.htm

-------------------------------------------------------------------------

WHY SETI IS DOOMED TO FAIL: A RADIO ENGINEER'S PERSPECTIVE ON THE PROJECT AND ITS PROSPECTS

Let me state right at the outset that I really, really want SETI, the Search for Extra-Terrestrial Intelligence, to succeed. The announcement that we are now known to not be alone, would be, in my opinion, one of the greatest news stories in the history of the human race, certainly the greatest news story of my life. And I really look forward to the day of that announcement, even if I don't hold out much hope that it will happen in my lifetime.

I hate to throw cold water on the various SETI projects and their enthusiastic supporters, because there are few people on the planet that are more desirous than I of knowing that we are not alone, that the potential for "contact" exists, and that we can finally discredit all the religious nonsense about "creationism." But if we are going to look at the SETI project's actual, real-world chances for success, we have to look at the basics of radio communications theory as applied to SETI, and see how that theory affects the chances for success. As a radio communications engineer with a career-long background in designing and implementing radio communications circuits (which is basically what SETI hopes to achieve), I come to this question with more than a few qualifications, and the basic understanding of what is needed for SETI to succeed - and what I know isn't offering much hope.

Lets begin by examining the basics, for those of you with a non-technical background. For a radio circuit to be established, it is necessary to generate a signal, transmit that signal through space, receive that signal in a reciever sensitive enough to perceive that signal amidst the "noise" (self-generated interference) that every receiver generates, or the radio noise that comes from natural cosmic sources (or, in some cases, man-made noise). That noise has to be overcome by making the signal strong enough to overpower the inherent noise of the receiver, which is the result of the laws of physics and can be minimized to some degree, but not eliminated. Additionally, the farther the signal travels through space, the more of it is lost through spreading out - a problem called "path loss." It is why distant headlights are more faint than close ones. So a communications circuit, whether man made or one being conducted with extraterrestrials, has to overcome these two problems - receiver noise and path loss.

There are several ways to overcome these problems. First, raise the transmitter power to a level sufficiently strong that the receiver's noise and signal spreading losses are simply overwhelmed. This is the method used by television and radio broadcast transmitters, which can operate at truly huge power levels - sometimes as much as a million watts or more. But generating this much radio frequency energy is very expensive, not just for the huge and complex radio transmitters involved, but for the sheer amount of electricity consumed, and so this method is employed only when the other methods cannot be used.

The second method is to use a "gain" antenna at the transmitter - an antenna that concentrates the signal it is radiating into a beam, much like the beam of a flashlight. In fact, this is a good analogy, because a flashlight is doing precisely that - concentrating radio waves (in the form of light which is a type of radio wave), by means of a concentrating reflector, into a beam consisting of nearly all of the light generated by the flashlight bulb. Microwave and satellite communication dish antennas do precisely the same thing - by concentrating all the energy into a beam, the signal inside the beam is made much stronger than it would be if the energy were simply permitted to escape in all directions. If you are transmitting the signal to a single point, or a variety of locations that are very close together as seen from the transmit antenna, this enables you to create a much stronger signal at the receive antenna than you otherwise could for a given amount of transmitter power. The ratio of how strong the signal is in the antenna beam to how strong it would be if it were radiating equally in all directions, is said to be the "gain" of the antenna.

The third method is employed at the receiver. It is to use a "gain" antenna to collect a large amount of weak energy coming from a certain direction and use (usually) a parabolic reflector to concentrate that signal in a small area, where it can be efficiently collected and utilized by the radio antenna. This is how a reflecting telescope works, and it is why making the telescope larger makes it more sensitive. This method is exactly the opposite as the second method explained above, but it is also compatible with the second method and can be used along with it. In fact, the use of methods two and three together is the normal method of setting up microwave communications links - the amount of transmitter power required for a link extending many miles can be so low that the power generated by the transmitter can be roughly the same as the amount of power used by the indicator light on your stereo that tells you your stereo is turned on. Used together, methods two and three are very effective indeed at reducing the amount of power required for a successful communications link.

The fourth method is to narrow the "bandwidth" of the signal. By restricting the signal to a narrow range of frequencies (or limiting the number of channels we are listening to), the receiver can filter out much of its own noise by employing a filter that simply removes all the noise contained in frequencies that are not of interest. Hence, the narrower the range of frequencies, the more noise can be excluded, and the more favorable is the ratio of signal to noise. Radio receivers can be made very sensitive by this method, but the problem is that the more information a signal carries, the wider the range of freqencies of which it consists, and that limits how much filtering can be done - filter out too much, and you filter out the information the link is carrying.

Which brings us to SETI. Since we don't know what power levels are being used by the hypothetical extraterrestrial civilization for its internal communications, nor can we control the gain of the antenna they may be using for transmitting their power, we have only two options - increasing our own receiver antenna gain (method three), and reducing the bandwidth of the receiver we are employing (method four). Both methods are employed by SETI. The antennas are as big as the financial resources available to the SETI projects permit. A trick of technology enables SETI to use high-power computing to examine millions, even billions, of single channels, one at a time, filtering out extraneous noise, though this is done by mathematical calculations, and is therefore not done in real time. This accomplishes the filtering out of unwanted noise, as is explained in method four above, but there are practical limits as to how much this technique can be applied. If you have a "SETI at home" screen saver running on your computer, this is what it is actually doing.

Because the distant solar systems that may be inhabited by SETI-candidate civilizations are really, really far away, the path loss between them and us is really, really big. We are, in effect, trying to see headlights on a car that is millions of miles distant. So for a SETI receiver to successfully detect these signals, the receivers have to be really, really sensitive (and there are limits imposed by the laws of physics on how sensitive a receiver can be made to be), the bandwidth being searched has to be really, really small, or the antennas have to be really, really big - or better yet, all three of these techniques applied at once. We have to assume that it is extremely unlikely that an extraterrestrial civilization will be aiming its signal at us with the specific intent that we hear it, and that we will be aiming our very narrow-beam antennas in just the right direction, and tuning our receivers to just the right frequency all at just the right time in order for us to detect the signal. Therefore, for us to hear an alien civilization, we are going to have to get extremely lucky. Recognition of this reality is why so much effort is being expended by the SETI projects to increase the number of antennas and receivers being used for the search.

We also have to assume that extraterrestrial civilizations are going to use signals similar to the kinds we know about and use ourselves, since we cannot otherwise predict what signal charactaristics will be, and we would not know what to look for. But if they are even a few decades of years more advanced than ourselves, then they are probably going to use signals of a nature that we cannot understand with our current limited technology. Now remember, we have been using radio communications for only about a century, and the principal method we have employed up until now has been "analog" communications. That means varying the strength or frequency of a signal to represent the information we wish the link to carry. That's a great system and it works well, but analog methods have some serious limitations, including fidelity of signal quality. But until recently, it has been all we have, and so we have always assumed it is all that other civilizations have too, and that is why SETI has been looking for signals that are AM (amplitude modulated - carrying information by varying the strength of the signal) or FM (frequency modulated - carrying information by varying the frequency of the signal). That's fine, but an extraterrestrial civilization is not likely to use AM or FM for any period of time much longer than we have. There is a new modulation technique that we have recently developed (mathematically, it is related to both AM and FM, but the resulting signal is so unlike either that it can for all intents and purposes be considered a new method). It is called "digital." Using a variety of techniques that result in both strenth and frequency shifts at the same time, the new digital methods produce a signal that, to an incompatible receiver, is totally indistinguishable from that receiver's own internally-generated noise.

Digital communications techniques offer such overwhelming advantages over analog methods, that it is extremely unlikely that a civilization that knows about digital would use analog methods for very long - in fact, our own civilization is converting very rapidly. Much of the world has already converted to digital radio and television broadcasting, and the last major industrial holdout, the United States, has begun the process. The vast majority of all satellite communications signals are now digital. So any civilization that has been using radio waves for communications for more than about a century is likely to be using digital communications techniques - and producing signals that are indistinguishable from noise, since we don't know how to build compatible receivers. So SETI is going to have a hard time distinguishing these signals from noise sources.

And for SETI, that is the rub. How can SETI detect a digital signal, when it doesn't know what to look for? In terms of digital signals, it can't, because unless a receiver is designed specifically to be compatible with the transmitted signal, the signal will appear to be ordinary radio noise.

There is yet another gotcha. One of these new digital techniques that has been developed in recent decades is a unique system called "direct sequence spread spectrum." Without going into a lot of physics here, let it be said simply that it is a method of adding what amounts to a fifth technique for overcoming spreading loss and receiver noise. Instead of narrowing the receiver bandwidth and excluding noise by filtering it out as is done in method four above, this technique uses a unique mathematical algorithm to do just the opposite - by spreading the signal out, and then mathematically "collapsing" the spread-out signal at the receiver, the receiver noise is also mathematically "collapsed" - and the signal is recovered, even though it may be much weaker than the noise - orders of magnitude in some cases. In order for us to "collapse the noise" and recover the signal, we need to know the precise spreading sequence code - and there is essentially an infinite number of possible codes to start searching through. Use of the wrong code results in just more noise. So even though a casual tuning search across the radio dial may show no signal (because it is hidden by the noise), a usable signal may actually be present (this is why this technique is of great interest to intelligence and military services). The signal is not only not decipherable by incompatible receivers, but isn't even detectable by them. So if our extraterrestrial civilizaton is using a direct-sequence spread spectrum signal, we won't even know the signal is there, much less be able to decipher it. And even, if through some good fortune, we are able to detect the presence of a signal, it will surely look like just plain noise to our receivers, and we are likely to dismiss it as yet another natural galactic noise source, such as the many natural microwave noise sources known to exist in the cosmos.

It gets worse. Remember where I talked above about the use of "beam" antennas to direct radio signals into narrow beams? Well, these beams can be very narrow indeed. Satellite transmitting earth stations use beams that are so narrow, that by the time the signal has traveled from the earth station to the satellite 22,300 miles away, the signal has spread out into a beam only about 150 miles wide for relatively small earth stations, and for the really big dishes used by international earth stations, that beam can be as narrow as 10 miles. The aiming of such antennas has to be very, very precise for the satellite to hear the signal, and for the ground station to hear the weak, solar-powered signal from the satellite. While indeed, the signal from the ground station is very concentrated (these are the strongest signals produced by planet earth's civilizations), the fact that they are concentrated into such narrow beams represents a problem. The chance that such a beam, generated by an extraterrestrial civilization would happened to be aimed at earth, is vanishingly small. If one imagines a sphere (representing all the possible directions to earth) with a radius of 26,300 miles (the radius from the center of the earth to the distance of the satellite), that represents a sphere with a surface area of 2,176,000,000 square miles. When that sphere is illuminated by a beam from our transmitting satellite earth station, only 78 square miles of that surface is lit up by a ten-mile wide beam - that represents the chances that the dish will happen by chance to be aimed in our direction. In other words, the chances of that such a dish being aimed in our direction randomly are represented by the fraction 78/2,176,000,000. Or just once chance in about 27,000,000. Pretty small.

And that's not all the bad news. It is highly unlikely that our hypothetical civilization would have evolved on a planet that is not rotating with respect to the galaxy as a whole. Taking the only case that we know about for sure - our own, if we assume that our aliens are on a planet rotating at a rate of once every 24 hours, that means the tiny, narrow beam isn't illuminating a given direction in space for very long, it is sweeping around in a circle once roughly every 24 hours with respect to its star. So how long is such a dish on a distant planet going to be aimed in our direction? With respect to the galaxy, our planet rotates once every 23 hours, 56 minutes and 10 seconds. If the satellite dish is located on the equator and is aimed at a satellite precisely above the equator (nearly all communications satellites are), the antenna will be sweeping though the Clarke belt, where the satellites are located. Since the Clarke belt is 82,580 miles in circumference, and only ten miles of that is illuminated by our ten mile wide satellite earth station antenna beam at any one time, that means that during the day, any point behind that Clarke belt will be illuminted by our antenna for only 10/82,580 of the day, or about 10.4 seconds per day - barely enough time for our SETI researchers back here on Earth to confirm that the signal is from an extraterrestrial source, even if we notice the signal immediately when it appears. So even if, through some tiny chance we happen to have our dish aimed in the right direction and our receiver tuned to the right frequency, and we're listening for the right kind of emission, we've got to be extremely fortunate to be listening at just the right time. So if we happen to aim our dish at a candidate star and have a listen, even if we know the frequency and emission type, the chances of hearing anything are one in 27,000,000 (the chances of the alien antenna being aimed in our direction) times 8258, or the chances of it being aimed at us while we happen to check for a signal. That's one chance in 223 billion. And that doesn't even consider the likelihood of our listening on the right frequency for the right kind of signal.

Not all of this is lost in the SETI people. To compensate for some of these unfortunate numbers, the SETI approach has been to assume that the distant civilization is reaching out to us and transmitting, hoping that we'll hear - and maybe respond. Yet these same problems work against that theory. The path loss to another star is truly enormous - so to make up for it, the distant civilization has to be transmitting deliberately in our direction, using a high gain, very directional antenna, and maintains its pointing in our direction.

If our experience is typical, for any given alien civilization out there, we've got only about a hundred year window from the time they discover radio and are using modulation techniques that generate signals of a type we can detect, to the time they start using digital signals that to us are indistinguishable from noise. During that brief time, they have to use high-powered microwave transmitters and huge parabolic antennas that are sufficient to generate signals concentrated enough for us to detect them, and that means the chances of their transmitting in the right direction are one in tens of millions; we have to be listening at just the right time (during a window of a few seconds per day), a chance of one in thousands, and on the right frequency out of an infinity of frequencies, and our listening in just the right direction for just the right type of emission.

But are they transmitting in our direction deliberately? The possibility exists that another civilization may be attempting to attract our attention, and doing so with a beacon signal aimed our direction.

Why would they suppose that this very ordinary star of ours, with its very ordinary solar system is any different than most, in that it harbors a supposedly intelligent civilization? Our solar system does not really stand out in any way that would presuppose they would think it a likely candidate. A far more likely candidate would be a solar system with a Jupiter-like planet orbiting about the same distance from the sun as us, a planet with a large family of moons, one or more of which is likely to be the right size to be earthlike. That lets us out. Our solar system is more unlikely than most - an earth-sized body too close to its star to be habitable (Venus), and another body at just the right distance, but too small to retain an atmosphere that is thick enough to make it hospitable or even habitable (Mars). The planet in the middle between them has a large moon, meaning it has been the object of a collision with a sizeable planet in the distant past, and that reduces the likelihood of its habitability. Likely to be a desert planet, with a thin atmosphere, bombarded by lots of meteorites, enough to make a civilization unlikely. Guess we'd best pass that solar system up and aim our beacon beam elsewhere towards a more likely candidate solar system.

And how likely is it that a distant civilization would actually be transmitting a beacon, anyway? What if they are afraid that any attention they attract might be unfriendly? So they think the better of it and keep their electronic mouths shut, just to be safe?

What if they consider the likelihood of success to be sufficiently small as to not be worth the effort - the same reason that we are not transmitting a beacon? There are many candidate solar systems out there, but Earth is not transmitting a regular beacon signal in the direction of even one of them. Why should we suppose that they would be any different?

These are all numbers that are difficult to quantify. But they don't add up to a distant galactic civilization reaching out to our little rock. Even adding the tiny possibility of a deliberate beacon signal being sent our direction specifically to attract our attention, adding that to the numerator in our fraction of the likelihood of a successful SETI project, I am still not overly hopeful.

ADDING IT ALL UP

The chances of that all coming together? One in many trillions. Or, add in the possibility of us observing a beacon, one in hundreds of billions. So don't hold your breath. Building an optical interferometer telescope system on the moon that would enable us to directly image the planets we are discovering which are circling other, distant stars offers us much more hope of actually finding something out there. A reasonable effort could enable us to search a volume within about two hundred light-years of earth, which would offer us many thousands of candidate solar systems to look at. It would also offer many other exciting and potentially useful discoveries about the cosmos as well. It is where our space exploration efforts should be directed, in my opinion, and it would give us a legitimate reason for beginning the permanent colonization of space, the last, unending frontier worthy of the human species. But, alas, there is no political will even for this modest effort as long as there are empires for politicians to seize upon earth.

My hunch is that those bottles of champaign that are on the back shelves of the refrigerators at all the SETI sites are going to be sitting there for a really long time. I really wish it weren't so - no one would be more pleased for a big discovery announcement than I. But the chances are overwhelmingly against a big celebration anytime soon. The numbers just don't add up.

[Singular Intellect]

The whole arguement that intelligent extraterrestrial life would have advanced far beyond our level of long distance communication is certainly not new.

I recall not too long ago myself pointing out it's the equivalent of primitive tribes that use drums to communicate over long distances, and thinking that larger more advanced civilizations would use really big drums to communicate over even longer distances. Therefore if interested in validating the existence of these other civilizations, they listen for the drums of these other civilizations, while being utterly oblivious to radio wave communications all around them.

[Darth Wong]

If we really wanted to contact a far-off civilization, it would really help to know where it was in advance. Then we could point a powerful laser in their direction and blink it on and off in a pattern of prime numbers. That would pretty much prove to them that we're intelligent, because they'd quickly figure out that we're not just a really distant pulsar.

Unfortunately, we don't have that information.

[Junghalli]

Digital communication is hardly "far advanced beyond our level", but he does seem to think it would be a lot harder to detect.

I'd be interested in hearing the opinions of any people with technical knowledge about radio on this.

[Junghalli]

If we really wanted to contact a far-off civilization, it would really help to know where it was in advance.

I think he was more adressing the idea of using SETI to pick up leakage from another civilization's internal communications.

[Darth Wong]

If we really wanted to contact a far-off civilization, it would really help to know where it was in advance.

I think he was more adressing the idea of using SETI to pick up leakage from another civilization's internal communications.

It seems like a silly idea from the outset: given the staggering distances involved, the inverse square law becomes an insurmountable problem.

[SpacedTeddyBear]

The bolded part of his arguement assumes that any possible alien civilization hadn't already began to search and send signals into space using radio waves before switching from analogue to PCM signals.

Radio signals are pretty robust when it comes to physical interference because of their long wavelengths. It's when you send them through areas with strong AC is when you get signal degradation. Of course this is all pretty trivial given the fact that you have the vastness of space between star systems which can also contain large planetoids or charged particles.

[Adrian Laguna]

I think he was more adressing the idea of using SETI to pick up leakage from another civilization's internal communications.

It seems like a silly idea from the outset: given the staggering distances involved, the inverse square law becomes an insurmountable problem.

I think the math shows that our own transmissions become nonsense indistinguishable from the background noise after a couple of light years. In other words, even the nearest star (Alpha Centauri, about 4 light-years away) doesn't get anything coherent.

[Gerald Tarrant]

The bolded part of his arguement assumes that any possible alien civilization hadn't already began to search and send signals into space using radio waves before switching from analogue to PCM signals.
...

From the article

Digital communications techniques offer such overwhelming advantages over analog methods, that it is extremely unlikely that a civilization that knows about digital would use analog methods for very long - in fact, our own civilization is converting very rapidly. Much of the world has already converted to digital radio and television broadcasting, and the last major industrial holdout, the United States, has begun the process. The vast majority of all satellite communications signals are now digital. So any civilization that has been using radio waves for communications for more than about a century is likely to be using digital communications techniques - and producing signals that are indistinguishable from noise, since we don't know how to build compatible receivers. So SETI is going to have a hard time distinguishing these signals from noise sources.

The problem here is that there's no reason to expect those hundred odd years of analog communications to coincide with when we're listening; the odds are kind of against that actually. For all we know E.T's analog radio phase coincided with the Jurassic period, and reached Earth when we were still figuring out stone axes.

[sketerpot]

Digital communication is hardly "far advanced beyond our level", but he does seem to think it would be a lot harder to detect.

I'd be interested in hearing the opinions of any people with technical knowledge about radio on this.

What the article said makes sense. DSSS produces something very similar to white noise, except that you can make sense of it if you know the random number generation algorithm and the random seed and the starting time. It's similar (but I think less severe) for frequency-hopping spread spectrum.

Mind you, I'm not a communications engineer. I've taken a few signals classes, but it's possible that I'm making a mistake -- for example, maybe the aliens' DSSS uses a really crappy pseudorandom number generator and we could detect the lack of sufficient randomness in the signals. But I doubt it.

[Wyrm]

The real weakness of SETI is that it assumes that advanced civilizations would transmit unambiguously-patterned signals long enough and with enough power for us to detect. Basically, it requires such a civilization to return the favor of letting other, less fortunate civs the message that they are not alone. This has a starting problem, because, what if the civilization thinks that such a General Announcement to Alien Intelligences is futile because it's own search turned up nothing?

[Junghalli]

I think the math shows that our own transmissions become nonsense indistinguishable from the background noise after a couple of light years. In other words, even the nearest star (Alpha Centauri, about 4 light-years away) doesn't get anything coherent.

Current SETI is for the most part pretty short-ranged by galactic standards but it's not quite that bad.

If within around 50 light years of us they could, given sufficiently large antenna arrays, have detected the "leakage radiation " from the Earth such as Radars and TV signals. But the TV signals would be too weak for them to produce images.

We don't put much power into our TV or radio transmissions because their intended audience is not very far away. Our defense radars are much stronger and the current Phoenix search could detect that kind of signal out to the limit of our current search which is 155 light years. The strongest signal generated by our technology is the planetary radar on the Arecibo radio telescope and we could detect its signal half way to the center of the galaxy.

From here. I suspect in a lot of these cases we'd have to get very lucky and be in the path of the directional beam though. Especially the last one.