I may be mistaken about the descrambling of analog video, I suppose the listener could eventually figure out the repetition rate and parse and frame the signal. The DEW Line radar type signals would certainly be detectable over great distances, and by bandwidth and frequency, immediately recognizeable as artificial.

But the former would quickly descend into the noise by inverse-square law and not carry very far. The latter would only be detectable if the listener just happened to be listening in our direction when the beam swept past him. It's the temporal rather than the spatial consideration that matters. We have been using missile defense radars for only a few decades now. When we start erecting impactor warning systems they may be in operation for just a few centuries. But our sun is 4.6 billion years old, and a transmitting civilization may have been here (as far as they know) for an unknown duration at any time over the last several billion years. The chances they'd be listening just when we'd be transmitting are infinitesmal. And they would certainly know that.

Another issue is that the narrower the beam, the farther away it can be picked up. But it also means the less likely it will be to illuminate a receiver at any one time. Not only that, the slightest deviation in direction would mean they might pick up one quick flash and then nothing as either we or they moved out of alignment (the WOW! Signal?). The constraints on accidental signal interception are not shortcomings of radio technology, they are mostly problems with geometry and picking the right time window to the precise millisecond, over a span of Gigayears.

It is highly unlikely our technology will be detectable (regardless of how good theirs is) unless we carefully design our active program to deliberately optimize its effectiveness. Sure, a nearby civilization could look our way and say "Hey, there's an old stable G subdwarf; let's give it a listen, what have we got to lose." That's essentially what our SETI programs our doing now, which is why I feel they are not likely to get anywhere.

They might very well pick up some of our leakage, and vice versa. But it's extremely unlikely there are any tech civilizations any closer than several hundred to several thousand light years (depending on which numbers you plug into the Drake Equation). In stellar samples that large*, they'd have hundreds of thousand of potential targets to monitor over millions of years to have even the slightest probability of hearing us. And they still have to guess our frequency! No wonder it's a needle in a haystack problem. It only takes a moment to walk down a beach, but you can't examine every grain of sand.

*The rule of thumb is: there are approximately 10,000 stars within a hundred lightyears of earth, and the number of stars goes up as the cube of the radius. Within 200 light years there are approximately 80,000 stars. The galactic disk is about a thousand light years thick at Sol's distance from the nucleus.