LOL I do build outs every day, 6 days a week.
Your average fade at 800mhz will be between -10 to -20. Your average fade at 1.9ghz will be about -8 to -21. The reason that they are soo close is because of buildings are usually 5 to 8 inches thick, which blocks 800mhz better then 1.9ghz.
When you say you half the frequency and get 4 times better penetration only works in general, but specifics REAL WORLD effects says that anything between 700mhz to 3ghz will have about the same building fade averaged over the whole cell site.
There is always a huge difference between lab and mathematical testings and real world build outs.
Your argument is "trust me I do this for a living". I'm sorry I am newer to this forum and you may truly be an expert in the field and may have proven it to the others already but that's not enough for me. It doesn't help that your supporting material seems to contradict you.
The paper you posted as your evidence definitely seems to be pointing in the opposite direction. The conclusion section seems pretty clear to me:
"4) The attenuation into the buildings increased only moderately with frequency in four of the buildings with a mean slope of about 1 to 3 dB/GHz, but increased fastest in the least attenuating building with a mean slope of 6 dB/GHz. The mean slope was near zero for the glass walled MER building."
This says 5 of the 6 building types tested had a moderate to large increase in attenuation due to building penetration as the frequency increased. Am I missing something? This is a "REAL WORLD" test isn't it?
In my first post I had broken up the losses into two parts. I wasn't as clear as I should have been but by propagation loss I am referring to Free Space Path Loss.
[URL="http://en.wikipedia.org/wiki/Free-space_path_loss"]http://en.wikipedia.org/wiki/Free-space_path_loss[/URL]
This is a loss that a signal will have as it propagates over free space (air too), without any obstacles in the way (i.e. no buildings). This term is where I got the 4x from and is independent of building penetration.
Ideal != real world, I know. building size, materials, height, etc all make a difference in the real world so I can't make a blanket statement. But its not a stretch to say that generally the higher the frequency the worse off you are from a path loss (includes all losses) perspective, and that means given a fixed amount of power and fixed size you signal strength will also be worse off.
Go ask the military why UHF (~300 MHz) communications is still so critical for them. I'll tell you why, its all about the robustness. I don't need a directional antenna that I have to keep properly pointed (not really an issue at L-Band either), and because it doesn't matter if I am in a dense rain-forest with a thick canopy, in an open desert, or in a building chasing after a terrorist they'll be able to connect.