This impinges upon a thing that has been "niggling" for a while.
What real bandwidth does a USB signal occupy?
The USB spec has some 50 or 100 pages about the electrical signal in the main 600 page (or 1000 page, I don't remember) spec. The signal is specified via eye diagrams. [Guess why I asked the poster about his opinion of that part of the spec?]
From the eye diagram you should be able to derive the minimum required signal rise and fall time.
I don't mean "pretend" bandwidth in Mbits/sec,I'm talking about what amount of spectrum it would occupy when looked at with a Spectrum Analyser.
You should be able to estimate the spectrum from the signal.
All the references I can find on the 'Net are for IT people,& don't go into the actual implementation of the system.
Just read the original spec. You can download it from usb.org. It consists of some 20 documents, the main spec being 600 pages. And that was one of my points. If you want to know about USB you don't have to wait for a video blog, you can get the spec right now. If you are willing to invest some of your own time, instead of expecting a full service blog.
Everything about USB cables,etc indicate that it isn't really 400MHz wide,
High speed does require 90 Ohm (or was it symmetric 45 Ohm?), and it needs to be good enough to guarantee the required USB signal quality. I don't remember if the spec also gives a bandwidth.
so it would seem that it uses sophisticated modulation methods,
I don't think I saw that in the 2.0 spec. But you can check for yourself.
& isn't just a "Super RS232" sort of thing.
USB is for sure tricker than RS232. One of the reasons embedded programmers still like to stick to RS232 when they can. And if they can't, they like to stick a serial-to-USB chip into the system to avoid having to deal with all of USB.