On Sat, 11 Sep 2004 21:16:57 +0500, sherlock@vsnl.com sherlock@vsnl.com wrote:
On Saturday 11 September 2004 20:47, Adil Kodian wrote:
the higher the signal bandwidth, the higher the SNR required to extract info.
actually it is the other way around - you can trade off SNR for bandwidth and vice versa - Higher the bandwidth and you can go with
I said the signal bandwidth, not the channel bandwidth. If your signal bandwidth=channel bandwidth you require a SNR=infinity.
yeah - i dint read carefully enough.
very low SNR. The SNR of the signal from the Mars Rover is just slightly above white noise - so they use a terahertz bandwidth pipe for communication.
Exactly. Trahertz channel bandwidth for sending a few kilobits of data.
netflix teamed up to offer similar service.
to-peer stuff like bittorrent. With newer modulation techniques like
UWB (1Gbps >bw and relatively "simple" recieving frontends) becoming commercially >viable "unlimited" bw for the masses is only an antenna away.
ultrawideband is actually severely limited by distance.
Very true - for now.
also worth mentioning are the regulatory issues with UWB - shouldnt understimate the power of the stupid bureaucracy - which put of FM development by a few decades.
UWB is actually the perfect example of trading off bandwidth for SNR. UWB uses a spectrum that is many Ghz wide (not 1Gbps as data rate is not bandwidth)
and therefore can support an SNR almost at the level of the surrounding noise. One of the features that most uwb proponents talk about (incorrectly) is the anonymity that this low SNR guarantees. UWB is therefore a very good replacement for the 802.11g that is used to solve in-home connectivity issues - but is hardly something that can be used for public broadcast.
Not neccessarily true. For now the power used is the microwatts range. Pumping up the power and frequency into the 25Ghz range should change things dramatically. Moon Bounce communications is already being used by hams in the 25 ~ 50 GHz region. Pure analog ofcourse.
thats the problem - i would imagine that a device that radiates even a few watts of power at 25Ghz would be dangerous and would defeat the purpose of trading bandwidth for SNR. Today people have a problem with cellphones that work at a few hundred megahz and radiate only a few watts of power.
data rates are then defined according to shanon limit on how much information you can send in this slice of the spectrum.
As a general comment - with the population density in India, and the amount of fiber put in the ground by reliance, wouldnt fiber based services be the way to go ? nothing can ever beat the data rate capability of fiber. Fiber can support a spectrum width of about 30Thz and that is only a limit imposed by the switching speed of the laser, and not the fiber. In any case, even with 30 thz, you can theoretically transmit all of mumbai's internet, cable and telephone traffic through a single fiber pair, provided it is fully utilized. so if youve laid fiber all over the city, the last mile problem is the only thing youve got to address with video delivery networks. satellites are really overkill for anything today.
That is precisely where UWB comes in. U dont require everybody's data to go up and down 36km to get things working.
most new broadcast satellites are not geostationary - so its not 36kms but what the hey - youre preaching to the choir - i design fiber optic networks for a living :-) satellite is probably best relegated to security and weather not communication.