Bandwidth of a channel

 Bandwidth of a signal is the difference between the upper and the lower frequencies. What do by mean by bandwidth of a channel?

Answer 1

[Walter Roberson]

Generally speaking, bandwidth of a channel is the maximum information carrying capacity of the channel.

A "signal" might have to be encoded or altered in order to carry it over a "channel". For example, you might start with an analog signal that represents human speech, truncate it at 8000 samples per second, process that 8 KHz signal through a processor at (say) 100 MHz, and then multiplex it with a whole bunch of other speech signals to transmit it over a 1 Gigabit/second fibre optic cable, possibly with error correction codes added. As there might be multiple layers of multiplexing along the way, there might be multiple layers of "channels" whose capacity could be measured, so when you see a figure such as a "56 Kbit/s channel", you should not (without further information) assume that it is a physically distinct channel of that speed, just that it is somehow getting sent through a process that promises that much transmission bandwidth for that logical channel.

[i Venky]

In the Shannon-Hartley theorem for channel capacity

C = B log2(1+S/N)

what is the term "B" then?

[i Venky]

Wikipedia says "B" is the bandwidth of the channel and I understand that it is related to the channel capacity but what is "B" as channel capacity is the maximum information that the channel can carry.(which you defined for B)

Thanks in advance.