Why is the maximum bitrate in a channel often twice its bandwidth?

[eetuz]

Hi all,
this question about bandwidth has been bugging me and I haven't been able to find an answer yet.
Im not so familiar with the topic so this might be some glitch in my thinking which I hope you guys could clarify:

Why is max bitrate in a channel two times its bandwidth? For example, if I have two different frequency "levels" in a FM signal for representing 0 and 1, how does the extra bandwidth give me more bits per second? As far as I understand, it enables me to have more signal levels to represent more different bit sequences such as 01 and 10, but how this works in this case where I only have two levels.

And does higher frequency automatically transfer to higher bitrate or is this all just about larger bandwidth in general?

Thanks, eetuz

 

[carlgrace]

The "max bitrate is twice the bandwidth" is a rule of thumb and not a law of nature. The max bitrate in a bandlimited channel is twice the bandwidth only with a non-return-to-zero signal with a specific pulse shape such as raised cosine. If you are modulating the data using Frequency Shift Keying (FSK), which is the method you describe, you channel will need MUCH larger bandwidth than 1/2 the bitrate.

In general, extra bandwidth enables higher bitrate because you can send more pulses. However, if you send coded data you can in fact have a much higher bitrate than the bandwidth.

For example, a 56 kb/s audio modem (from 10 years ago) signals over a telephone channel with a bandwidth of about 8 kHz. This is because the data are coded and not simply 1 or 0.

The real "Max bitrate" is given by Shannon's formula, not the "twice bandwidth" rule of thumb.