Why do we consider zero mean for gaussian noise in a communication channel?

AI Thread Summary
Gaussian noise is typically considered to have a zero mean because a DC component would imply a constant power drain from the source, leading to eventual failure of the communication system. In practical applications, channels are often AC coupled to eliminate low-frequency DC levels, which are not relevant to most signaling rates. The presence of a DC component can complicate the analysis and design of communication systems, making zero-mean noise a more manageable assumption. Additionally, the term "matter of semantics" highlights that different interpretations of language can lead to varying understandings of concepts in noise analysis. Overall, zero-mean Gaussian noise simplifies communication channel modeling and is a standard assumption in the field.
dexterdev
Messages
194
Reaction score
1
Hi all,
I got an idea why noise is having a gaussian pdf but didnot understood why it should be having zero mean. Why can't noise contain a dc content.

-Devanand T
 
Engineering news on Phys.org
in most cases we consider 0 mean awgn , is it so...
 
dexterdev said:
Why can't noise contain a dc content.

DC component means that there is energy (power) permanently leaving the source, i.e. the source is "evaporating" so it would eventually seize to exist.
 
FailedLaunch said:
DC component means that there is energy (power) permanently leaving the source, i.e. the source is "evaporating" so it would eventually seize to exist.

Can you explain in simple english (Iam not good in english). I didnot got the idea of evaporate and seize etc..
 
Taking it back

FailedLaunch said:
DC component means that there is energy (power) permanently leaving the source, i.e. the source is "evaporating" so it would eventually seize to exist.
Actually, it looks like I'll need to take this explanation back since it doesn't appear to be to right reason while the used AWGN is zero-mean. :rolleyes:

Check also this thread: [thread=163759]White noise in communication channel[/thread]
 
dexterdev said:
Hi all,
I got an idea why noise is having a gaussian pdf but didnot understood why it should be having zero mean. Why can't noise contain a dc content.

-Devanand T

Your question is quite a reasonable one. It is largely a matter of semantics and practicalities, I think. In all analogue equipment, you can expect very low rate drifting of DC levels and these will mostly be due to other effects (due to thermal variations of gain etc.) than the normal Gaussian Noise that you can also expect. Most analogue channels are deliberately made 'AC' coupled so there is a lower frequency limit, below which we are not particularly interested and it is not very relevant, in any case. Also, if you think of noise power in W/Hz, there is very little of it in a small fraction of a Hz, compared with normal signalling rates.
 
Thankyou for your helpful replies. but what does "matter of semantics" mean
 
dexterdev said:
Thankyou for your helpful replies. but what does "matter of semantics" mean

It refers to the meaning we attribute to words. We make 'assumptions' when we read some words but different people make different assumptions. Mostly, we take it for granted that channels are AC coupled because DC is too difficult. (I can't think of an example, in fact. Now that's a challenge for you all.)
 
thanks again sir
 
Back
Top