Why Pe Is Not Proportional To Input SNR

Question

Why does the slide seem to say P_e is proportional to (S/N)_in?

Short Answer

Strictly, P_e should not be directly proportional to (S/N)_in.

If (S/N)_in means input signal-to-noise ratio, then higher input SNR normally makes the probability of error smaller:

[ (S/N)_{in} \uparrow \quad \Rightarrow \quad P_e \downarrow ]

So the slide’s last line is likely shorthand for “P_e depends on (S/N)_in”, or it is missing an inverse/decreasing relationship.

Claim status: INFERRED, based on the slide context and the standard meaning of signal-to-noise ratio.

Why Higher SNR Lowers Error Probability

Digital detection usually works by comparing a received signal value to a decision threshold.

An error happens when noise pushes the received value to the wrong side of that threshold.

  • stronger signal received levels are farther apart
  • weaker noise less random pushing
  • larger SNR less overlap between the possible received levels
  • less overlap lower probability of choosing the wrong code

That is why P_e is usually a decreasing function of SNR.

More Accurate Relationship

For many digital communication systems, P_e is not a simple linear proportion. It is more like:

[ P_e = Q(\sqrt{k \cdot SNR}) ]

or approximately decreases exponentially at high SNR:

[ P_e \approx e^{-k \cdot SNR} ]

The exact formula depends on the modulation and detection method, but the direction is the same:

[ SNR \uparrow \Rightarrow P_e \downarrow ]

How This Connects To The Slide Formula

The slide gives:

[ \left(\frac{S}{N}\right)_{out}

\frac{M^2}{1 + 4P_e(M^2 - 1)} ]

Here:

  • M is the number of quantization levels
  • P_e is the probability of a transmission/decision error
  • higher P_e makes the denominator larger
  • therefore higher P_e makes output SNR worse

If:

[ P_e = 0 ]

then:

[ \left(\frac{S}{N}\right)_{out} = M^2 ]

This is the best case shown by the formula.

Practical Exam Interpretation

Read the last line as:

[ P_e \text{ is controlled by } (S/N)_{in} ]

not as:

[ P_e \propto (S/N)_{in} ]

A safer statement is:

[ P_e \propto \frac{1}{(S/N)_{in}} ]

as a rough intuition, or more accurately:

[ P_e = f((S/N)_{in}) ]

where f is a decreasing function.