# Why are some Shannon's limit plots exponential rather than logarithmic?

This diagram again. This is clearly not the shape of $$\log_2(1+SNR)$$ but I very frequently see this instead of this:

But this is also a stupid diagram because it doesn't cross at the origin.

Also why do the modulation techniques veer away from $$\log_2(1+SNR)$$, they should stay on the curve and then get suddenly capped at their spectral efficiencies.

• The upper graph clearly IS the shape of $log_2(1+SNR)$. Why do you think it's not ? – Hilmar May 24 at 12:27
• @Hilmar imgur.com/a/DLHhKPb because that's the shape – Lewis Kelsey May 24 at 12:32
• @LewisKelsey the question was why do you think it's not? Seriously, insert the definition of dB into log(1+SNR), and come to your own conclusions, instead of citing other pictures! – Marcus Müller May 24 at 12:56
• @MarcusMüller well actually I typed that in desmos to illustrate what I was expecting, so that's my picture, but that's it, dB is logarithmic which I forgot about, thanks for the reminder. This is the answer – Lewis Kelsey May 24 at 13:06
• Morale of the story: read the axis labels! – Hilmar May 24 at 13:38