On p. 461 of the Sixth Edition of Hardy and Wright's An Introduction to the Theory of Numbers (Theorem 423). (see this post) it is written that - $$ \int_{n-1}^{n}\log^h{(\frac x t)}dt \geq \log^h{(\frac x n)} $$
My Understanding: We see that $\int_{n-1}^{n}f(x)\,dx$ is accumulating/summing/integrating all points from $n-1$ to $n$ for function $f$ whereas $\int_{n-1}^{n}f(n)\,dx = f(n)$ is just holding the value of $n$ for function $f$, thus it is always true that $$f(n)\leq \int_{n-1}^{n} f(t)dt $$
Is it correct?
Confusion: I have been told that it's generally true if $f$ is decreasing (and integrable), which $t\mapsto \log^h(x/t)$ is. If $f$ is decreasing, then $\int_{n-1}^n f(t)\textrm{d}t\geq \int_{n-1}^n f(n)\textrm{d}t=f(n).$ If $f$ is increasing, then $f(x) \le f(n)$ for all $x \in [n-1,n]$ Hence,$\int_{n-1}^{n}f(x)\,dx \le \int_{n-1}^{n}f(n)\,dx = f(n)$, but how does increasing or decreasing matter ? The function is given for a point and integration is done in between an interval.
Please clear my confusion in detail, in general,
How can I prove rigorously that -
$$f(n)\leq \int_{n-1}^{n} f(t)dt ?$$
what are the necessary conditions?
I'll just explain what they did there. Bro... Let $a = Re(s) > 0$. Then $\frac 1{t^{1+a}}$ is decreasing on $[n,n+1]$. So, $\frac 1{t^{1+a}}\le\frac 1{n^{1+a}}$ for $t\in [n,n+1]$. Thus, $\int_n^x\frac 1{t^{1+a}}\,dt\le\int_n^x\frac 1{n^{1+a}}\,dt$. Now, the integrand on the RHS is a constant, so you can pull it out of the interal: $$\int_n^x\frac 1{t^{1+a}}\,dt\le\int_n^x\frac 1{n^{1+a}}\,dt = \frac 1{n^{1+a}}\int_n^x\,1\,dt = \frac 1{n^{1+a}}(x-n)\le \frac 1{n^{1+a}},$$because $x-n\le 1$.