Raabe's integral is usually stated in the form
$$ I(\alpha):=\int_0^1 \log\left(\,\Gamma\left(x+\alpha\right)\,\right)\,{\rm d}x =\frac{\log\left(2\pi\right)}{2}+\alpha \log\left(\alpha\right) -\alpha\,,\qquad \alpha \geq 0. $$
(see for example this question.) I was wondering whether the condition $\alpha\geq0$ is necessary. Since $\Gamma(x+\alpha)$ is a meromorphic function, so is $I(\alpha)$, and therefore differentiation yields
$$I'(\alpha)=\int_0^1\psi(x+\alpha)\,dx =\log\Gamma(1+\alpha)-\log\Gamma(\alpha)$$
from where we deduce that
$$I(\alpha)=\int_0^1(\log\Gamma(1+\alpha)-\log\Gamma(\alpha))\,d\alpha+C$$ for some appropriate constant $C$. If $\alpha$ were real $\log\Gamma(1+\alpha)-\log\Gamma(\alpha)=\log\alpha$ trivially, but when $\alpha$ is complex some care should be taken. We can write $\log\Gamma(1+\alpha)=\log\Gamma(\alpha)+\log\alpha+2\pi i k(\alpha)$ with $k(\alpha)$ an integral-valued function, and hence
$$I(\alpha)=C'+\alpha\log\alpha-\alpha+2\pi i\int k(\alpha)\,d\alpha.$$
How one can show that $k(\alpha)=0$ (if this is really true), or at least, by any other method, compute $I(\alpha)$ for (some) complex arguments?