Statistically-targeted forcing (STF) is a recently proposed method for synthetic turbulence generation that uses a relatively simple, deterministic linear forcing term to produce a desired target Reynolds stress distribution based on either time- or spatially-averaged statistics. This paper presents two updates to the original STF method. The first is the use of a dynamic mapping tensor in the forcing term to allow more effective control of second-order statistics without the need for a user-determined forcing time scale. The second is adoption of combined low- and high-pass filtering to selectively input energy primarily near a single scale, improving the spectral behavior of the generated turbulence. Initial results are presented for synthetic homogeneous isotropic turbulence with varying characteristic length and time scales.