Some tools are optimized for dialog processing, so there are some built-in profiles and optimizations for typical frequency content and dynamic nature of human voice. Then there is usually also some frequency smoothing of that threshold curve, which can prevent another type of reduction artifacts.
and to be as natural as possible without jumps in noise level and signal cutout artifacts.
Important thing in case of dynamic processing is also timing behavior. ReaFir is quite general FIR processor, which can work as multiband gate or subtract captured noise profile from signal.īoth will work to certain basic degree, but specialized tools for noise treatment have for instance floating threshold detection, so it automatically changes in time and adapts to incoming signal. I don't think, you'll find mentioned ReaFir or ReaGate as effective in bg noise removal as commercial tools from Cedar, Izotope, Sonnox or Algorithmix.
