Originally Posted by guitfnky
an interesting idea, but can't you do this already to a large degree just by making combinators using existing stuff? I haven't had a chance to read through your thread, eXode, and I'm not familiar with Reaktor, so forgive me if that's a dumb question. I am most definitely ignorant on the subject.
Combinators only combines high level existing devices while Reaktor allows low level and mid level DSP processing modules be combined.
For example if an engineer wants to implement a high order custom FIR (Finite Impulse Response) Filter, QMF Filters set (Quadrature Mirror Filters) or IIR (Infinite Impulse Response) Filter using Butterworth, Bessel, ... design techniques or wants to fragment an audio file on sample blocks to do a FFT (Fast Fourier Transform), wavelet transform,... and then windowing using Hanning, Bartlett, ... actually is completely impossible to do using combinators.
There are an infinite list of examples that actually can not be done using combinators and current Racks and Rack extensions.
Imagine for example AutoTune Pitch correction algorithm by Antares. It was designed using Time domain tecniques and has very low latency for pitch correction in real time.
Reason's Neptune was designed using frequency domain techniques and latency is higher than Antares Auto-Tune for real-time use.
Advanced users can do for example time domain pitch correctors (for example) using low level DSP operators like NI Reaktor does.