"Saturation" or "clipping" function: clip(x)=min(max(x, -1), 1)

The context where I encountered this is LTI (linear time-invariant) system analysis (in classic control theory).

I tried to take its laplace transformation and mix it into the system's transfer function as usual, but soon figured out that I can't perform the inverse laplace transformation once it's mixed in.

Fortunately, an analytically accurate solution is not necessary. So I figured I need a simple yet effective function that does pretty much the same as clip(x), with a simpler Laplace Transformed form.

Or... Are there any simpler ways of doing it (other than doing an discrete-time simulation)?


clip() is not a LTI filter (it's not linear). So I don't understand how you'd expect to use Laplace/Fourier here.

  • $\begingroup$ That's the point -- is there a "good enough" approximation to it that CAN be laplace'd? $\endgroup$ – smilekzs Jun 16 '13 at 1:37
  • $\begingroup$ No. I think you are missing the point of Laplace/Fourier transform in this context. They are primary related to the temporal characterization of a filter. Eg, if you have a memoryless filter, it must be linear and its Laplace transform is trivial (a constant). That would be the only approximation. $\endgroup$ – leonbloy Jun 16 '13 at 14:02
  • $\begingroup$ Only in very special cases (for very restricted input signals), you could approximate the effect of memoryless alinear filter by a linear filter with memory: eg, saturation "produces/enhances" high frecuencies $\endgroup$ – leonbloy Jun 16 '13 at 14:05
  • $\begingroup$ Ah I see. So basically waveshapers are of a different kind from LTI filters. Thanks! $\endgroup$ – smilekzs Jun 17 '13 at 14:46

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.