Non linear sliders/knobs, best practice?

[Morphoice]

would you rather change the skew factor / middle Position on a knob/slider or use cable_tables in a network to make them non-linear/exponential/logarithmic? or is there even another method I don't know yet?
Is the skew factor curved or just a linear knee?
Which method is more precise on 0-1 sliders?

the nodes also offer a curve on each knob, but that's reset once a faust is recompiled... also I noticed it can lead to snapping of knobs

[ustk]

@Morphoice I don't know about Faust resetting the curve but:

• The curve you see does nothing but adjusting the skew factor behind the scene. So no it's not linear, it is exactly what you see.

• I personally only use the skew number, that's my habit, and it's one less node in the graph! But a cable_table would do the same (but of course is necessary for more complex behaviour)

• in the end, most of the time the curves I need are either linear or log, so the skew is often either 1.0 or approx 0.3 for a kinda log behaviour

• If the slider is snapping, it's probably because StepSize isn't 0.0

[Morphoice]

@ustk does that apply to sliders in the UI too? I noticed there is no skew factor there but a mid point, so is it still a curved metric?

[griffinboy]

@Morphoice

There isn't a consistent system.
I always interpret the knobs at the last step.

I do it manually rather than use hise's skew.
Make everything work standardized behind the scenes (percentages when possible, or standard ranges) and then I use the UI script to interpret all of them with different curve functions. That way there is consistency and it's easy to change anything.

Then again I work with c++ so I'm able to be quite picky with how I want parameters to be interpreted in the code