Phijel

does anybody know if there's a way to "invert" an exp scaling so that it goes back to linear like in the following example.

scaling-back-to-linear

https://cycling74.com/forums/logarithmic-to-lineair

If the first lin->exp scaling is in the form x^y and you know the value used for y then you can reverse it by using x^(1/y) e.g. x^2 -> x^0.5

thanks, I'll look into that, missed that thread...

Any idea tho on how the exp base value used as the fifth arg in scale would translate into an expr using pow ? my maths are far away ;-)

Any idea tho on how the exp base value used as the fifth arg in scale would translate into an expr using pow ?  my maths are far away ;-)

I checked your patch earlier and used my suggestion of taking the reciprocal of the exponent as the new exponent in a reverse conversion, and I got a straight line....

So I think if you were using [expr pow] it translates to how the curve would behave if [scale] were first converting it's input to 0->1, raising the result to a power, then scaling/offsetting as required by [scale]'s output behaviour. If that makes sense....

I checked your patch earlier and used my suggestion of taking the reciprocal of the exponent as the new exponent in a reverse conversion, and I got a straight line....  

So I think if you were using [expr pow] it translates to how the curve would behave if [scale] were first converting it's input to 0->1, raising the result to a power, then scaling/offsetting as required by [scale]'s output behaviour.  If that makes sense....


scaling back to linear ?