Simple code question for Urs

[EvilDragon]

If it's not a big secret, could you please share the pseudocode for "make unipolar" and "soften" functions in your modmappers? I'm trying to implement something similar in a Kontakt script, but to no avail so far.


Thank you very much!

[EvilDragon]

As a side-question, is "make unipolar" even working correctly?

Currently if we have the whole modmapper set to -100, it never changes any values when you select "make unipolar". Isn't the point of "make unipolar" to move all values to positive side of the modmapper?

[V0RT3X]

you should just PM him?

[EvilDragon]

Oh well, somehow didn't think of that.

[Howard]

As a side-question, is "make unipolar" even working correctly?
Currently if we have the whole modmapper set to -100, it never changes any values when you select "make unipolar". Isn't the point of "make unipolar" to move all values to positive side of the modmapper?

1) The current function name is misleading - it should read "normalize unipolar" (and "normalize" should read "normalize bipolar").
2) In your example case, "normalize unipolar" should actually set all values to zero. Bug-ette

[EvilDragon]

That makes sense, Howard.

I have managed to figure out the "make unipolar" thing. It's just normalizing everything to 50% and adding 50 to it.

Code: Select all

{ find maximum absolute value }
for i := 0 to Steps - 1
	if max < abs(Table[i])
		max := abs(Table[i])
	end if
end for
{ scale to 50% and offset by 50 to move everything to positive range }
for i := 0 to Steps - 1
	Table[i] := ((Table[i] * 50) / max) + 50
end for

I'd say that "make unipolar" function definitely behaves weirdly. At least it seems it doesn't follow the code above...

[Urs]

make unipolar:

Code: Select all

float min = 1.f;
float max = -1.f;
														
for ( int i = 0; i < NumSteps; i++ )
{
	if ( steps[ i ] < min ) min = steps[ i ];
	if ( steps[ i ] > max ) max = steps[ i ];
}
							
							
if ( min != max )
{
	float add = ( max + min ) * 0.5f;
	float mul = 1.f / ( max - min );
	
	for ( int i = 0; i < NumSteps; i++ )
	{
		steps[ i ] -= add;
		steps[ i ] *= mul;
		steps[ i ] += 0.5f;
	}
}

soften (simple lowpass filter):

Code: Select all

float tmp = steps[ 0 ];

for ( int i = 1; i < NumSteps-1; i++ )
{
	float v = steps[ i ] * 0.6f + 0.2f * ( tmp + steps[ i + 1 ] );
	tmp = steps[ i ];
	steps[ i ] = v;
}

[EvilDragon]

Thanks a million!

Simple lowpass, duuuuh I'm stupid. -_-


My "make unipolar" works nicely though (code from previous post)! Sorry for not PMing you instead

[SJ_Digriz]

Sorry for not PMing you instead

Actually, I appreciate the open question and answer! Super interesting and informative with no drama ... I thought I woke up in a parallel universe with a different internet there for a sec.