# Modulation values: 0 is not really 0?

If you doing a pitch modulation by adding a constant operand to the pitch, then the value “0.000” will not result in 0, but something like near to 0 />/>

Steps to reproduce:

2. Go to sampler and move macro knob “PWN modulation speed” to 0. You will still hear a modulating sound. Since both layers have the same loop length and finetune, they should eraze themselves, since -1 + +1 = 0 and so on.

3. Now remove the pitch operand on set 02. You will now hear how it should sound like - BUT why do I still hear some quiet (non modulating) sound? If math would do a precise job here, we should hear exactly nothing, right? Is it because of the sample-reverse function maybe?

Expected result: Exactly no sound at all

Seems to be a tiny inaccuracy with the macro when setting the bipolar pitch operand. Even though the macro shows 0.0, the operand value is actually off by a very tiny amount (off by 0.00049996, to be precise). If you manually set the operand to 0.0, then it will truly be 0.0 and it will not exhibit any modulation. We’ll have to take a closer look at possible ways to improve the macro accuracy here.

In the meantime, if you set your macro range so that it has an equal amount of positive and negative travel, ie. -0.5 to +0.5, then a macro value of 50.0 will result in an operand value of precisely 0.0 with zero movement.

Your sample is not perfectly symmetrical, so reversing it will not produce a perfectly phase inverted result in this particular case. When it’s not a perfect phase inversion, you will get some residual noise remaining.

So instead of reversing the 2nd sample, you should clone the 1st sample and then phase invert it instead.

Here’s a fixed instrument with the changes I’ve mentioned above:

Doubleclick on the pitch operand’s value-slider to set it to zero. It seems you’ve accidently set some tiny value beyond 3 digits. After the doubleclick everything works like expected.

Edit: uhm… works for the first keystroke, then indeed is out of tune again. :-S

@dblue: Still this is kinda strange. I just copied the first sample and phase-inverted it (to make sure there are no unexpected differences in the samles), but still the samples were out of tune after an initial keystroke

Hm, okay… indeed, when the bipolar parameter range is symmectrical on the 0-axis, everything works fine.

It changes to 0.0005. The floats of the rotators should either be limited to 3 digit decimals or the absolute drag down value from the mouse should really reach 0.00000000000

[quote=“dblue, post:2, topic:40577”]
Your sample is not perfectly symmetrical, so reversing it will not produce a perfectly phase inverted result in this particular case. When it’s not a perfect phase inversion, you will get some residual noise remaining.

So instead of reversing the 2nd sample, you should clone the 1st sample and then phase invert it instead.
4585 Jurek-Simple-PWM-fixed.xrni
Thanks dblue. That sounds completely logical

It’s the min mapping of the macro, not the macro value itself which is not exactly 0 here, but 0.00000something. If you type in “0” for the min value, it will work correctly.
Will try to improve dragging/changing of such values. Indeed is pretty confusing…