I have noticed that, when you try to humanize very high or very low values, it’s very likely that the value won’t change, I think because there is bound checking preventing that.
C-4 01 00
humanize range: 0A
as far as I can understand, if you humanize the 00 value, Renoise tries to add a random value between the -0A and 0A to 00. If the added value is negative, then of course the old value will not change.
what I propose is that, when humanizing range is higher than maxvalue-currentvalue or currentvalue-minvalue, the humanizing value would be taken in absolute value.
The only exception in my opinion is the humanization of the delay column.
I still think it’s a flaw that negative numbers will not move the note to previous line (with the exception of line 00 and when the previous line is occupied where in case it should use absolute values as well).
I was going to post this when I read the first post. Yes, if the delay column would be humanized in the way It-Alien described, it somehow ends up with a slight total delay, which is of course unsatisfactory.
An instrument player unintentionally plays some notes slightly sooner and some later than a perfect quantization. I think that the delay column can also be humanized in a more efficient way:
Before delay column humanization:
After delay column humanization:
Mmm, negative delays would be really cool. Renoise could, internally, check the next line it is about to play: if it has a negative delay, play the note as if it were on the current line (with delay FF minus whatever the negative delay was).
This is totally blue-sky feature-creep, of course. But, very useful for slow-attack instruments and humanising.
Or, you can use the PDC negative track delay to do a similar thing.