Not sure if there is a way except triggering the LFO-reset.
But has anyone created a script tool or some such that will let you have a unipolar LFO in renoise?
Problem with reset, isif you have really slow changes it will be a mess trying to find thespots to reset it.
Essentially would like to be able to set a high or low threshold and the LFO should only change the value in one direction.
What’s wrong with:
using offset
using graph custom lfo
using a hydra in between to make sure it travels only in the wanted range
formula device, on its own or in between the lfo and whatever you modulate
what exactly do you want to achieve? I.e. how do you want a sine or saw or so act on another parameter? Try to see it visual, as curves.
But I also know your pain, the lfo is a bit off the other modulation devices with its silly amplitude and offset tunings. I normally make it cycle in the full range and then use a hydra in between to set the actual range I want to modulate.
An artificial threshold, with a formula device, would alter the shape of the modulation. But is possible too, either clipping the original shape or wrapping it around between positive/negative at center or a threshold you want.
This stuff can be a bit confusing after all, because the meta devices act not absolute, but mostly relative.
As OopsIFly points out, the LFO reset/offset command can start the LFO from whatever position you want:
image source: this post
Problem with reset, isif you have really slow changes it will be a mess trying to find thespots to reset it.
Not really, because reset/offset is always relating to the position - no matter how much time has passed, x80 is always the middle.
But maybe both I and OopsIFly have misunderstood what you’re trying to achieve here?
That’s why make the LFO reset a automat-able slider 
Using the picture as reference.
Copy x800 to x880 and paste it on x880.
So a “bouncing” ball graph.
Using the picture as reference.
Copy x800 to x880 and paste it on x880.
So a “bouncing” ball graph.
I cannot attach things here, is that normal? I would like to show via an xrnt file with 4 devices how to achieve this.
In the mean time, you can try to set it up yourself:
add sine LFO at wanted speed, keep offset at zero and set amplitude to 100%
link the LFO into the first parameter of an empy formula device, enter the formula " abs(A-0.5)*2 ".
link the formula device into a hydra’s input - set the hydra outputs to what you wish to modulate, setting min/max.
In my example I linked it into a filter, and the cutoff was nicely bouncing like a half sine.
if you use a “saw” lfo, you will get a triangle out of it. A Square outputs 1.0 all the time.
you can swap the hydra’s min and max to reverse the motion i.e. put it upside down.
Thanks OopslFly, haven’t tried the formula device before.
But that seems to do the trick.
May have jumped the conclusion.
As far as I can tell the problem is still the bipolar oscilation of the LFO.
Unless I’ve made a misstake somewhere you get a flatline between x880 and x900 (if using the picture above as reference).
Not sure if you can view the values that gets processed by the formula device, that would have helped.
But what I’m guessing is that with the offset set to 0 on the LFO it will send 0 during the whole period of x881-x8ff.
Which will give a flatline for that period of time and then move up again.
Ok.
In this subforum I may not attach files as it seems, but I can put the copy&paste code for my setup into a spoiler. Open it up, try to mark everything and then copy and paste the whole text into an empty track in renoise. Select within the device chain with right mouse click “device chain->paste”. Yes it is a lot code, sorry for that. It should replace the device chain with 4 devices I set up demonstrating the halfsine-bouncer-boogey.
Do not mess with the lfo or formula device settings. Configure what you want to control only in the hydra.
fold_lfo
true
Mixer
true
false
Init
Bundled Content
false
1.0
Device only
0.5
Device only
1.0
Device only
0.0
Device only
0.5
Device only
1.0
Device only
true
true
true
Init
Bundled Content
true
1.0
Device only
-1
Device only
2
Device only
1.0
Device only
1.0
Device only
0.0
Device only
3.75
Device only
0.0
Device only
Curve
64
0.0
Unipolar
0,0.0
63,1.0
false
true
true
false
blank
User Library
true
1.0
Device only
abs(A-0.5)*2
A
B
C
true
0.915734768
Device only
0.0
Device only
0.0
Device only
-1
Device only
3
Device only
1.0
Device only
true
false
Init
Bundled Content
true
1.0
Device only
1
0.831469536
Mixer and Device
-1
Device only
4
Device only
2
Device only
0.0620457865
Device only
0.651299477
Device only
Linear
-1
Device only
-1
Device only
-1
Device only
0.0
Device only
1.0
Device only
Linear
-1
Device only
-1
Device only
-1
Device only
0.0
Device only
1.0
Device only
Linear
-1
Device only
-1
Device only
-1
Device only
0.0
Device only
1.0
Device only
Linear
-1
Device only
-1
Device only
-1
Device only
0.0
Device only
1.0
Device only
Linear
-1
Device only
-1
Device only
-1
Device only
0.0
Device only
1.0
Device only
Linear
-1
Device only
-1
Device only
-1
Device only
0.0
Device only
1.0
Device only
Linear
-1
Device only
-1
Device only
-1
Device only
0.0
Device only
1.0
Device only
Linear
-1
Device only
-1
Device only
-1
Device only
0.0
Device only
1.0
Device only
Linear
true
false
Init
Bundled Content
true
1.0
Device only
2x
2P K35
0.0
Device only
0.551992297
Device only
0.581000149
Device only
0.0078125
Device only
0.0
Device only
true
18
Edit: hope that helped.
Yes, you must have done something different, because for me it properly folds up the lower half of the sine, resulting in the pattern you suggested without any flat lines.
You can see to the right of the formula device a bar (like also with the lfo) showing the output, or you can link it to some parameter/slider and watch/read out the values there.
Not with offset 0% on the LFO.
With the offset set to 50% (or value 0.0 as you use in the code) you can just skip the formula and hydra device as far as I can tell.
It looks the same at least, except being twice as fast with the formula device 
Ah ok, sorry about the offset.
But it really does what you wanted. Repeating a half sine circle in unipolar fashion, resulting in a graph that looks a bit similar to a bouncing ball. Of Course it will be twice as fast as a standard sine, because the lower part (below zero) just gets mirrored into the positive range. I thought this is what you want.
And I’ve tested it, it actually works like this. I made a 4 second sample with 1.0 DC (drawing tool along the upper edge of the sample editor, disable audio output, dc is bad for your speakers…). Then I replaced the analog filter with a gainer cycling between -inf and 0.0 db. And resampled it in the track. Then you will see the shape in the sample editor.
Yeah the motion is fine, only problemis the starting point.
Could possible solve that with automating the initial rise so that the target will have the same value as 50% offset.
Then enable the LFO.
But it’s not as clean
Sorry I don’t quite so much get what modulation shape you are trying to achieve.
Maybe a little, but I’m not quite sure what you have on your mind.
You want the old center point of the lfo (in the pic at 800 & 880) modulation be preserved as new “lowest” point of the modulation?
Well then why not go to the hydra, and choose that point (old middle new lowest) as minimum value & you’re ready to go?
You could also change the Formula from “abs(A-0.5)*2” to “abs(A-0.5)+0.5” to get the effect in a different way.
I just thought it would be better to keep the modulation under more exact control via using the whole lfo range 0…1 and then defining the limits it should act on in the hydra. Like I almost always use a hydra between the lfo and what is modulated, to have this amount of control that is harder to adjust with the amp/offset lfo controls.