Nice online converter if anyone wants to experiment with assignments:
I just had a thought. The word "Assignments" might be the key. Perhaps there are 8 Bezier curves described in BezierCurvesEntry and [Gesture | Event]CurveAssignmentsEntry simply assign them to various parameters. GestureCurveAssignmentsEntry contains 47 digits which appear to be in the range 1-8.
So for example for Drum, all gesture parameters point to Bezier curve 1 except for parameter 42 which points to Bezier curve 4:
Drum:
AQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEEAQEBAQE=
[1,
1,1,1,1,1,1,1,1,
1,1,1,1,1,1,1,1,
1,1,1,1,1,1,1,1,
1,1,1,1,1,1,1,1,
1,1,1,1,1,1,1,1,
4,1,1,1,1,1]
and for EventCurveAssignmentsEntry, it contains 14 parameter mappings:
AQEBAQEBAQEBAQEBAQE=
[1,1,1,1,1,1,1,1,1,1,1,1,1,1]
I did much the same when I first started thinking about the bezier curves. I first tried swapping the curveAssignmentsData string only - I can't recall exactly why but I had a suspicion that it related to the modes (lead,chords etc.) somehow. Perhaps these can be a red herring if they are being ignored in favour of the settings in one of the other modes and so seem to have no effect? This, or something like it, would possibly explain some of the odd things I have experienced.
I don't think I initially paid enough attention to the various curves when I started down this path. There are actually 3 types of curves: BezierCurvesEntry, GestureCurveAssignmentsEntry and EventCurveAssignmentsEntry. For example:
<BezierCurvesEntry name="Default BezierCurves" bezierCurvesData="AAAAAAAAAABx9ChGcfQoRnH0qEZx9KhGAP7/RgD+/0YAAAAAAAAAAHH0qEaNj21EUt/6RutF4UUA/v9GAP7/RgAAAAAAAAAAZ5gZRsLVo0N7UDhGUt/6RgD+/0YA/v9GAAAAAAAAAABx9ChGcfQoRnH0qEZx9KhGAP7/RgD+/0YAAAAAAAAAAHH0KEZx9ChGcfSoRnH0qEYA/v9GAP7/RgAAAAAAAAAAcfQoRnH0KEZx9KhGcfSoRgD+/0YA/v9GAAAAAAAAAABx9ChGcfQoRnH0qEZx9KhGAP7/RgD+/0YAAAAAAAAAAHH0KEZx9ChGcfSoRnH0qEYA/v9GAP7/Rg=="/> <GestureCurveAssignmentsEntry name="Default GestureCurveAssignments" curveAssignmentsData="AQUFBQUFBQUFAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEFAQEBAQEEAQEBAQE="/> <EventCurveAssignmentsEntry name="Default EventCurveAssignments" curveAssignmentsData="AQEBAQEBAQEBAQEBAQE="/>
I in correctly assumed they were all Bezier curves, but only the first node actually mentions "Bezier". The data from the others (Gesture & Event) is certainly something else. I'll come back to these since I now have a clear path on Bezier curves, but I haven't quite figured out what the other curves are.
I was going to experiment further with the patterns you listed before but I got distracted with trying to get a couple of other presets working better or at all. I've sort of done with that now. (Although not entirely successfully - with a sample based preset I found that I lost the character of the instrument sound - perhaps because I'd had to pitch correct the samples.)
The second pattern is the closest to one that might do what I was looking for - but would probably want something flatter at the start. But I guess different curves will work better with different patches.
I'm finding the Orba to be a bit erratic when loading modified presets and sometimes find I'm not actually getting to hear the effects of what I've done when I expect to.
I'm working on a new utility to edit Bezier curves. Here is a screen shot of what is defined in the Chord Ambeeant Preset. It appears that only 2 of 8 curves are defined and the rest are all set to linear Bezier curves. This is the data from Pattern7.txt above. I'm going to load in the other 6 this weekend. The Bezier widget I found allows me to edit the curves so I can insert them into other Presets to see if they actually do anything?!
I had another thought this morning - I have been trying lead presets in isolation. But perhaps, just as chords can be "inherited" from the underlying song (Having the Ambieent song at the moment on the Orba my Dulcihack preset now plays in minor - which is quite nice actually), possibly the bezier curves could also be inherited. It would make some sense since gestures may affect chords, bass and drums as well as the lead.
I've not been able to try out this idea properly yet though.
I've tried swapping out the GestureCurveAssignmentsEntry alone and BezierCurvesEntry alone and also the whole Bezier caboodle on a preset with seekers only for pitch bend on tilt and vanilla synthPatch - but without any effect.
I'm not convinced that my experiments have been thorough enough though at this point, but I'm beginning to wonder whether this is something that's not actually being used by the Orba.
As an aside, I was poking around some of the sample sets and found a couple of sets that had samples at such a high pitch they were painful, and a couple that might have been above my (aged) hearing capability! Which sort of re-enforces the idea that there's stuff that just isn't used in the data and presets.
But I will do a bit more experimenting.
From the end users point of view this could make the Orba a better, expressive instrument to play, (along with what you did with the samples on the Pandrum preset) - to me that seems a big step up from being just a toy to a real instrument.
I can only describe my thinking - which I imagine may well be too simplistic.
I am thinking, in terms of a graph with the x axis being either the rate of change of the application of the effect or the total amount of effect at each point in time (which would be constrained by parameters in the patch probably) - because I see the Y axis as time (although this might be better thought of as the passage through a gesture).
I assume only a simple curve - a C shape and perhaps S - is used. Something more complex might be considered desirable, but would you really need it? (I also thought why make it more complicated than needed when the computing power is limited?)
Going back to my vector art thinking only 4 coordinates would be required - a start and end position and the co-ordinates of the two "hooks" that define the actual curve. This would require 8 numbers to be passed to whatever routine actual does the calculations. It seems to me that the start and end point in time (Y) could be derived from the parameters in the patch, perhaps the weight (X) as well. But they may be included for convenience. (Quite why they should be obfuscated in a base64 string rather than just hex I don't know.)
Sticking with that idea there could be four hooks, as I have sometimes seen, and that would tie in with your breakdown of the numbers as in your previous post.
The straight line you found seems to fit with what I would expect - an even addition to the effect over the passage of the gesture. Perhaps the other strings simply define a basic set of curves which could be used, keeping the whole thing simple. To me that seems a sensible approach - giving options that can be simply applied by preset designers. If they are lazy a designer might simply try a straight line first and not bother to try others to see if they work better - which would have been a good idea on the Tilt Lead preset in my opinion. This would be good for us too, because we could stick to known safe strings to use and with some degree of understanding what they will do.
And I agree you may have cracked it!
I think I've got it! It appears to be using the Float (IEEE754 Single precision 32-bit) standard and is recorded in reverse (many programs do this to simplify bit-shifting). I converted the curve:
[0,0,0,0],[0,0,0,0],
[113,244,40,70],[113,244,40,70],
[113,244,168,70],[113,244,168,70],
[0,254,255,70],[0,254,255,70]
to
0.0 , 0.0
1.08131103515625E4 , 1.08131103515625E4
2.1626220703125E4 , 2.1626220703125E4
3.2767E4, 3.2767E4 (Note this value is the max value for a 16-bit signed int)
..which is completely linear! This can't be a coincidence. I'll need to convert the other curves this week to see what all the curves look like. The first one is here:
Based on the above, my assumption is that there are several Bezier curves defined. I would guess that one Bezier is defined like this:
[0,0,0,0],[0,0,0,0],
[113,244,40,70],[113,244,40,70],
[113,244,168,70],[113,244,168,70],
[0,254,255,70],[0,254,255,70]
as
[P0x],[P0y],
[P1x],[P1y],
[P2x],[P2y],
[P3x],[P3y]
I'm then going to speculate that each of the 4 integers somehow represent a 32 bit number (likely a float). I'm going to research some different ways of storing these numbers and can hopefully find the right schema. I feel like I'm on the right trail.
Here are the 7 BezierCurvesEntries along with the Presets they are used with. Note: pattern 7 published above is unique to the Ambient Chord Preset. I played with this sound to find out what was so special about it. I found this sound has NoteOff velocity which I hadn't noticed from any sound before. I'm not sure how this is achieved or even if the Bezier curve(s) have any impact.
I'm also looking at BezierCurvesEntry which has 7 unique Base64 strings:
bezierCurvesData="AAAAAAAAAABx9ChGcfQoRnH0qEZx9KhGAP7/RgD+/0YAAAAAAAAAAHH0qEaOj21EUt/6Ru1F4UUA/v9GAP7/RgAAAAAAAAAAZ5gZRsLVo0N7UDhGUt/6RgD+/0YA/v9GAAAAAAAAAABx9ChGcfQoRnH0qEZx9KhGAP7/RgD+/0YAAAAAAAAAAHH0KEZx9ChGcfSoRnH0qEYA/v9GAP7/RgAAAAAAAAAAcfQoRnH0KEZx9KhGcfSoRgD+/0YA/v9GAAAAAAAAAABx9ChGcfQoRnH0qEZx9KhGAP7/RgD+/0YAAAAAAAAAAHH0KEZx9ChGcfSoRnH0qEYA/v9GAP7/Rg=="/> bezierCurvesData="AAAAAAAAAAAbIdtFuHmURkLVY0atYNBGAP7/RgD+/0YAAAAAAAAAAHH0qEajwHVEUt/6Ru1F4UUA/v9GAP7/RgAAAAAAAAAAZ5gZRtDVo0N7UDhGUt/6RgD+/0YA/v9GAAAAAAAAAAC+J5xGwtWjQ6PA9UbC1SNEAP7/RgD+f0YAAAAAAAAAAHH0KEZx9ChGcfSoRnH0qEYA/v9GAP7/RgAAAAAAAAAAcfQoRnH0KEZx9KhGcfSoRgD+/0YA/v9GAAAAAAAAAABx9ChGcfQoRnH0qEZx9KhGAP7/RgD+/0YAAAAAAAAAAHH0KEZx9ChGcfSoRnH0qEYA/v9GAP7/Rg=="/> bezierCurvesData="AAAAAAAAAABx9ChGcfQoRnH0qEZx9KhGAP7/RgD+/0YAAAAAAAAAAHH0qEajwHVEUt/6Ru1F4UUA/v9GAP7/RgAAAAAAAAAAZ5gZRtDVo0N7UDhGUt/6RgD+/0YA/v9GAAAAAAAAAAC+J5xGWDKiQ00x80aspBtEAP7/RgD+f0YAAAAAAAAAAHH0KEZx9ChGcfSoRnH0qEYA/v9GAP7/RgAAAAAAAAAAcfQoRnH0KEZx9KhGcfSoRgD+/0YA/v9GAAAAAAAAAABx9ChGcfQoRnH0qEZx9KhGAP7/RgD+/0YAAAAAAAAAAHH0KEZx9ChGcfSoRnH0qEYA/v9GAP7/Rg=="/> bezierCurvesData="AAAAAAAAAAAbIdtFMDiURifIY0atYNBGAP7/RgD+/0YAAAAAAAAAAHH0qEaNj21EUt/6RutF4UUA/v9GAP7/RgAAAAAAAAAAZ5gZRsLVo0N7UDhGUt/6RgD+/0YA/v9GAAAAAAAAAAC+J5xG0qujQxe69UYNBCNEAP7/RgD+f0YAAAAAAAAAAHH0KEZx9ChGcfSoRnH0qEYA/v9GAP7/RgAAAAAAAAAAcfQoRnH0KEZx9KhGcfSoRgD+/0YA/v9GAAAAAAAAAABx9ChGcfQoRnH0qEZx9KhGAP7/RgD+/0YAAAAAAAAAAHH0KEZx9ChGcfSoRnH0qEYA/v9GAP7/Rg=="/> bezierCurvesData="AAAAAAAAAABx9ChGcfQoRnH0qEZx9KhGAP7/RgD+/0YAAAAAAAAAAHH0qEaNj21EUt/6RutF4UUA/v9GAP7/RgAAAAAAAAAAZ5gZRsLVo0N7UDhGUt/6RgD+/0YA/v9GAAAAAAAAAAC+J5xGWDKiQxe69Ub/LSNEAP7/RgD+f0YAAAAAAAAAAHH0KEZx9ChGcfSoRnH0qEYA/v9GAP7/RgAAAAAAAAAAcfQoRnH0KEZx9KhGcfSoRgD+/0YA/v9GAAAAAAAAAABx9ChGcfQoRnH0qEZx9KhGAP7/RgD+/0YAAAAAAAAAAHH0KEZx9ChGcfSoRnH0qEYA/v9GAP7/Rg=="/> bezierCurvesData="AAAAAAAAAAAbIdtFMDiURifIY0atYNBGAP7/RgD+/0YAAAAAAAAAAHH0qEaNj21EUt/6RutF4UUA/v9GAP7/RgAAAAAAAAAAZ5gZRsLVo0N7UDhGUt/6RgD+/0YA/v9GAAAAAAAAAAC+J5xGWDKiQxt/9UaspBtEAP7/RgD+f0YAAAAAAAAAAHH0KEZx9ChGcfSoRnH0qEYA/v9GAP7/RgAAAAAAAAAAcfQoRnH0KEZx9KhGcfSoRgD+/0YA/v9GAAAAAAAAAABx9ChGcfQoRnH0qEZx9KhGAP7/RgD+/0YAAAAAAAAAAHH0KEZx9ChGcfSoRnH0qEYA/v9GAP7/Rg=="/> bezierCurvesData="AAAAAAAAAABx9ChGcfQoRnH0qEZx9KhGAP7/RgD+/0YAAAAAAAAAAHH0qEaNj21EUt/6RutF4UUA/v9GAP7/RgAAAAAAAAAAZ5gZRsLVo0N7UDhGUt/6RgD+/0YA/v9GAAAAAAAAAABx9ChGcfQoRnH0qEZx9KhGAP7/RgD+/0YAAAAAAAAAAHH0KEZx9ChGcfSoRnH0qEYA/v9GAP7/RgAAAAAAAAAAcfQoRnH0KEZx9KhGcfSoRgD+/0YA/v9GAAAAAAAAAABx9ChGcfQoRnH0qEZx9KhGAP7/RgD+/0YAAAAAAAAAAHH0KEZx9ChGcfSoRnH0qEYA/v9GAP7/Rg=="/>
Each decode to 256 values. Here I split the last one in groups of 8:
0,0,0,0,0,0,0,0,
113,244,40,70,113,244,40,70,
113,244,168,70,113,244,168,70,
0,254,255,70,0,254,255,70,
0,0,0,0,0,0,0,0,
113,244,168,70,141,143,109,68,
82,223,250,70,235,69,225,69,
0,254,255,70,0,254,255,70,
0,0,0,0,0,0,0,0,
103,152,25,70,194,213,163,67,
123,80,56,70,82,223,250,70,
0,254,255,70,0,254,255,70,
0,0,0,0,0,0,0,0,
113,244,40,70,113,244,40,70,
113,244,168,70,113,244,168,70,
0,254,255,70,0,254,255,70,
0,0,0,0,0,0,0,0,
113,244,40,70,113,244,40,70,
113,244,168,70,113,244,168,70,
0,254,255,70,0,254,255,70,
0,0,0,0,0,0,0,0,
113,244,40,70,113,244,40,70,
113,244,168,70,113,244,168,70,
0,254,255,70,0,254,255,70,
0,0,0,0,0,0,0,0,
113,244,40,70,113,244,40,70,
113,244,168,70,113,244,168,70,
0,254,255,70,0,254,255,70,
0,0,0,0,0,0,0,0,
113,244,40,70,113,244,40,70,
113,244,168,70,113,244,168,70,
0,254,255,70,0,254,255,70
There's definitely some patterns here. I'm not really sure what these do yet but I'm going to try to build a list of Preset types/names that use each of the 7 strings to look for patterns.
I found this image online:
Drum:
AQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEEAQEBAQE=
[1,
1,1,1,1,1,1,1,1,
1,1,1,1,1,1,1,1,
1,1,1,1,1,1,1,1,
1,1,1,1,1,1,1,1,
1,1,1,1,1,1,1,1,
4,1,1,1,1,1]
Lead & Bass:
AQUFBQUFBQUFAQEBAQEBAQEBAQEBAQEBAQYGBgYGBgYGAQEFAQEBAQEEAQEBAQE=
[1,
5,5,5,5,5,5,5,5,
1,1,1,1,1,1,1,1,
1,1,1,1,1,1,1,1,
6,6,6,6,6,6,6,6,
1,1,5,1,1,1,1,1,
4,1,1,1,1,1]
Chord:
AQUFBQUFBQUFAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEFAQEBAQEEAQEBAQE=
[1,
5,5,5,5,5,5,5,5,
1,1,1,1,1,1,1,1,
1,1,1,1,1,1,1,1,
1,1,1,1,1,1,1,1,
1,1,5,1,1,1,1,1,
4,1,1,1,1,1]
I'm not quite sure if this is correct?! One thing is for sure, don't spend any time trying to find patterns in the Base64 string. I went down that rabbit hole last year and any patterns you find are just coincidence. The magic is in the number arrays somewhere..
Subskybox
This forum is intended to share Orba 2 hacking tips amongst the Orba 2 community. NOTE: Please post facts that are well understood & useful. If you have theories to discuss, please start another forum and link to it here.
2 people like this idea