Welcome to MOK Forum, this is a public forum where you can ask questions, discuss issues, and receive answers from other members of the community.

Note-on per-voice random?

+1 vote
Is there an LFO setting that will give a per-voice-unique random value at the start of each incoming midi note (not time based, so that it's unique per-voice even for notes triggered at the same exact time)?

Also, some kind of real-time visual indication on each modulator of its final output value would be very helpful. The ability to copy and paste oscillators, filters, and efx blocks would be great too to make creating stereo patches easier.

Also, maybe a bug? I was trying to make some stereo patches by running filt1 and filt2 in parallel and panning them left and right in the voice amp section, and I found that even with the exact same filter settings, one filter's output is always quieter than the other. I could be missing something in one of the mixers but I tried to make things as simple as possible to search out the issue. I just have oscillator 1 running into filt1 and filt2, both at %100 in the omix section, and both filters outputting at %100 in amp>voice amp>envelope 1 mixer but one is quieter than the other somehow.
 
I'm the only person writing on the forum... haha sorry if it's annoying. I'm really enjoying the synth though, it's super unique and I'll definitely tell others about it. Really impressive piece of work. Not like any other synth I've used, and I'm getting really musical results. The oscillators, filter options, and modulation possibilities make it really easy to make unpredictable, "organic" instruments that behave more like acoustic instruments.
asked Apr 6, 2018 in Waverazor by ipzid (200 points)
edited Apr 6, 2018 by ipzid

1 Answer

+1 vote

Not annoying at all!  More great stuff, and important to discuss.  Many thanks!  Let's parse this out again:

LFO unique random per voice - Yes, there is a way to do that, but trying it now, I think there may be a bug.  Let me explain...  First, to set it up, make an LFO with Shape set to Random S0-9 (doesn't matter which number), Trig Mode set to Random, and Mode set to Unipolar or Rectified.   Now, the frequency of the LFO doesn't go slow enough, so you have to freeze it.  The easiest way to do this is to set up a mod route with your Amp Env "latchout atk" as the source.  That output is a Gate signal that goes high and stays high as soon as the voice starts.  Then for the destination, choose the "freeze gate" input of the random LFO.  That tells the LFO to freeze when it sees a gate high input (with gate low, it's unaffected, so hitting it with a square wave or other shape from a different LFO is interesting).  With all that set up, you'll get a unique, static, random value for each voice you launch.

However, with some random generations, I'm getting intermodulation that results in audio artifacts.  Sometimes it's clicking.  Sometimes weird beating.  I don't know what that is, so I'm logging a bug for it.  Honestly, I like how it sounds, but I don't like that I can't control it.  We'll take a look at it.

Additionally, it might be easy enough to create another type of modulation source that's a random number generator with a static output.  That way, you don't have to freeze the LFO when all you want is one random value.  This is going on the list as well.

Modulator visual indication - The issue with this is choosing which polyphonic voice to visualize, and then when you're playing multiple voices, what you're seeing may not match what you're hearing.  We could give some choices about which voice gets visualized, but that adds some user interface complexity, and with everything involved, it's a bit of hit to the CPU.  Doesn't mean it can't be done, though.  We'll put it on the list and see if Rob has a conniption.

Copy and Paste - Yes, very much agreed.  It's already on our list as the ability to copy from one module to another module of the same type.  I'll add your vote.

Parallel Filter Bug? - I actually think this is not a bug.  Here's why...  When I went to verify this, I set up the stereo split filter routing (which you did correctly, btw), and instantiated filter 2 with the same filter as filter 1.  They definitely sounded different as you said, but I realized that filter 1 was already programmed with some different filter settings and modulation routing in the patch that I called up to test (Lead-Bass Chop Saw).  With filter 1 in focus, I selected MTRX on the Input column so I could view the mods going specifically to filter 1, and then I muted them all.  Finally, I double clicked on the parameters in Filter 1 to set them to default.  (I suppose a quicker way would be to change filter types and then recall the original filter type.)  At that point, the parallel filters sounded identical to me.  If you're getting different results after trying this let me know.

Organic and Acoustic - Yes!  I'm glad you notice that.  I'm a big fan of that vibe in my own sound design style, so I love to run a lot of random mods to different places, even if it's just for a subtle effect.  It gives the sound life!

Once again, thanks so much for taking the time to speak with us.  We appreciate the collaboration!

All the best,

Taiho

answered Apr 6, 2018 by taiho (9,320 points)
Further note on the LFO Unique Random issue...  I was hearing some artifacts when routing to osc tuning, but routing to filter cutoff seems okay (tested with high resonance).  The artifacts may be more of an issue frequency interaction in the oscillator, so I'll have to dig deeper.
I tried to replicate the filter split problem I was having and now it seems to work as it should. I don't know how that happened, I thought I replicated it 2 times earlier, but now it seems to work. I must have been doing something wrong.

A static, per voice random number modulator would be super useful. I think a static, alternating modulator is also really useful, i.e. a modulator that outputs a high value when triggered the first time, then a low value the next time, then high, and so on. Very useful for things like per-voice ping pong panning.

On the visual feedback, I can understand not wanting it if it would take up too much CPU. Would making it hideable/showable help that at all? As for which voice to display, a simple last-note priority for retriggering mode would be fine I think. Even just a simple 2 pixel line that goes up and down over a line marking zero would be fine. It would just be helpful for seeing the slopes and shapes of the wave you're using in real time, without having to go into the oscillator or something to see what the waves are like. Also useful to very quickly understand exactly how fast the rate of an LFO or the slope/time of an envelope stage is, especially if you're doing modulations that are audibly subtle.
All good stuff!

@Taiho, that's a very clever solution to a random number generator, though it would be a lot easier if we just had a Random Number generator block to do that, or maybe more simple to just have the ability to start the LFO in Freeze Mode.

@ipzid :  I love the alternating static modulator thing too... not limited to just high and low, but could be a feature of the LFO, so you can step through any shape (and some rate) with each note.

Regarding visual feedback, I agree that would be really helpful.   It's not a CPU problem, it just means that if you played multiple notes, the visual display would be for only one voice, so it would only be accurate for one voice at a time, but it would certainly give you insight into what is happening.   No conniption here ;)

Many thanks for keeping the forum lively, @ipzid!

Related questions

0 votes
1 answer
asked Feb 6, 2018 in Waverazor by andyd (160 points)
+1 vote
4 answers
0 votes
1 answer
0 votes
2 answers
asked Apr 17, 2017 in Waverazor by rob (14,780 points)
0 votes
1 answer
asked Apr 17, 2017 in Waverazor by bbrylow (160 points)
...