Some basic questions

Levitanus

Can't you use an envelope for playing attack samples?

Afraid, not. Or i have to have 3 envelopes, 0.7 sec for mixing with attack, 0,5 with gliss, 0.3 with leg
And Every time i want to change the main mixing properties should be edited three times...
Or i have poor logic)

in KSP I would make something like this:

on init
  declare constant $attack_time := 700000
  declare constant $leg_time := 300000
  declare constant $gliss_time := 500000
  declare constant $fade_time := 100000
end on
on note
  if (<first note trigger>)
    <attack groups allowing>
    $attack_id :=    play_note ()
    <sus groups allowing>
    $new_id := play_note()
    change_vol($new_id,0) {if i not mistake}
    wait ($attack_time - $fade_time)
    change_vol ($new_id, <0dB lvl>)
    fade_out ($attack_id,$fade_time,1)
    fade_in ($new_id,$fade_time)
  end if
{+some other stuff is not included
The same with leg and gliss}
end on

d.healey

I think you could do it with a table envelope. Or use 3 samplers

Levitanus

Or use 3 samplers

for sustain? The better to ave 3 envelopes, i think, but the problem of wasting time is here...

d.healey

It really depends on how you're laying out other things like RRs and Dynamics but I see no reason why you could have 4 samplers (3 was a typo), one for attacks, one for sustains, one for legatos, and one for gliss. They could each have their own envelopes. Or you can one sampler with 3 envelopes and change the attack/release time on the fly.

Levitanus

I have dedicated samplers for every phase. But The point is I need different attack delay (not attack time) on sustain depends on attack, leg, or gliss triggering.
And it's quite simple example... May be i'm trying to make things too complicated...

d.healey

What do you mean by attack delay? - I think I'm not understanding correctly

Levitanus

look. If I use delay event function with 200 ms delay, and take a note shorter than 200 ms, i'll get stucking note, because note off came earlier than note has started.
Even this code doesn't works in this case

var exit = 0;
function onNoteOn()
{
	exit = 0;
	if (exit == 0){
	Message.delayEvent(Engine.getSamplesForMilliSeconds(200));
	}
}
function onNoteOff()
{
	exit = 1;
}

And such problems will be worse with midi-routing instead of global events. So the better practise on my taste is take everything we need at one moment, just "mute" some things for a while. Than fade in. But if these notes are off yet they would not fade in because of they are not already exists) Simple, transparent logic. But in this case on every mixing event we need fade_in delay. It can be produced by the table envelope. But for several different mixing types we need different delay with the same fade_in (attack) time.
So just one envelope with changing attack time will not work both for 300ms and for 700ms, attack (fade_in) time will be very different in this case.
And it's just note events. More complicated logic - more request for exact timing...

Levitanus

well. lets try to understand how the timer works.
There's audio thread,which is executed one time per sample, timer function (callback) gets worldclock, or samples from start and is called every amount of samples. If i not mistake...
I don't think KSP behind logic works different. Just need to understand how does it came back to the callback.
I was thinking of null while loop, but it's terminated by too many iterations, i think...
Maybe there's a little bit more elegant solution? I really afraid of making garbage from the timer callback... especcially if some unpredictable usage will be produced

Levitanus

AAAA) getUptime!
I was confused by the "seconds" word in the API, but it uses ms

d.healey

Yeah that confused me too :)

Levitanus

emmm.... The same problem with while loop. Even don't know who could expect)))
Well, it seems, the unique timer is the best solution at the moment)

Christoph Hart

A key difference to KSP is that there is no wait() instruction. Instead you delay events using Message.delayEvent(). The problem with stuck notes can be solved by delaying the note off command with the same amount (it's hardly noticeable on note off commands).

Don't use the Engine.createTimerObject() for audio stuff. This is running in the GUI thread and has absolutely no predictability or accuracy.
There are scripting callbacks for adding volume fades. Check out Synth.addVolumeFade().
the onTimer callback runs in the audio thread and has sample accuracy.

Thanks for the hint, I'll fix the getUptime doc in the API...

Levitanus

Ok, I think, the last question in this direction))
How about JUCE API? Can it work inside the script processor?

Levitanus

can't find are reg global, and if not, 32 per plugin or 32 per script?

Christoph Hart

32 per namespace:

namespace n1
{
    reg n1_1 = 1;
    // ...
    reg n1_32 = 32;
}

namespace n2
{
    reg n2_1 = 1;
    // ...
    reg n2_32 = 32;
}

Levitanus

Very interesting)))
So i can use 32 pieces of
reg atck_var
and 32 of
reg sus_var &

And they are global, am i right?

Another question. I placed Synth.noteOffByEventId(atck_id); into timer callback, as well as to the note_off callback. Now, pretty predictable i receive error in console about missing event by timer or by noteOff callback depends on situation. Is it bad? Or I can eat it as usefull information and newer mind?

and more, more questions ))))
When does envelope work? I know that once triggered by the voice KONTAKT envelope works as it was before triggering, even if it has been changed during the note playback.
For example, if I move some points in noteOn, will it "eat" them respectivelly to the current note and not for any other already triggered?
or it will work with all currentely produced samples? Or new values will be produced only with next notes, or with artifically played with a bit of delay?

P.S. and.. I have been confused a little by the dates od requests, so can;t understand actual things. Is timer global for the plugin, or local for each script?

Levitanus

Well I have almost done the "Hello world", but can't find how to set persistence of wigets... Or they will be saved automatically (as fbx)?

Christoph Hart

Widgets are persistent by default. If you don't want this, you can call

widget.set("saveInPreset", false);

Levitanus

Thanks! But I can't get the widget value by simple:

modwheel = Content.addKnob("modwheel", 346, 110);// [JSON modwheel]
Content.setPropertiesFromJSON("modwheel", {
  "max": 127,
  "stepSize": "1"
});
// [/JSON modwheel]
modwheel.showControl(false);
Globals.MW = modwheel.getValue();

What should be here else?

Christoph Hart

The persistent data is not available in the onInit callback. This happens behind the scenes:

The persistent widget values are stored.
The script gets compiled.
The persistent widget values are restored.
For all persistent widgets, the onControl callback gets executed.

In your case, you need to move the

Globals.MW = modwheel.getValue();

line in the onControl callback.

Some basic questions

15

1.8k

10.2k

89.1k