I just wanted to make a note on how helpful David Healey is in this community, and that his tireless efforts to upgrade everyone’s knowledge are noticed and appreciated!

Thank you, David :)

Happy New Year!

Christoph,

First off, I would like to thank you for this amazing development tool for music :)

I am quite happy to be taking the plunge and learning as much about coding in HISE every day since watching and following the David Healey tutorial video: "How to make a synth"

Now, my main goals with HISE involve making MIDI plugins; some to be released as free utilities, and others as a paid product.

In my tests it seems that HISE does not receive or pass (thru) some MIDI message types:
MIDI clock
MTC
MMC
SysEx
etc.

Then there is Program Change, which HISE can accept and use, but not generate or pass (thru).

My concern:
When using HISE to make MIDI plugins, blocking these messages may deny users access to certain features that are needed for the next plugin in the chain (often a VSTi).

Therefore, if I am to produce MIDI plugins I am apprehensive about blocking those messages, and I, myself as a MIDI plugin customer would not be happy if a plugin that I purchased were to deny the passage of those messages to the output stage.

Personally, I do a lot of orchestral template mockups in a system where the articulation switching of VSTis is managed by expression maps (in Dorico, Nuendo, Bidule and Reaper) which are changed via incoming program change messages. Currently, I can't actually use any of my HISE produced MIDI plugins within this system, as without PC messages, it blocks the articulation switching.

In addition, there are many VSTis that can utilize program change messages directly to change patches. Kontakt, Omnisphere and VSL for starters.

And from my observation, it seems that many electronic producers with outboard gear setups require MTC (MIDI time code) and/or MIDI clock to sync devices.

I also have use-cases for plugin designs involving MIDI clock, MTC and SysEx and would love to have access to them for making custom MIDI controller managers.

Given that, I would like to add a feature request of three different levels (sorted from lowest to highest complexity):

1. a function to pass all unused MIDI messages directly to the MIDI output (without any processing)
2. a method of allowing HISE users to get and set MIDI messages in raw MIDI hex (so we can roll our own)
3. a feature set in the Message class for these other MIDI message types, similar to the ones available for note on, note off and CC.

Thanks for your time,
Craig

@clevername27 Thank you very much for your time!
You have a lot of incite to provide and I found the instruction as well as conversation to be quite worthwhile. I would certainly sign up again for more!

@d-healey said in HISE Meet Up:

@HISEnberg said in HISE Meet Up:

I don't think you could do spectral morphing in realtime

https://www.zynaptiq.com/morph/
https://web.archive.org/web/20180705061655/http://www.hakenaudio.com/RealTimeMorph/

I find the Kyma to be rather convincing for smooth morphing:

I would love to hear this kind of thing used to interpolate between sample sets for articulations that have capabilities for continuous control of timbre:
bow position, fluttertongue, growl, stick position on percussion etc.

There are a few different algorithms to choose from for morphing on the Kyma, and I don't know the details, but I think these are precomputed FFTs for resynthesis.

I am not sure what is causing this, but sometimes when I open a project/preset I am not getting any MIDI input and when I check the settings the MIDI inputs have been turned off:

Whereas I am certain I have saved it like this:

Yes, you can send MIDI out from button clicks.
I am working on a series of MIDI plugins at the moment.

You need to enable MIDI out in settings:

When using UI components to generate MIDI notes, you need to use Synth.addNoteOn() and Synth.addNoteOff() on the callback to you component.

Here I have pads (made with panels) which play chords when you use a mouse click.
It turns the notes off when the mouse is up.

			// mouse callbacks for the pads -- 
			pad.setMouseCallback(function[unitSize](event)
			{
				var ps = this.data.pitchSet;
				var l = ps.length;			
				
				if (event.clicked)
				{
					var velocity = Math.round(event.y / unitSize * 127);							
					for (k = 0; k < l; k++)
						Synth.addNoteOn(1, ps[k], velocity, 0);									
				}
				
				if (event.mouseUp)
				{				
					for (k = 0; k < l; k++)
						Synth.addNoteOff(1, ps[k], 0);
				}				
			});		

To get the MIDI to output from the plugin,
as others have said you must use Message.sendToMidiOut();

To make sure it catches everything (and not some intermediate stage of MIDI),
I always place these functions in a script processor and container after all of my other containers:

And place the Message.sendToMidiOut() on each of the three MIDI callbacks:

@HISEnberg said in Let’s Build the “League of Newbies”:

@HISEnberg Done, I created the group chat. I guess anyone else who is interested can post here and we can add them in.

Add me :)
I'll check it out.

@d-healey It needs both ceil and floor to work, so I wrote an inline function to do the job:

inline function roundFix(value)
{
	local remainder = Math.fmod(value, 1.0);
	local roundedValue;
	remainder >= 0.5 ? roundedValue = Math.ceil(value) : roundedValue = Math.floor(value);
	return roundedValue;
};

const testRound = roundFix(4.49);
Console.print("test round: " + testRound);
const testRound1 = roundFix(4.5);
Console.print("test round: " + testRound1);

@d-healey said in Is there a method to rotate a path?:

@VirtualVirgin Instead of making a function to rotate a path, why not make a function to create a rotated path. You can pass in the data you would have used to create the initial path.

So based on your suggestion I came up with this and it seems to work:

inline function createRotatedPath(pathArray, angle, area)
{
    local rad = Math.toRadians(angle);
    local cosA = Math.cos(rad);
    local sinA = Math.sin(rad);
    
    local path = Content.createPath();
    
    local rotationCenterX = area[0] + (area[2] * 0.5);
    local rotationCenterY = area[1] + (area[3] * 0.5);
    
    for (i = 0; i < pathArray.length; i++)
    {
        local x = pathArray[i][0];
        local y = pathArray[i][1];
        
        local newX = (x - rotationCenterX) * cosA - (y - rotationCenterY) * sinA + rotationCenterX;
        local newY = (x - rotationCenterX) * sinA + (y - rotationCenterY) * cosA + rotationCenterY;
        
        if (i == 0)
            path.startNewSubPath(newX, newY);
        else
            path.lineTo(newX, newY);
    }
    
    path.closeSubPath();
    return path;
}

The pathArray is a 2D array of x,y points
i.e.
[[x,y],[x,y] etc..]

@d-healey I'm imagining that you answered that question before I even submitted it! Lightning fast response.

@ustk apparently it has to do with transferring an array reference. I tried this

local newValue = scribeData.clone();

instead of this

local newValue = scribeData;

And now it works (the old value is properly being retrieved)...
But I am not quite sure why as I am lost in the puzzle of how the array references operate in this case.

I have been scratching my head for a couple of hours on this, and I just don't understand!

I am trying to use the panel with the undo system, and here I am getting the old panel value before setting the new panel value, but my old value is always identical to the new value!!!!

The line below the Console prints is

NotationPanel.setPanelValueWithUndo(oldValue, newValue , "set");

This line is the only place in the entire project that is setting the panel value.
There are no other calls to .setPanelValueWithUndo or .setValue.

How is it at all possible that NotationPanel.getValue is always returning the new value before it is ever even set?

@d-healey Actually, I think I just found the error:

local MPN = direction == "up" ? noteObject.y[0] : noteobject.y[noteObject.y.length -1];

vs.

local MPN = direction == "up" ? noteObject.y[0] : noteObject.y[noteObject.y.length -1];

I forgot the camelCase on noteObject for the false branch :P

@d-healey I think this project is a bit too involved to make a snippet. Maybe I can try later. Here is the variable in question a little earlier in the chain, and it is also perplexing:

The variable "MPN" is being set here from a ternary and the print just before that clearly shows that noteObject.y[0] has a value: "F5"
And it shows that the direction is "up".

Then again, it goes on to print a value that it says is "undefined".
A bit perplexing... seems contradictory.

And now it is printing the result while still saying that the function doesn't exist!!!

What is going on???

This function has been doing the job contently for some time,
but now it doesn't...

Logically, this should be getting the "B" from "B6".
How would this not work?

@bendurso Thanks!

@bendurso Ah! I've never doe that before. How do I "get" the container reference?

const TestContainer = ???

@ulrik Is that supposed to work with any container?

I just tried a test and I get this: