@CyberGen Oh, I never thought there is a third place where I have to turn it on. :grinning_face_with_sweat:

Nevertheless, everything works fine finally! Thanks a lot!

Excellent, it works perfectly!
Thanks @Lindon and @DabDab!

@Christoph-Hart
Here's an update about the actual status.

The JSON exception issue disappeared after I put all the file content into the script as an object. Now the plugin always opens, and the "plugin is not recognized" issue disappeared. Yay!

However, the neural stuff I am using still acts as bypassed in the exported plugin. In HISE it works properly, in the exported plugin it doesn't.

HISE_INCLUDE_RT_NEURAL=1 is used during the export as well.

Any tips for possible next steps?
Is there anyone who is facing the same issue, or who has a working setup with the same settings?

@Christoph-Hart
I could not solve the problem yet, but I have more details.

When I run the plugin in debug mode, I get this error message:

I tried the same with the example project (NeuralNetworkExample):

1. add script to create neural network, load the model layout weights (in this case model_tf.json), and load Tensorflow model
2. add the model into a ScriptNode FX in math.neural
3. export file to DLL and load it as a Hardcoded Master FX
4. export plugin

I got the same error message as above. The JSON files pass JSON Validation, and the models work properly inside HISE, they are sounding in the way as expected.

Do you have any ideas on which direction should I move forward?
Your help would be greatly appreciated. :folded_hands:

Hi everyone,

I'm not sure if it is a bug or just my lack of experience, but since I updated my MacOS to Sequoia, I had to update Hise to 4.1 and also Xcode to 16.2 (16.3 did not work with HISE 4.1 somehow), the plugin I built cannot be used with Intel machines. It builds only an arm64-compatible plugin, which is my native architecture.

I tried several things:

• Exporting the plugin through the Autogenerated Projucer file, with all the right architectures checked
• Change the Build Architectures in Xcode (however, it was already on the right setting to support "Standard Architectures (Apple Silicon, Intel)")
• Export a normal JUCE plugin not made in HISE, which was fine. In this case, it worked with Intel as well.

No luck so far with HISE and Intel. Do you have any idea what the issue is here? Any support would be greatly appreciated.

Best,
Gábor

@d-healey I will have a look, but I guess this option will not work at this time due to time pressure. Thanks anyway, at least I know the limitations.

Thanks @d-healey !

Oh damn. How can I do that?

Hi everyone,

I would like to ask for a little help.

I am currently working on a preset browser, and I changed the look of the dialog with drawPresetBrowserDialog.

But I can not change the position of the text input. My goal is to follow the changes of obj.labelArea, and somehow make it centered, but it doesn't work.

My code:

laf.registerFunction("drawPresetBrowserDialog", function(g, obj)
{	
                obj.area = [obj.area[0] + 75.0, obj.area[1], obj.area[2] -150.0, obj.area[3]];
		obj.labelArea = [obj.labelArea[0] + 40.0, obj.labelArea[1], obj.labelArea[2] -80.0, obj.labelArea[3]];
	
		g.fillAll(0x00000000);
		g.setColour(0xF0CECECE);
		g.fillRoundedRectangle(obj.area, 5.0);
		g.setColour(0x76000000);
		g.fillRoundedRectangle(obj.labelArea, 2.5);
		g.setColour(0xFF222222);
		g.drawAlignedText(obj.text, [-32, 160, 350, 4], "centred");
});

Is there any workaround for this?

Thanks!
Gábor

@Dan-Korneff Thanks, Dan! Let me know if it will work on your end.

@Christoph-Hart
Here's an update about the actual status.

The JSON exception issue disappeared after I put all the file content into the script as an object. Now the plugin always opens, and the "plugin is not recognized" issue disappeared. Yay!

However, the neural stuff I am using still acts as bypassed in the exported plugin. In HISE it works properly, in the exported plugin it doesn't.

HISE_INCLUDE_RT_NEURAL=1 is used during the export as well.

Any tips for possible next steps?
Is there anyone who is facing the same issue, or who has a working setup with the same settings?

@Christoph-Hart
I could not solve the problem yet, but I have more details.

When I run the plugin in debug mode, I get this error message:

I tried the same with the example project (NeuralNetworkExample):

1. add script to create neural network, load the model layout weights (in this case model_tf.json), and load Tensorflow model
2. add the model into a ScriptNode FX in math.neural
3. export file to DLL and load it as a Hardcoded Master FX
4. export plugin

I got the same error message as above. The JSON files pass JSON Validation, and the models work properly inside HISE, they are sounding in the way as expected.

Do you have any ideas on which direction should I move forward?
Your help would be greatly appreciated. :folded_hands:

@Christoph-Hart
I've added this to the extra definitions, but it's still the same.

Oh, and I did the same with the Example Neural Network snippet, and I had the same issue there as well.

Hi everyone,

I'm currently building a plugin with a distortion neural network inside.

I used Tensorflow, and added my model in the same way as in the example snippet:

const var tf = Engine.createNeuralNetwork("ExampleNetwork");;
const var weights_tf = pythonRoot.getChildFile("model_tf.json").loadAsObject();
tf.loadTensorFlowModel(weights_tf);

then I load it in Scriptnode into the Math.neural node, export it as dll, and use it as a Hardcoded Master FX. Everything works fine in HISE.

However, as soon as I export my plugin, and load it into my DAW, the plugin is not recognized.

If I remove the script (see above), and export again, the plugin loads up, but the neural network sounds like it's bypassed (it's hardcoded, so it's weird). This way it acts the same way inside HISE as well (bypassed).
Then I added the code above again, it works well in HISE, but the plugin is not recognized again when it is exported.

What can be the problem here? Is there something I am missing?

Thanks!
Gábor