Hush Noise Reduction "ML Model not Loaded"
Title
Hush Noise Reduction "ML Model not Loaded"
What do you expect to happen when you run the script/macro?
I am trying to create a Macro for Hush Noise Reduction where it will call up the plugin and render. Simple enough. Hush opens, but throws an error at the render action. "ML model not loaded. Wait for model to finish loading before previewing or rendering audio." Assuming the program needs time to load an ML model, is there a way in Soundflow to introduce a delay between actions?
Are you seeing an error?
ML model not loaded. Wait for the model to finsih loading before previewing or rendering audio.
Hush Noise Reduction AudioSuite Plugin will trying to render.
What happens when you run this script?
Hush Noise Reduction opens and sets parameters properly. But Hush give me the "ML model not loaded" error at the render action step.
How were you running this script?
I used a Stream Deck button
How important is this issue to you?
4
Details
{
"inputExpected": "I am trying to create a Macro for Hush Noise Reduction where it will call up the plugin and render. Simple enough. Hush opens, but throws an error at the render action. \"ML model not loaded. Wait for model to finish loading before previewing or rendering audio.\" Assuming the program needs time to load an ML model, is there a way in Soundflow to introduce a delay between actions?",
"inputIsError": true,
"inputError": "ML model not loaded. Wait for the model to finsih loading before previewing or rendering audio.\n\nHush Noise Reduction AudioSuite Plugin will trying to render.",
"inputWhatHappens": "Hush Noise Reduction opens and sets parameters properly. But Hush give me the \"ML model not loaded\" error at the render action step.",
"inputHowRun": {
"key": "-MpfwmPg-2Sb-HxHQAff",
"title": "I used a Stream Deck button"
},
"inputImportance": 4,
"inputTitle": "Hush Noise Reduction \"ML Model not Loaded\""
}
Source
//Macro converted to script
sf.ui.proTools.audioSuiteOpenPlugin({
category: "Noise Reduction",
name: "Hush Mix",
});
sf.ui.proTools.firstAudioSuiteWindow.audioSuiteSetOptions({
processingInputMode: "ClipByClip",
processingOutputMode: "CreateIndividualFiles",
});
sf.ui.proTools.firstAudioSuiteWindow.audioSuiteRender();
Links
User UID: V4pxjelonDSg9cwaZeo1xHQOnY63
Feedback Key: sffeedback:V4pxjelonDSg9cwaZeo1xHQOnY63:-O-MmdWmKiOYKso7e-bA
Feedback ZIP: 6/KtF0KPyAyf5LJLNb2BSKbHtMjXFj9BFoG7ba0dpD88Z/3Ew5ehuV3AZIG/mPmBLOk2ouY6wVmxFZzZ/dMNwsuJUmZRFiFRMUdyNGi6qwBh0Wb5T6l4tkYDi3EB6+JdjpvEuU8crecJwegswmt0R8S7lv7j3MaPmic4E7W3Rf1+JCqVZsES2432RDTiDfYWAdhNfeywe0rYyHBe87J388KXxzvMFYr57y4pvGznoM9TzP0lERrJnxonz2ocDlgHBk54GhMOinRmG0yjuXF9wz4LrBhxAfL0krOqdmnUeh5kcwqjIk2K1WT+eYmGqE0nGyEalFWMMMgiitsrVzB152hggqdVpWVTPrVu2d8K4z4=
- BBen Rauscher @Ben_Rauscher
This would be great if it could work.
Chris Shaw @Chris_Shaw2024-08-08 18:50:56.381ZIs there a progress bar or other indicator that the ML model is loading?
- BBen Rauscher @Ben_Rauscher
No. In Keyboard Maestro, the fix would be to wait for the ML Loading graphic to disappear via screen recording. Don’t know if SF has the capability to detect items in that manner.
- Chris Shaw @Chris_Shaw2024-08-08 20:57:51.118Z
Unfortunately SF doesn't support checking in that manner.
Does the ML model only load when Hush window is first opened or each time yo render?
If the ML only loads when the AS window is opened you could check if the AS window is already opened at the top of your script then put a wait at line 8 if it isn't.
- Open the hush AS window, create a new macro and add a 'Click UI Element' action.
- Click the "Pick" button, switch over to PT and click the title bar of the Hush window.
- Go back to the editor and click the "First AudioSuite Window" text which will open a popup window with different options (I'm using the EQ3 1-band as an example):
-
That second option should show you the window name of the Hush plugin.
-
Use that window name to replace the window name in line 2
-
adjust the time in line 14 if needed to wait for the ML to load
Here's the script:
// Check if Hush is already open const isHushOpen = sf.ui.proTools.windows.whoseTitle.is("Audio Suite: EQ3 1-Band").first.exists // If not open it if (!isHushOpen) { sf.ui.proTools.audioSuiteOpenPlugin({ category: "Noise Reduction", name: "Hush Mix", }); // Wait for ML to load (Adjust the time to allow for ML to load) sf.wait({ intervalMs: 500 }) } // Hush is open - continue sf.ui.proTools.firstAudioSuiteWindow.audioSuiteSetOptions({ processingInputMode: "ClipByClip", processingOutputMode: "CreateIndividualFiles", }); sf.ui.proTools.firstAudioSuiteWindow.audioSuiteRender();
- BIn reply toJason_Shablik⬆:Ben Rauscher @Ben_Rauscher
The ML message displays every time the Hush audio suite window is opened. Never bothered to time it, but the result would probably be different for everyone - especially silicon users. Without a modal or screen capture, a set wait interval would indeed be the only method to get it working. Thanks for trying.
- BIn reply toJason_Shablik⬆:Ben Rauscher @Ben_Rauscher
One thing that would help Hush Pro function smoother, at least on Intels, is the ability to set Mono mode from the Set Audiosuite Options macro. Rendering while in Multi-Mono causes the plugin to process all selected tracks simultaneously, which can bog down the system and / or stop out the plugin. In Mono mode, Hush will process the clips one at a time. So having the macro option to set this would be great for non-coders.