MAX Audio Classifier Demo

This flow illustrates how to identify sounds in short audio clips using a deep learning model. The model used in this example is the MAX Audio Classifier, which comes in the form of a node known as node-red-contrib-max-audio-classifier.

In the flow, sample usage of the node is shown, which accepts input in the form of audio WAV file data, sends it to an instance of the Audio Classifier model from the Model Asset eXchange, and then returns what is identified in msg.payload.

The featured node was created with help from the amazing node-red-nodegen package (Node-RED Node Generator) and if you're interested to learn more, the process of creating a node from a deep learning microservice was documented here.

More information from the flow's comment node:

This flow captures audio using your computer's microphone and sends it to the MAX Audio Classifier Model, which identifies what types of things are being heard.

To begin recording, click the button on the left side of the 'microphone' node. To stop, click the button again.

You should see a status message while the audio classifier node is processing the request, and then the output will appear in the right-hand sidebar, in the 'Debug' console.

Alternatively, you can upload a WAV file by double clicking the 'file inject' node. The output will be the same.

NOTE: This is only a demo of what this node can do. With a little experimentation, you will find many new ways to control the capture of input, and ways to process the response and display it as output.

[{"id":"6c24d4d6.57a15c","type":"tab","label":"MAX Audio Classifier Demo","disabled":false,"info":""},{"id":"a0f4732b.adb67","type":"audio-classifier","z":"6c24d4d6.57a15c","service":"d6b5216.e5535e","method":"predict","predict_audio":"","predict_audioType":"str","name":"","x":413,"y":262,"wires":[["efd9424f.a04e4"]]},{"id":"88a099b5.42dc18","type":"fileinject","z":"6c24d4d6.57a15c","name":"","x":131,"y":330,"wires":[["a0f4732b.adb67"]]},{"id":"efd9424f.a04e4","type":"debug","z":"6c24d4d6.57a15c","name":"","active":true,"tosidebar":true,"console":false,"tostatus":false,"complete":"false","x":659,"y":262,"wires":[]},{"id":"aed4fef4.194c9","type":"microphone","z":"6c24d4d6.57a15c","name":"","x":140.5,"y":153,"wires":[["a0f4732b.adb67"]]},{"id":"5298fd04.1ffb44","type":"comment","z":"6c24d4d6.57a15c","name":"This flow shows the usage of the MAX Audio Classifier node (double-click for more...)","info":"This flow captures audio using your \ncomputer's microphone and sends it to \nthe MAX Audio Classifier Model, which \nidentifies what types of things are \nbeing heard. \n\nTo begin recording, click the button on the\nleft side of the 'microphone' node. To stop,\nclick the button again. \n\nYou should see a status message while the \naudio classifier node is processing the\nrequest, and then the output will appear in\nthe right-hand sidebar, in the 'Debug' console.\n\nAlternatively, you can upload a WAV file\nby double clicking the 'file inject' node. The\noutput will be the same.\n\nNOTE: This is only a demo of what this node \ncan do. With a little experimentation, you\nwill find many new ways to control the \ncapture of input, and ways to process the \nresponse and display it as output.","x":401.5,"y":69,"wires":[]},{"id":"d6b5216.e5535e","type":"audio-classifier-service","z":"","host":"https://max-audio-classifier.max.us-south.containers.appdomain.cloud","name":"Cloud Instance (Experimental)"}]

Flow Info

Created 1 year, 7 months ago
Updated 1 year, 6 months ago
Rating: not yet rated

Owner

Node Types

Core
  • comment (x1)
  • debug (x1)
  • tab (x1)
Other

Tags

  • node-red-nodegen
  • model-asset-exchange
  • Model Asset eXchange
  • deep learning
  • artificial intelligence
  • audio classifier
  • sound
Copy this flow JSON to your clipboard and then import into Node-RED using the Import From > Clipboard (Ctrl-I) menu option