Create a Google Chrome extension that detects sounds like Shazam

Job ID: 36832546

Budget: $30 – $250 USD

You're already familiar with an app on your phone called Shazam. It lets the user detect what song they're listening to.

You'll create a Chrome extension that does something similar. It will be always listening, and it will parse 5-6 second clips of ambient audio to determine if a specific audio signature has been detected. If nothing is detected, the audio is discarded and the process starts all over again, and keeps running.

You'll use the algorithm Fast Fourier Transform to accomplish this. There are libraries to do it. You'll want to create a couple hashes for testing, like dog barking or something like that, using a canned audio file. The key is, the initial processing has to happen client side/browser side (not cloud/server side).

The extension will simply return the confidence level if and when the pre-determined sound is detected. For example, if it's listening, and "hears" dog barking, it will return (the algorithm does) the confidence level (e.g. 74%). Just show that confidence level on the extension in some way, on the icon, a popup, whatever. This is a good sound to use to create the signature hash, and then test to ensure that the extension properly detects it. https://freesound.org/people/Anton/sounds/157322/

This is specific and nuanced. Let me know if you have any questions.

This needs to function without the cloud. It needs to be entirely local and needs to work on Chromebooks. It needs to be a Chrome extension without any sort of companion app for Mac or PC. All in the browser.
Related categories: JavaScript Python Algorithm Analysis