The AI Show: Ep 55 | Audio analytics with Azure Automated ML

Serge Retkowsky is back to demonstrate how audio processing can consist of extracting audio signal information into spectrograms (time vs frequency vs Db) images that we can use to build a custom vision model with Azure. We'll also extract some audio components and use a generic classification model with Azure ML and its AutoML features.

Jump to:
00:00 Welcome to the AI Show
01:27 Intro to Audio analytics with Azure Automated ML
03:02 Fast Fourier Transform (FFT) - Time Domain Signal to Frequency Domain Signal
05:10 Spectogram of how the spectrum of of frequencies vary over time
07:50 Demo - Music Genre Classification
13:15 Creating a Chromagram
14:05 Spectograms generation
15:04 Music Genre Classification with Automated ML for Images
16:34 Calling AutoML CV Model
22:42 Demo 2
25:29 Learn more

Recommended resources:
AutoML for Images doc https://aka.ms/AIShow/AutoML/AzureML
AutoML for Images Algorithms https://aka.ms/AIShow/AutoMLImages/Al...
AutoML for Images Tutorial https://aka.ms/AIShow/AutoMLImages/Tu...
Create a Free account (Azure) https://aka.ms/aishow-seth-azurefree

Connect with Seth   / sethjuarez  

Don't miss new episodes, subscribe to the AI Show https://aka.ms/AIShowsubscribe
AI Show Playlist https://aka.ms/AIShowPlaylist

Join us every other Friday, for an AI Show livestream on Learn TV and YouTube https://aka.ms/LearnTV - https://aka.ms/AIShowLive