Error converting content: marked is not a function
- What a fucking match! The best in last 5 years! #cricket - 6 - 7:30 PM P/Berklee AI in Music Assignment 1. Done 1.5 hr. That was fun. 🍅🍅🍅 collapsed:: true - #system -forcing the learning. #loveit❤️ - Analysis {{renderer :wordcount_}} - Digital representation of music is fascinating mix of science and art. What we perceive is subjective so each representation is a different way to look at the underlying audio signal and gain insights into underlying audio without actually listening to it. Waveform is visually intuitive representation as it operates in time domain. Like ocean waves going rising up and down in intensity we can perceive the loudness of an audio track with variations in amplitude. A 5 minute long song will take us 5 minutes to listen to, whereas a waveform analysis gives us a quick snapshot view that can be grasped in a few seconds. Abstract representation of underlying phenomenon always limit what they represent. These limitations are key to bringing down the complexity to the level of human understanding. The simplicity of the audio waveform makes them palatable to our eyes but in this process we limit what richness of information contained in it. Encoding data in such useful representation is key to learning as well. To add more information to visual representation of audio, we could add frequency data of the signal in an understandable representation. Spectrograms are used for such representations. The core mathematical foundation pinning spectrograms is the Fourier transform of complex audio signals converting them from time domain to frequency domain by splitting the signals into it's constituent frequencies. Spectrogram gives us a 2D representation which combines both time and frequency domain in a single view. It's like a picture of the sound track (across time domain) showing it's underlying constituent frequencies and their intensities which are helpful for both trained human eyes and machine learning models. The fascinating human element of spectrogram analysis was the realization that our perception of pitch is logarithmic i.e it is not linear. Therefore, we plot frequency in log scale to see variations in pitch. - - Tinkering - 8:30 PM P/12 Week Legacy 2024 . Week 9 Chest cont - TODO P/Saie Birthday 2024 (33) planning - so I head - 'Sarah is very curiious what you will do for her birthday' - Saie's Friend