Data Analytics in Music and Hydrologic Modelling

 

Data analytics in Music and Hydrologic Modelling

The ability to leverage hobbies and a career, where the two work synergistically to inform each other offers benefits such as skill development, creativity, improved psychological well-being, and a more fulfilling, well-rounded life. A hobby can provide opportunities for personal growth that can translate into career development and greater job satisfaction.

Music production has been one of my cherished hobbies. Over time, it has ben interesting to discover how stochastic modelling, including time series modelling can be leveraged with music production.

Both music production and hydrologic modelling entail data analytics, based on signal processing techniques such as analyzing and manipulating the hydrologic inputs, or (in music audio signal) to understand its characteristics and quality. For instance, data filtering is used to remove unwanted frequencies, spectral analysis visualizes the signal's frequency content over time (often via a spectrogram), and pulse response (Cf. the unit hydrograph) measures how a system reacts to a very short sound wave (rainstorm in hydrology), revealing its acoustic (hydrologic response) properties. These methods help evaluate the audio chain, improve the sound quality by identifying and correcting issues, and inform data analytics for applications like audio classification or room acoustics analysis. 

1. Data Analytics

What it is: Like hydrologic analysis, data analytics involves the process of examining recorded audio data to extract meaningful information, identify patterns, and gain insights into the signal's characteristics. Examples here are timeseries analysis and modelling.

How it's applied:

Quality Assessment: Analyzing spectral and time-domain data to identify issues in an audio chain.

Audio Classification: Developing models to categorize different types of environmental sounds using features extracted from spectral analysis.

Acoustic Analysis: Using the pulse response to understand how a room's acoustics influence the sound.

2. Filtering

What it is: A signal processing technique that modifies the signal's frequency content, allowing some frequencies to pass through while attenuating or blocking others. This method is used in hydrometeorological time series analysis for frequency characterization.

How it's applied:

Noise Reduction: Applying low-pass filters to remove high-frequency noise and high-pass filters to eliminate unwanted low-frequency drift.

Antialiasing: A low-pass filter is used before an analog-to-digital converter (ADC) to prevent unwanted artifacts from high frequencies. Data aliasing helps to identify short climatic fluctuations that can erroneously be interpreted as climate change signals.

3. Spectral Analysis (Time Series Modelling)

What it is: A method that decomposes a complex signal into its constituent frequencies, revealing the amplitude (or power) of each frequency component at different points in time. The method can be used in frequency domain analysis to identification of hydrometeorological data periodicities such as those caused atmospheric quasi-biennial oscillations and sunspot cycles.

How it's applied:

Spectrograms: A time-frequency (T-F) representation that displays the evolution of the signal's spectrum over time, showing where different frequencies are present.

Diagnosing Issues: Identifying problematic frequencies in a mix or recording, such as resonant frequencies caused by room acoustics.

Understanding Musical Content: Analyzing which musical tones are playing at which moments and how they change.

4. Pulse Response (Cf. the Unit hydrograph)

What it is: The output of a system when it is subjected to a very short, sharp input (an impulse).

How it's applied:

In hydrology the impulse response function is used in rainfall-runoff routing process.

Acoustic Characterization: The impulse response of a room or a piece of audio equipment provides detailed information about its frequency response and how it affects the sound.

System Identification: It helps to understand the inherent properties of the recording equipment or the acoustic environment, such as reverberation time.

Comparison: By comparing the recorded pulse response of a system to a known reference, one can identify differences and potential flaws in the audio chain.