This repository contains a simple Rust tool for estimating the Beats Per Minute (BPM) of an audio file. The tool processes the audio file, calculates its spectrogram, detects peaks, and utilizes histogram analysis to provide a BPM estimation.
I built it to learn a little bit about digital signal processing.
Estimating BPM (Beats Per Minute) in music is a fundamental task in audio processing and analysis. This tool uses Rust to provide a basic yet effective implementation for estimating BPM from audio files.
A spectrogram is a visual representation of the frequencies present in a signal over time. In the context of audio, it helps us see how different frequencies contribute to the overall sound. Think of it as a musical landscape where the x-axis represents time, the y-axis represents frequency, and the color or intensity represents the amplitude (loudness) of each frequency at a given moment.
To create a spectrogram, we use a mathematical tool called the Fast Fourier Transform (FFT). The FFT takes a snapshot of the audio at small intervals, converts it from the time domain to the frequency domain, and helps identify which frequencies are present in that snapshot.
Additionally, we use a Hamming window to smooth out the edges of these snapshots. The windowing process helps prevent a phenomenon known as spectral leakage, which can introduce artifacts into the spectrogram.
Peaks in the context of audio represent moments where certain frequencies are particularly strong or pronounced. In our case, we're interested in detecting peaks in the spectrogram that correspond to beats or rhythmic elements in the music.
We apply a simple thresholding approach to identify peaks. If the amplitude of a particular frequency at a given time exceeds a certain threshold, we consider it a peak. These peaks help us locate significant moments in the audio where beats are likely to occur.
A histogram is like a bar chart that shows the distribution of values within a certain range. In our case, we're interested in the distribution of time intervals between detected peaks.
We calculate the time intervals between consecutive peaks and create a histogram to see which intervals occur most frequently. The idea is that the most common interval likely corresponds to the underlying tempo of the music.
By analyzing this histogram, we can identify the dominant tempo and calculate the Beats Per Minute (BPM) based on the average time interval between peaks. This approach provides a robust estimation even if the tempo of the music varies.
These algorithms work together to analyze the audio and provide an estimate of the BPM, helping us understand the rhythm and tempo of the music.
-
Install Rust:
Ensure that you have Rust installed on your system. You can install it by following the instructions on the official Rust website.
-
Clone the Repository:
git clone https://github.com/simpajj/audio-bpm-estimation.git cd audio-bpm-estimation -
Build and Run:
cargo build --release cargo run --release -- <path_to_your_audio_file.wav>
Replace <path_to_your_audio_file.wav> with the actual path to your audio file in WAV format.
The tool will process the audio file and output the estimated BPM.