spyctrum

My attempt to create a spectrum analyser.

---

What is Spyctrum?

Spyctrum is still in active development! See the Roadmap for details on the current progress.

Spyctrum is intended to be a simple spectrum analyser written in Python. My motivation for writing it lies primarily in music, as a tool to aid in "reverse-engineering" music, however I want the project to remain useful for other purposes, including spectral analysis of scientific data and visualisation of music or other signals.

Spyctrum primarily uses numpy to perform Fourier analysis on small chunks of your input signal. This allows you to examine the frequencies present in a small time range, which, if you are musically-motivated, should allow you to pick out the dominant notes and associated harmonics in an audio signal to help in transcribing music. If you struggle to recognise musical notes by ear, this could be invaluable in learning to play a piece of music you can't find suitable sheet music for.

Supported Formats

The scipy module is used for converting audio data to numpy arrays. This by default can only read .WAV files, but if you are only intending to read .WAV files you will not require any external software -- numpy and scipy will be sufficient. In order to provide support for other formats, ffmpeg is used to convert tracks into a format scipy can read. This can be done either in-memory or using tempfiles. This means that spyctrum should support any format ffmpeg supports, including most popular formats, such as MP3, OGG, AAC, FLAC, etc.

Roadmap

• Implement reading of audio files
• Write a proper README about the project
• Fourier analysis functions
  • Frequencies corresponding to musical notes This will be done at the histogram level by definining appropriate bins
  • Conversion between frequencies and musical notes
  • Variable chunk size
  • Variable chunk position
  • Apply window functions to a chunk
• Graphing functions
  • Musical notes on x-axis
  • Linear amplitude
  • Logarithmic amplitude (dB scale?)
• Peak identification
• Be useable as a module
• Playback from array
• Write a GUI
  • Graphs (linear and log amplitude)
  • Seekbar to select point of interest
  • Playback with live graphs
  • Mouse over bar on graph highlights the same note at different octaves
• Customisable spectrum animations