dowser

dowser is (presently) a quick and simple low-level utility for performing spectral analysis of sound files.

usage

dowser [in] [out] <options>

options include:

• --minhz, lower frequency bound of analysis, in hz
• --maxhz, upper frequency bound of analysis, in hz
• --minmag, peak magnitude threshold
• --maxpeaks, max number of peaks per frame

the output is a supercollider script, defining a list of dictionaries. each dictionary contains data for a single spectral analysis frame.

per-frame measurements include:

• peaks: list of events with (hz, mag) keys, listing spectral peaks for the given frame
• papr: peak-to-average-power ratio, a measure of "tonal-ness"
• flatness: AKA weiner entropy, geometric mean / arithmetic mean. another measure of "tonalness"
• centroid: spectral centroid, a measure of "brightness"
• fluxPositive: spectral flux, in a positive direction
• fluxNegative: spectral flux, in a negative direction

requirements

cmake and a suitable c++ compilation toolset (requires c++17)

build

1. fetch dependencies (JUCE, which is lame but so it goes):

git submodule update --init --recursive

2. create build folder:

mkdir cmake-build-release

3. configure and build:

cd cmake-build-release

cmake -DCMAKE_BUILD_TYPE=Release .. (note the ..)

cmake --build .

the freshly compiled executable should then be located at: cmake_build_release/dowser_artefacts/Release/dowser (or .../dowser.exe on windows)

TODO

• other output formats (csv, python, etc)
• expose more parameters (hz range, max peaks per frame, FFT config)
• more analyses? (crest factor? autocorrelation peaks?)
• drag/drop GUI option
• maybe some higher-level analysis (e.g. clustering)