Python implementation for calculating and analyzing Wow & Flutter (W&F) in audio recordings according to the AES6-2008 standard (compatible with IEC/DIN variants). This tool follows the preferred Two Sigma method, as described in the document:
📄 Wow and Flutter Measurement using Multi-Instrument
This script was created for educational purposes, in the absence of publicly available code implementing precise W&F measurements as defined by international standards.
⚠️ If you're only interested in using the tool (without diving into the technical details), you can download the standalone Windows executable from thedist/folder. No Python environment required.
Comparison of wnf_meter.py results with Virtins Multi-Instrument using a 3150 Hz test tone recorded from a Reloop RP-4000MK2 turntable.
wnf_meter.py: Main script that performs Wow & Flutter analysis in 7 clear steps, outputting both unweighted and weighted 2σ (Two Sigma) results.weighting_filter.py: Generates the IIR filter coefficients that implement the AES6 weighting curve, designed as a simple asymmetric bandpass filter.weighting_filter_a.csv,weighting_filter_b.csv: Coefficients used to apply the weighting filter in the analysis script.assets/: Contains sample reference audio files for testing, including synthetic modulated tones and a test signal captured from a Reloop RP-4000MK2 turntable.
The following table compares the theoretical weighted W&F values for the synthetic test files (as defined in the Virtins guide) with the measured values obtained using this tool:
| Test File | Modulation | Theoretical WnF | Measured WnF |
|---|---|---|---|
| 3150Hz.wav | none | 0.00000% | 0.0000% |
| 3150HzModulatedBy4HzAt0.00997Percent0.315Hz.wav | 4 Hz | 0.00997% | 0.0100% |
| 3150HzModulatedBy4HzAt0.00997Percent0.315Hz.wav | 4 Hz | 0.0997% | 0.0997% |
| 3150HzModulatedBy4HzAt0.0997Percent3.15Hz.wav | 4 Hz | 0.997% | 0.9969% |
| 3150HzModulatedBy4HzAt9.97Percent315Hz.wav | 4 Hz | 9.97% | 9.9687% |
| 3150HzModulatedBy0.8HzAt0.0499Percent3.15Hz.wav | 0.8 Hz | 0.0499% | 0.0567% |
| 3150HzModulatedBy20HzAt0.0506Percent3.15Hz.wav | 20 Hz | 0.0506% | 0.0496% |
| 3150HzModulatedBy0.2HzAt0.003Percent3.15Hz.wav | 0.2 Hz | 0.0030% | 0.0035% |
Note: You can download the full reference test set from the links provided in the Virtins document.
The weighting_filter.py script implements a straightforward analog-inspired design using bilinear-transformed high-pass and low-pass stages, resulting in an IIR bandpass filter that conforms to the tolerance envelope defined in AES6. The filter is designed to operate at a sample rate of 2 kHz, which is sufficient for the frequency range of interest in W&F analysis (typically <200 Hz).
Users can extend this script to design and validate more complex filters using the same framework.
This project uses standard scientific Python libraries:
numpyscipymatplotlibtkinter(for file selection dialog)
To install them:
pip install numpy scipy matplotlib