wav2csv

Converts a WAV recording into a frequency/SPL/phase CSV suitable for use as a measurement input to autoeq or roomeq.

Synopsis

wav2csv <INPUT> [OPTIONS]

Arguments

Argument	Required	Description
INPUT	Yes	Input WAV file

Flags

Flag	Default	Description
-o / --output	(input with .csv)	Output CSV file path
-n / --num-points	2000	Number of frequency points in the output
--min-freq	20.0	Minimum frequency (Hz)
--max-freq	20000.0	Maximum frequency (Hz)
--fft-size	auto	FFT size (default chosen from file length)
--overlap	0.5	Window overlap ratio for Welch’s method (0.0–1.0)
--single-fft	off	Use a single FFT instead of Welch’s averaged method
--pink-compensation	off	Apply −3 dB/octave compensation (for log sweeps)
--no-window	off	Use rectangular window instead of Hann

Choosing the Right Mode

The correct flag combination depends on the type of signal in the WAV file:

Signal type	Recommended flags
Music / stationary noise	(defaults — Welch’s method with Hann window)
Log sweep (room measurement)	--single-fft --pink-compensation --no-window
Impulse response	--single-fft

Output Format

The CSV has three columns:

freq,spl,phase
20.0,68.3,12.5
25.1,71.0,9.8
...
20000.0,55.2,-87.4

This matches the format accepted by both autoeq --curve and roomeq speaker measurement paths.

Examples

Analyze a room measurement sweep:

wav2csv room-sweep.wav --single-fft --pink-compensation --no-window
# writes room-sweep.csv

Analyze with a custom output path and frequency range:

wav2csv measurement.wav -o speaker-left.csv --min-freq 50 --max-freq 16000

Convert an impulse response:

wav2csv ir.wav --single-fft -o ir.csv

Typical Workflow

# 1. Record a log sweep with generate-audio-tests or your measurement software
generate-audio-tests --signals sweep --channels 1 --sample-rates 48000 --bits 24

# 2. Play the sweep through your speaker and record the microphone input
#    (use SotF Recording screen or any DAW)

# 3. Convert the recording to CSV
wav2csv recorded-sweep.wav --single-fft --pink-compensation --no-window

# 4. Feed the CSV into autoeq or roomeq
autoeq --curve recorded-sweep.csv --loss speaker-flat -n 7