A frequency-response graph can reveal the shape of a sound, but it does not automatically reveal perceived loudness. Loudness depends on the sound level at each frequency, the way hearing groups nearby frequencies, and the listening field. Stationary Zwicker loudness combines those ideas into a total loudness result in sones and a detailed pattern in sone per Bark.
What does Zwicker loudness calculate?
The stationary Zwicker method starts with an absolute spectrum: usually one-third-octave sound-pressure levels in dB SPL. It applies hearing-related corrections, forms excitation across critical bands, accounts for level-dependent masking, and integrates the resulting specific-loudness pattern. The total is reported in sones.
One sone is conventionally the perceived loudness of a 1 kHz tone at 40 phons. A useful rule of thumb is that doubling sones corresponds to a large, roughly doubling change in perceived loudness. Phons are a loudness-level scale referenced to a 1 kHz tone; this calculator derives them from the calculated total loudness.
What input does a stationary calculation need?
The most defensible input is a measured, calibrated set of 28 one-third-octave band levels from 25 Hz to 12.5 kHz. The calculator asks whether the displayed chart represents that kind of absolute SPL data, a narrowband SPL trace, a transfer-function estimate, or an uncalibrated relative response.
- Calibrated one-third-octave SPL: the direct input path for stationary Zwicker analysis.
- Narrowband SPL trace: can be sampled at third-octave centres, but this is an estimate rather than energy-summing the original signal into bands.
- Transfer function: can be reconstructed only when you declare an absolute SPL at 1 kHz. The app normalizes the trace at its actual 1 kHz value before applying that anchor.
- Relative response: useful for digitizing and comparison, but it cannot produce a real loudness value without an absolute reference level.
How the calculator turns a chart into loudness
- Upload a clean screenshot of the frequency-response chart.
- Check the suggested plot rectangle and enter the frequency and dB axis limits.
- Extract the coloured curve, then correct any points that picked up a legend, grid line, or wrong trace.
- Declare what the y-axis means and choose free-field or diffuse-field listening before reading the output.
The calculated output includes the one-third-octave spectrum used by the method, total loudness in sones, loudness level in phons, and a specific-loudness curve in sone/Bark.
Free field and diffuse field
Free field describes sound arriving principally from one direction with little reflection at the listening position. Diffuse field describes sound arriving from many directions, as can happen in a strongly reverberant environment. The stationary method uses different field corrections, so select the condition that best matches the measurement or intended listening setup.
What this tool does not claim
Image digitization introduces uncertainty before the loudness calculation begins. A screenshot may omit calibration details, bandwidth information, microphone conditions, or room effects. The calculator therefore labels transfer-function and narrowband routes as estimates and blocks an uncalibrated response from being treated as absolute SPL. It is designed for stationary loudness, not for time-varying, binaural, or certification use.
When is Zwicker loudness useful?
It is useful when comparing stationary noise spectra, inspecting how spectral shaping changes perception, explaining why two sounds with similar SPL can feel different, or translating a calibrated frequency-response measurement into a more human-centred summary. For equal-loudness contours themselves, see Fletcher–Munson versus Zwicker loudness; they answer a related but different question.