Metrics › Psychoacoustic annoyance

Psychoacoustic annoyance (Widmann, Zwicker & Fastl, More, Di et al.)

Unit: dimensionless · Also known as PA, psychoacoustic annoyance, Zwicker annoyance, Widmann annoyance, More annoyance, Di annoyance

Psychoacoustic annoyance (PA) combines percentile loudness N5 with sharpness, roughness and fluctuation strength (Widmann 1992; Zwicker & Fastl 1999). More (2010) added tonality and Di et al. (2016) adapted the model for tonal noise. None of these is standardised; ISO/AWI 26573, registered in 2026, covers the assessment of annoyance through listening tests, including prediction models.

Editions and who implements them

Project map

The project map, with only the lines for psychoacoustic annoyance between the projects listed for it. Agreement between a project and the code it was ported or adapted from only checks the port.

• ported or adapted from another project
• results checked against
• same contributor
• SQAT SQAT, MoSQITo, and sottek-hearing-model
• legacy legacy project

Scroll the picture sideways.

Implementations

Projects in bold and other established projects come first, then newly released, developing and legacy ones. Validation is what each project says about its own testing. The details are below, and the kinds of evidence are explained on the About page. The project map above shows which project was ported or adapted from which.

How they were validated

As reported by each project; LISQM has not run the code. Two implementations that agree can still share the same error. Kinds of evidence.

SQAT · Di et al. (2016)

SQAT · More (2010)

not stated

• No verification routine (README).

refmap-psychoacoustics · Widmann (1992)

standard or paper data

• A validation script (validation/metrics/mlab/acousticPsychAnnoyValidation.m) compares the MATLAB version with values from Widmann (1992).

phonometry new · Zwicker & Fastl (1999)

own tests only

• Tests compare one set of inputs (N5 = 30 sone, S = 2.0 acum, F = 0.5 vacil, R = 0.3 asper) with the value worked out by hand from Eqs. 16.2–16.4 (37.0478, within 0.001) and check the terms of the formula; no published example is used.

PsychoacousticMetrics.jl new · Widmann (1992)

compared with SQAT

• Cross-checked against SQAT's PsychoacousticAnnoyance_Widmann1992: an 840-case formula grid, reproduced bit-exactly, and six signal-level cases with SQAT's own N5, S5, R5 and FS5 vendored, so that every deviation is attributed to a component.
• Checked at the published 1 au anchor (a 1 kHz tone at 40 dB SPL): on the 4 s cross-check case, the percentile convention lands 0.1 % from it and the stationary approximation 0.8 %; on the README's 5 s examples, 0.3 % and 1.4 %.

pySQAT developing · Di et al. (2016)

pySQAT developing · More (2010)

pySQAT developing · Zwicker & Fastl (1999)

compared with SQAT (its source)

Ported or adapted from SQAT. It was compared with the code it was ported from, which checks the port only.

• Compared with SQAT v1.3, run in MATLAB, by validation/FINAL_VAL.py; the generated reports are not committed.

Not stated by SQAT for Zwicker & Fastl (1999) and Widmann (1992); refmap-psychoacoustics for Di et al. (2016) and More (2010); PsychoacousticParametersMeasurer.

Before you compare numbers

Implementations of the same metric can give different numbers without either being wrong. Check these choices first.

• Inputs. The models combine percentile values of loudness, sharpness, roughness and fluctuation strength (N5, S5, R5 and F5 in SQAT and PsychoacousticMetrics.jl). The percentile definition and the sharpness, roughness and fluctuation-strength models used change the result, and so does using whole-signal values instead (a stationary approximation).
• Same model, two names. Zwicker & Fastl (1999) give Widmann's (1992) model; SQAT lists it under both names. More (2010) and Di et al. (2016) are different models.

What the projects state:

SQAT

• Statistics such as N5 take the nearest rank of the sorted time series, without interpolation (get_exceeded_value), and leave out the first time_skip seconds of the signal.

PsychoacousticMetrics.jl new

• The percentile form needs 48 kHz input.
• Percentiles are Hyndman–Fan type 7 quantiles (linear interpolation); SQAT takes the nearest rank, which the README says differs by a full order-statistic step on the coarse fluctuation-strength track.

References

1. Widmann, U. (1992). Ein Modell der psychoakustischen Lästigkeit von Schallen und seine Anwendung in der Praxis der Lärmbeurteilung [Doctoral dissertation, Technische Universität München].
2. Zwicker, E., & Fastl, H. (1999). Psychoacoustics: Facts and models (2nd ed.). Springer. https://doi.org/10.1007/978-3-662-09562-1
3. More, S. R. (2010). Aircraft noise characteristics and metrics [Doctoral dissertation, Purdue University]. PARTNER report PARTNER-COE-2011-004. https://docs.lib.purdue.edu/dissertations/AAI3453255
4. Di, G., Chen, X., Song, K., Zhou, B., & Pei, C. (2016). Improvement of Zwicker's psychoacoustic annoyance model aiming at tonal noises. Applied Acoustics, 105, 164–170. https://doi.org/10.1016/j.apacoust.2015.12.006
5. ISO/AWI 26573. Acoustics — Procedures for assessment and prediction of noise-induced perceptual annoyance through laboratory-based listening tests. https://www.iso.org/standard/93922.html