The validation_Sharpness_DIN45692_narrowband_and_broadband_signals.m code is used to verify the sharpness implementation according to DIN 45692 [1] (see Sharpness_DIN45692code here). The verification is performed considering test signals as specified by the DIN norm, which consist of:
- 21 narrowband sounds (one-critical-band wide) with center frequencies in Bark (from
$2.5~\mathrm{Bark}$ to$22.5~\mathrm{Bark}$ in$1~\mathrm{Bark}$ steps), and - 20 broadband sounds with variable lower frequencies in Bark (from
$2.5~\mathrm{Bark}$ to$21.5~\mathrm{Bark}$ in$1~\mathrm{Bark}$ steps) and fixed upper frequency of$10~\mathrm{kHz}$ ($\approx 22.4~\mathrm{Bark}$ ).
All test signals are stationary noises with a total loudness of
In order to run this code and reproduce the figures available in the figs folder, the user needs to download the dataset of sound files from zenodo here. The obtained folder called validation_SQAT_v1_0 has to be included in the sound_files folder of the toolbox.
The figures below compare the results obtained using the Sharpness_DIN45692 implementation in SQAT with reference values from the DIN norm [1]. The DIN norm stipulates a maximum tolerance of 0.05 acum from the provided reference values. The loudness of the test signals is obtained using the loudness method for stationary signals according to ISO 532-1 [2], as implemented in SQAT (see Loudness_ISO532_1 here).
| Absolute values | Relative error |
|---|---|
 |
 |
The DIN standard [1] defines three weighting functions for the calculation of sharpness. The standard weighting function
As the sharpness results may significantly differ depending on the weighting function and the test signals, it is recommended to adequately report the method used for the sake of clarity. This is also the case for the method used to compute the loudness. Moreover, the selection of the weighting function has to be carefully made depending on the test sounds [6]:
- The level-independent weighting functions
$g(z)$ and$g_{\mathrm{B}}(z)$ are recommended if the sounds to be compared have similar loudness. - The weighting function
$g_{\mathrm{A}}(z)$ accounts for the influence of the total loudness and is thus recommended if the test sounds have a significant loudness difference.
The figure below presents a comparison of sharpness values obtained using the three different weighting functions. The results are obtained using the Sharpness_DIN45692 implementation in SQAT and the narrowband and broadband test signals specified by the DIN standard [1].
| Narrowband signals | Broadband signals |
|---|---|
 |
 |
[1] Deutsches Institut für Normung. (2009). Measurement technique for the simulation of the auditory sensation of sharpness (DIN Standard No. 45692).
[2] International Organization for Standardization. (2017). Acoustics - Methods for calculating loudness - Part 1: Zwicker method (ISO Standard No. 532-1).
[3] Widmann, U. (1993). Untersuchungen zur Schärfe und zur Lästigkeit von Rauschen unterschiedlicher Spektralverteilung, DAGA 93, S. 644-647.
[4] von Bismarck, G. (1974). Sharpness as an Attribute of the Timbre of Steady Sounds. Acta Acustica united with Acustica, 30(3), 159-172.
[5] Aures, W. (1985). Berechnungsverfahren für den sensorischen Wohlklang beliebiger Schallsignale (A model for calculating the sensory euphony of various sounds). Acta Acustica united with Acustica, 59(2), 130-141.
[6] Head acoustics. (2018). Psychoacoustic analyses I, Application note. https://cdn.head-acoustics.com/fileadmin/data/global/Application-Notes/SVP/Psychoacoustic-Analyses-I-02.2018.pdf (Last viewed: May 25, 2023)
This code was released in SQAT v1.0, 14.05.2023