Supplementary Material for "Data-driven Auditory Contrast Enhancement for Everyday Sounds and Sonifications"

We introduce Auditory Contrast Enhancement (ACE) as a technique to enhance sounds at hand of a given collection of sound or sonification examples that belong to different classes, such as sounds of machines with and without a certain malfunction, or medical data sonifications for different pathologies/conditions. A frequent use case in inductive data mining is the discovery of patterns in which such groups can be discerned, to guide subsequent paths for modelling and feature extraction. ACE provides researchers with a set of methods to render focused auditory perspectives that accentuate inter-group differences and in turn also enhance the intra-group similarity, i.e, it warps sounds so that our human built-in metrics for assessing differences between sounds is better aligned to systematic differences between sounds belonging to different classes. We unfold and detail the concept along three different lines: temporal, spectral and spectrotemporal auditory contrast enhancement and we demonstrate their performance at hand of given sound and sonification collections. #### **ACE of Impact Sounds: table interaction sounds** ##### **S1.0-table-stimuli.wav** see article for further details. ##### **S1.1-table__spectral__mode=median_filtered_ts__order=2.5__medfilt=9__nlgain=0.3.wav** see article for further details. ##### **S1.2-table__spectral__mode=tanh_gained_ts__order=6.25__medfilt=25__nlgain=0.16.wav** see article for further details. ##### **S1.3-table__spectral__mode=tanh_gained_ts__order=26.75__medfilt=17__nlgain=0.17.wav** see article for further details. ##### **S1.4-table__spectral__mode=db_median_filtered_ps__order=47.5__medfilt=25__nlgain=0.3.wav** see article for further details. ##### **S1.5-table__temporal__mode=median_filtered_ts__order=3.5__medfilt=5__nlgain=0.3.wav** see article for further details. ##### **S1.6-table__spectrotemporal__mode=only_highest_ts__ts_threshold=3.311__sigma=0.316__mix=1.0.wav** see article for further details. ##### **S1.7-table__spectrotemporal__mode=only_highest_ts__ts_threshold=13.18__sigma=0.316__mix=1.0.wav** see article for further details. ##### **S1.8-table__spectrotemporal__mode=blurred_ts_mask__ts_threshold=6.0256__sigma=0.331__mix=1.0.wav** see article for further details. ##### **S1.9-table__spectrotemporal__mode=blurred_ts_mask__ts_threshold=6.0256__sigma=0.871__mix=0.998.wav** see article for further details. #### **ACE of Impact Sounds: finger snap interaction sounds** ##### **S2.0-snap-stimuli.wav** see article for further details. ##### **S2.1-snap__spectral__mode=median_filtered_ts__order=1.0__medfilt=1__nlgain=0.3.wav** see article for further details. ##### **S2.2-snap__spectral__mode=tanh_gained_ts__order=4.5__medfilt=1__nlgain=0.3.wav** see article for further details. ##### **S2.3-snap__spectral__mode=tanh_gained_ts__order=4.5__medfilt=1__nlgain=0.12.wav** see article for further details. ##### **S2.4-snap__spectral__mode=median_filtered_ts__order=4.5__medfilt=1__nlgain=0.3.wav** see article for further details. ##### **S2.5-snap__spectrotemporal__mode=only_highest_ts__ts_threshold=3.02__sigma=0.316__mix=1.0.wav** see article for further details. #### **ACE of synthesized sounds: SuperCollider Klank plus Noise** ##### **S3.0-klanks-stimuli.wav** see article for further details. ##### **S3.1-klanks__spectral__mode=median_filtered_ts__medfilt=15__order=1.75__nlgain=0.0.wav** see article for further details. ##### **S3.2-klanks__spectral__mode=tanh_gained_ts__order=3.0__medfilt=9__nlgain=0.21.wav** see article for further details. ##### **S3.3-klanks__spectral__mode=tanh_gained_ts__medfilt=27__order=7.0__nlgain=0.23.wav** see article for further details. ##### **S3.4-klanks__spectral__mode=median_filtered_ts__order=9.0__medfilt=1__nlgain=0.04.wav** see article for further details. ##### **S3.5-klanks__spectrotemporal__mode=blurred_ts_mask__ts_threshold=0.832__sigma=0.661__mix=0.998.wav** see article for further details. #### **ACE of continuous sounds: Vowel sounds class a vs. ä** ##### **S4.0-vowel-stimuli.wav** see article for further details. ##### **S4.1-vowel__spectral__mode=tanh_gained_ts__order=4.25__medfilt=39__nlgain=0.09.wav** see article for further details. ##### **S4.2-vowel__spectral__mode=median_filtered_ts__order=10.75__medfilt=39__nlgain=0.01.wav** see article for further details. ##### **S4.3-vowel__spectrotemporal__mode=only_highest_ts__ts_threshold=2.291__sigma=0.912__mix=1.0.wav** see article for further details. ##### **S4.4-vowel-a-ae-transition-orig.wav** see article for further details. ##### **S4.5-vowel-a-ae-transition-aced.wav** see article for further details. #### **ACE of time series sonifications: daily energy consumption profiles of a building** ##### **S5.0-building-stimuli.wav** see article for further details. ##### **S5.1-building__spectral__mode=median_filtered_ts__medfilt=5__order=1.0__nlgain=0.01.wav** see article for further details. ##### **S5.2-building__spectral__mode=tanh_gained_ts__medfilt=3__order=0.5__nlgain=0.01.wav** see article for further details. ##### **S5.3-building__temporal__mode=median_filtered_ts__medfilt=7__order=2.75__nlgain=0.3.wav** see article for further details. ##### **S5.4-building__spectrotemporal__mode=only_highest_ts__ts_threshold=3.981__sigma=0.603__mix=0.995.wav** see article for further details. #### **ACE of manipulated music recordings: Temporal ACE gives condensed difference thumbnails** ##### **S6.0-miles-stimuli.wav** see article for further details. ##### **S6.1-miles__temporal__mode=median_filtered_ts__order=0.5__medfilt=7__nlgain=0.3.wav** see article for further details. ##### **S6.2-miles__temporal__mode=tanh_gained_ts__order=0.5__medfilt=7__nlgain=0.12.wav** see article for further details.