Speech Adjusted Noises (SAN) for German speech recognition tests
Christina Zinner, Alexandra Winkler, Inga Holube · Zenodo (CERN European Organization for Nuclear Research) · 2021
These sound files were created to use as speech simulating noises for five different German speech test materials. Every speech adjusted noise was generated by superimposing the corresponding speech material 30 times according to [1]. The used speech test materials were: Freiburg monosyllabic speech test [2], Oldenburg sentence test (German matrix test [3]), Göttingen sentence test [4], Hochmair-Schulz-Moser sentence test [5], and monosyllabic rhyme test according to von Wallenberg and Kollmeier [6]. A more detailed description and speech recognition results are available in [7]. File Speech material Reference to speech test SAN-FBE.wav monosyllables [2] SAN-OLSA.wav matrix sentences [3] SAN-GOESA.wav everyday sentences [4] SAN-HSM.wav everyday sentences [5] SAN-WAKO.wav monosyllabic rhyme words [6]