Alcohol language corpus: the first public corpus of alcoholized German speech

Abstract

The Alcohol Language Corpus (ALC) is the first publicly available speech corpus comprising intoxicated and sober speech of 162 female and male German speakers. Recordings are done in the automotive environment to allow for the development of automatic alcohol detection and to ensure a consistent acoustic environment for the alcoholized and the sober recording. The recorded speech covers a variety of contents and speech styles. Breath and blood alcohol concentration measurements are provided for all speakers. A transcription according to SpeechDat/Verbmobil standards and disfluency tagging as well as an automatic phonetic segmentation are part of the corpus. An Emu version of ALC allows easy access to basic speech parameters as well as the us of R for statistical analysis of selected parts of ALC. ALC is available without restriction for scientific or commercial use at the Bavarian Archive for Speech Signals.

The influence of alcoholic intoxication on the fundamental frequency of female and male speakers

“…In an earlier study the authors of the present paper estimated the Pearson correlation between BrAC and BAC based on samples taken from 152 intoxicated persons to be r ¼ 0:89. From the observed distribution of differences the chance for a deviation between BrAC and BAC of more than 0.0001 can be estimated to at least 29% (Schiel et al, 2012). In the present study, therefore, only BAC measurements are used as a reference.…”

Proceedings of the 2nd Workshop on Representation Learning for NLP

“…Therefore, our knowledge of the hierarchical nature of linguistic structures and the theory of linguistic compositionality have been shown to be biologically plausible. Previous works have applied this hierarchical structure to RNNs in movement tracking (Paine and Tani, 2004), sensorimotor control systems (Yamashita and Tani, 2008) and speech recognition (Heinrich et al, 2012). Based on the above conclusions, we adopt the multiple timescales concept to implement the temporal hierarchy architecture for representing multiple compositionalities which will help in handling longer sequences for our CLM.…”

“…Number of speakers S LER S WER TL LER TL WER ALC (Schiel et al, 2012) 54.54h 162 13.48% 32.83% 8.23% 21.14% HEMPEL (Draxler and Schiel, 2002) 14 (Wahlster, 1993) 32.40h 654 ----VM2 (Wahlster, 1993) 43.90h 214 ----ZIPTEL (Draxler and Schiel, 2002) 12 Table 2: Information on the kind of speech data contained in each corpus.…”

“…In order to improve the performance of the CLMs, there is a need for better representation of the additional levels of compositionality and the richer discourse structure found in CLMs. Heinrich et al (2012) used multiple timescale RNNs to learn the linguistic hierarchy for speech related tasks and Ding et al (2016) demonstrated that, during listening to connected speech, cortical activity of different timescales concurrently tracked the time course of abstract linguistic compositionality at different hierarchical levels, such as words, phrases and sentences. In this work, we propose a character-level recurrent neural network (RNN) LM that employs an adaptive multiple timescales approach to incorporate temporal hierarchies in the architecture to enhance the representation of multiple compositionalities.…”

Articulatory and Acoustic Characteristics of German Fricative Clusters

“…For our highly controlled, semantically artificial compounds, word frequency cannot be assessed, but we use the published statistics of the Bavarian Archive for Speech Signals to take an initial look at whether phonotactic probability might be an influencing factor. Phoneme bigram statistics of the Bavarian Archive for Speech Signals (Schiel, 2010) are given in table 5. 1 The bigram statistic includes any sequence of C1C2, irrespective of an intervening boundary (syllable, word, phrase, etc. ).…”