14:00 - 14:15

Opening Session

14:15 - 15:15 Keynote by

Bhiksha Raj

on "Understanding the application of neural networks for signal enhancement"

15:15 - 15:45 Coffee Break

15:45 - 17:15 Oral Session: Iterative Algorithms & Machine Learning for Speech Enhancement

         EXIT Charts for Turbo Automatic Speech Recognition: A Case Study
        Timo Lohrenz, Simon Receveur and Tim Fingscheidt, TU Braunschweig

         Introducing Block-Wise Processing into Turbo Viterbi ASR
        Simon Receveur, Timo Lohrenz and Tim Fingscheidt, TU Braunschweig


Noise-Presence-Probability-Based Noise PSD Estimation by Using DNNs

        Aleksej Chinaev, Jahn Heymann, Lukas Drude and Reinhold Haeb-Umbach, Universität Paderborn


Iterative Harmonic Speech Enhancement

        Johannes Stahl and Pejman Mowlaee, Graz University of Technology


Factor Graph Decoding for Speech Presence Probability Estimation

        Thomas Glarner, Mohammad Mahdi Momenzadeh, Lukas Drude and Reinhold Haeb-Umbach, Universität Paderborn


New Insights into Turbo-Decoding-Based AVSR with Dynamic Stream Weights

        Sebastian Gergen1, Steffen Zeiler1, Ahmed Hussen Abdelaziz2 and Dorothea Kolossa1,
        1Ruhr-Universität Bochum, 2International Computer Science Institute Berkeley

17:15 - 18:45 Poster Session: Iterative Algorithms & Machine Learning for Speech Enhancement


Unsupervised Classification of Voiced Speech and Pitch Tracking Using Forward-Backward Kalman Filtering

        Benedikt Bönninghoff1, Robert Nickel2, Steffen Zeiler1 and Dorothea Kolossa1
         Ruhr-Universität Bochum, 2Bucknell University 


A Combination of Pre-Trained Approaches and Generic Methods for an Improved Speech Enhancement 

        Robert Rehr and Timo Gerkmann, Universität Oldenburg


Balancing Gaussianity and sparseness in feature-space speaker adaptation for word prominence detection

        Andrea Schnall1 and Martin Heckmann2,
        1 TU Darmstadt, 2Honda Research Institute Europe GmbH

17:15 - 18:45 Poster Session: Selected Topics in Speech Processing


Evaluation of Enhanced F0-Trajectories for Speech Detection and Classification in Acoustic Monitoring

        Frank Kurth and Alessia Cornaggia-Urrigshardt, Fraunhofer FKIE


General Detection of Speech Signals in the Time-Frequency Plane

        Sebastian Urrigshardt, Sebastian Kreuzer and Frank Kurth, Fraunhofer FKIE


Improving Vector Quantization-Based Decoders for Correlated Processes in Error-Free Transmission

        Ziyue Zhao, Sai Han and Tim Fingscheidt, TU Braunschweig


Head-Orientation-Based Device Selection: Are You Talking to Me?

        Menno Müller1, Steven van de Par2 and Jörg Bitzer1,
        1Jade Hochschule, 2Universität Oldenburg


Voice Activity Detection Based on Modulation-Phase Differences

        Simon Graf1, Tobias Herbig1, Markus Buck1 and Gerhard Schmidt2,
        1Nuance Communications Deutschland GmbH, 2Universität Kiel


A Method to Analyze the Spatial Response of Informed Spatial Filters

        Soumitro Chakrabarty, Oliver Thiergart and Emanuël A.P. Habets, International Audio Laboratories Erlangen


Estimating Source Dominated Microphone Clusters in Ad-Hoc Microphone Arrays by Fuzzy Clustering in the Feature Space

        Sebastian Gergen and Rainer Martin, Ruhr-Universität Bochum


On the Bias of Direction of Arrival Estimation Using Linear Microphone Arrays

        Florian Jacob and Reinhold Haeb-Umbach, Universität Paderborn


Coding of Parametric Models with Randomized Quantization in a Distributed Speech and Audio Codec

        Tom Bäckström and Johannes Fischer, International Audio Laboratories Erlangen

17:15 - 18:45 Poster Session: Emerging Topics and Applications


“Listen, Follow me”: The Transformational Leadership Corpus (TLC)

        Chia-Chun Hsu1, Jarek Krajewski1, Jörg Felfe2, Joachim Mrnka2, Andre Wiggerich3 and Sebastian Schnieder1,
        1Universität Wuppertal, 2Helmut-Schmidt-Universität, 3TU Dresden


Towards Opaque Audio Features for Privacy in Acoustic Sensor Networks

        Alexandru Nelus, Sebastian Gergen, Jalal Taghia and Rainer Martin, Ruhr-Universität Bochum


The Fraunhofer IAIS Audio Mining System: Current State and Future Directions

        Christoph Schmidt, Michael Stadtschnitzer and Joachim Köhler, Fraunhofer IAIS


Personalized News Event Retrieval for Small Talk in Social Dialog Systems

        Lucas Bechberger1, Maria Schmidt1, Alex Waibel1 and Marcello Federico2,
        1Karlsruher Institut für Technologie, 2Human Language Technology Fondazione Bruno Kessler


Using Tweets as "Ice-Breaking" Sentences in a Social Dialog System

        Aleksandar Andonov, Maria Schmidt, Jan Niehues and Alex Waibel, Karlsruher Institut für Technologie

18:30 - 19:30 ITG Fachgruppensitzung

19:00 - 21:00 Welcome Reception

8:30-9:30 Keynote by

Richard C. Hendriks



Optimizing Speech Intelligibility in Noisy Environments Using a Simple Model of Communication"

9:30 - 10:00 Coffee Break

10:00 - 11:00 Oral Session: Speech Processing for ear-mounted devices


Performance Comparison of Bilateral and Binaural MVDR-based Noise Reduction Algorithms in the Presence of DOA Estimation Errors

        Daniel Marquardt and Simon Doclo, Universität Oldenburg


Active Cancellation of the Occlusion Effect in Hearing Aids by Time Invariant Robust Feedback

        Stefan Liebich, Peter Jax and Peter Vary, RWTH Aachen


A Model-Based Placement Strategy for a Nearby External Microphone for Speech Enhancement in Hearing Aids

        Dianna Yee1, Rainer Martin2, Homayoun Kamkar-Parsi1 and Henning Puder1
        1Sivantos GmbH, 2Ruhr-Universität Bochum


On the Use of Beamforming Approaches for Binaural Speaker Localization

        Mehdi Zohourian, Gerald Enzner and Rainer Martin, Ruhr-Universität Bochum

11:00 - 12:30 Poster Session: Speech Processing for ear-mounted devices


Probabilistic Spatial Filter Estimation for Multi-Channel Signal Enhancement in Hearing Aids

        Hendrik Kayser and Jörn Anemüller, Universität Oldenburg

        Development of a Sound Coding Strategy based on a Deep Recurrent Neural Network for Monaural Source Separation in Cochlear Implants

        Waldo Nogueira1, Tom Gajęcki2, Benjamin Krüger1, Jordi Janer2 and Andreas Büchner1,
        1Medizinische Hochschule Hannover, 2Universitat Pompeu Fabra


On The Impact of Quantization on Binaural MVDR Beamforming

        Jamal Amini1, Richard C. Hendriks1, Richard Heusdens1, Meng Guo2 and Jesper Jensen2,
        1TU Delft, 2Aalborg University


A Robust Null-Steering Beamformer for Acoustic Feedback Cancellation for a Multi-Microphone Earpiece

        Henning Schepker1, Linh Tran2, Sven Nordholm2 and Simon Doclo1
        1Universität Oldenburg, 2Curtin University


Two-channel Coherence-Based Own Voice Detection for Privacy-aware Long-term Acoustic Measurements

        Joerg Bitzer and Sven Kissner, Jade Hochschule

11:00 - 12:30 Poster Session: Quality Evaluation


Method for analyzing personalized telephone speech in quiet and noisy environments in normal-hearing and hearing-impaired listeners 

        Tobias Bruns1, Dirk Oetting1, Markus Meis2 and Jan Rennies1,
        1Fraunhofer IDMT, 2Hörzentrum Oldenburg GmbH


Design of Double Talk Sequences in Different Languages to Harmonize Third Party Listening Test Results

        Frank Kettler, Silvia Poschen and Radi Serafimov, HEAD acoustics GmbH


Towards VoIP quality testing with real-life devices and degradations

        Michal Soloducha1, Alexander Raake1, Frank Kettler2, Nils Rohrer2, Eva Parotat3, Marcel Wältermann3, Sven Trevisany3 and Peter Voigt3,
        1TU Ilmenau, 2HEAD acoustics GmbH, 3AVM GmbH


Instrumental speech and noise quality assessment for super-wideband and fullband transmission

        Jan Reimes and H. W. Gierlich, HEAD acoustics GmbH


Emotion Intelligibility within Codec-Compressed and Reduced Bandwith Speech

        Ingo Siegert1, Alicia Flores Lotz1, Michael Maruschke2, Jokisch Oliver2 and Andreas Wendemuth1
        1Otto von Guericke Universität, 2Hochschule für Telekommunikation Leipzig


Voice and Speech Assessment From Telephone Recordings Using Prosodic Analysis Based on mu-Law-Companded Features

        Tino Haderlein, Anne Schützenberger, Michael Döllinger and Elmar Nöth, Universität Erlangen-Nürnberg


Evaluation of Communication Systems for Full-Face Firefighter Masks

        Michael Brodersen1, Moritz Juengling2 and Gerhard Schmidt2,
        1Dräger Safety AG, 2Universität Kiel

11:00 - 12:30 Poster Session: Speech & Diagnostics


Large Sleepy Reading Corpus (LSRC): Applying Read Speech for Detecting Sleepiness

        Jarek Krajewski1, Christopher Monschau2, Raphael Titt3, Martin Golz4, David Sommer4 and Sebastian Schnieder1,
        1Bergische Universität Wuppertal, 2Rheinische Fachhochschule Köln, 3Universität Tübingen, 4FH Schmalkalden


An Analysis of Perplexity to Reveal the Effects of Alzheimer's Disease on Language 

        Sebastian Wankerl, Elmar Nöth and Stefan Evert, Universität Erlangen


Gender–dependent GMM–UBM for tracking Parkinson’s disease progression from speech

        Tomás Arias1, Juan Camilo Vasquez1, Juan Rafael Orozco1, Jesús Francisco Vargas1, Tino Haderlein2 and Elmar Nöth2
        1Universidad de Antioquia, 2Universität Nürnberg-Erlangen


Towards Cross-lingual Automatic Diagnosis of Autism Spectrum Condition in Children's Voices

        Maximilian Schmitt1, Erik Marchi2, Fabien Ringeval3 and Björn Schuller1,
        1Universität Passau, 2Universität München, 3Université Grenoble Alpes


Acoustic and grammatical characterization of crisis-related babblings in Italian persons undergoing Courts-of-Law examinations

        Licia Sbattella, Roberto Tedesco and Sonia Cenceschi, Politecnico di Milano


Non-invasive photoglottography for use in the lab and the field

        Eike Suthau1, Peter Birkholz1, Alexander Mainka1 and Adrian P. Simpson2
        1TU Dresden, 2Universität Jena


On the Role of the Limbic Brain System in Recognizing Emotions From Paralinguistic Speech Features

        Sascha Frühholz, Universität Zürich

12:30 - 13:30 Lunch Break

13:30 - 14:30 Oral Session: Quality Evaluation


Non-Intrusive Estimation Model for the Speech-Quality Dimension Loudness

        Friedemann Köster, Victor Cercos-Llombart, Gabriel Mittag and Sebastian Möller, TU Berlin


Predicting the quality of processed speech by combining modulation based features and model-trees

        Benjamin Cauchi1, Joao F. Santos2, Kai Siedenburg3, Tiago H. Falk2, Patrick A. Naylor4, Simon Doclo3 and Stefan Goetze1,
        1Fraunhofer IDMT, 2Institut National de la Recherche Scientifique, 3Universität Oldenburg, 4Imperial College London


A Paired-Comparison Listening Test for Collecting Voice Likability Scores

        Laura Fernández Gallardo, TU Berlin


Objective Assessment of Artificial Speech Bandwidth Extension Approaches

        Johannes Abel1, Magdalena Kaniewska2, Cyril Guillaumé2, Wouter Tirry2 and Tim Fingscheidt1,
        1TU Braunschweig, 2NXP Software

14:30 - 15:30 Oral Session: Speech & Diagnostics


A Bag-of-Audio-Words Approach for Snore Sounds’ Excitation Localisation

        Maximilian Schmitt1, Christoph Janott2, Kun Qian1, Vedhas Pandit1, Clemens Heiser2, Werner Hemmert2 and Björn Schuller1,
        1 Universität Passau, 2 TU München


Wavelet-Based Time-Frequency Representations for Automatic Recognition of Emotions from Speech

        Juan Camilo Vasquez-Correa1, Tomas Arias-Vergara1, Juan Rafael Orozco Arroyave1, Jesus Francisco Vargas Bonilla1 and Elmar Nöth2,
        1Universidad de Antioquia, 2Universität Erlangen-Nürnberg


Detection of Intra-Personal Development of Cognitive Impairment From Conversational Speech

        Jochen Weiner and Tanja Schultz, Universität Bremen


Parkinson-Speech Analysis: Methods and Aims

        Christin Baasch, Gerhard Schmidt, Ulrich Heute, Adelheid Nebel and Günther Deuschl,
        Universität Kiel

15:30 - 16:00 Coffee Break

16:00 - 18:45 Excursion Town & HNF Computer Museum

19:00 - 23:00 Dinner Gut Ringelsbruch

8:30-9:30 Keynote by

Hynek Hermansky

on "Multistream Recognition of Speech"

9:30 - 10:00 Coffee Break

10:00 - 11:30 Oral Session: Speech Enhancement in Dynamic Acoustic Scenarios


Time Domain Approach for Listening Enhancement in Noisy Environments

        Markus Niermann, Christian Thierfeld, Peter Jax and Peter Vary, RWTH Aachen


Multiframe Echo Suppression Based on Orthogonal Signal Decompositions

        Hai Huang1, Christian Hofmann2, Walter Kellermann2, Jingdong Chen1 and Jacob Benesty3,
        1Northwestern Polytechnical University, 2Universität Nürnberg-Erlangen, 3University of Quebec


Combined Single-Microphone Wiener and MVDR Filtering based on Speech Interframe Correlations and Speech Presence Probability

        Dörte Fischer1, Timo Gerkmann1, Simon Doclo1 and Emanuel A. P. Habets2,
        1Universität Oldenburg, 2International Audio Laboratories Erlangen


A Priori SNR Estimation Using Weibull Mixture Model

        Aleksej Chinaev, Jens Heitkaemper and Reinhold Haeb-Umbach, Universität Paderborn

        Maximum-Likelihood Approach to Multichannel-Wiener-Postfiltering for Wind-Noise 


        Philipp Thüne and Gerald Enzner, Ruhr-Universität Bochum


Kurtosis-Controlled Babble Noise Suppression

        Simon Graf1, Tobias Herbig1, Markus Buck1 and Gerhard Schmidt2
        1 Nuance Communications Deutschland GmbH, 2 Universität Kiel

11:30 - 13:00 Poster Session: Speech Enhancement in Dynamic Acoustic Scenarios


Combined Linear and Nonlinear Residual Echo Suppression Using a Deficient Distortion Model - A Proof of Concept

        Ingo Schalk-Schupp1, Friedrich Faubel1, Markus Buck1 and Andreas Wendemuth2,
        1 Nuance Communications Deutschland GmbH, 2 Otto-von-Guericke Universität


On the Performance of LPTV Coherence Reduction Methods in the Sub-band Domain for Stereophonic Acoustic Echo Cancellation

        Maria Luis Valero and Emanuel A. P. Habets, International Audio Laboratories Erlangen


Spectral Envelope Statistics for Source Modelling in Speech Enhancement

        Sneha Das, Alexandra Craciun, Tobias Jähnel and Tom Bäckström, International Audio Laboratories Erlangen


A Practical Beamformer-Postfilter System for Microphone Arrays on Seat Belts

        Mohammed Krini1, Zafar-Baig Mirza2 and Klaus Rodemer2
        1Hochschule Aschaffenburg, 2Paragon AG


HMM Embedded Conditional Vector Estimation Applied to Noisy Line Spectral Frequencies

        André Klein and Stefan Feldes, Hochschule Mannheim


Acoustic Feedback Compensation with Reverb-based Stepsize Control for In-car      Communication Systems

        Philipp Bulling1, Klaus Linhard1, Arthur Wolf1 and Gerhard Schmidt2,
        1Daimler AG, 2Universität Kiel


Noise Reduction in the Time Domain Using ARMA Filtering

        Florian Heese, Richard Steinbiss, Peter Jax and Peter Vary, RWTH Aachen

11:30 - 13:00 Poster Session: Efficient Modeling ASR


Phoneme Boundary Detection using Deep Bidirectional LSTMs

        Jörg Franke1, Markus Müller1, Fatima Hamlaoui2, Sebastian Stüker1 and Alex Waibel1
        1Karlsruher Institut für Technologie, 2Zentrum für Allgemeine Sprachwissenschaft


Training Deep Neural Networks for Reverberation Robust Speech Recognition

        Marvin Ritter, Markus Müller, Sebastian Stüker, Florian Metze and Alex Waibel
        Karlsruher Institut für Technologie

11:30 - 13:00 Poster Session: Show & Tell


Binaural Noise Reduction using Raspberry Pi

        Masoumeh Azarpour and Gerald Enzner, Ruhr-Universität Bochum


3PASS & HHP IV - up-to-date speech quality tests of terminals

        Christian Schüring, HEAD acoustics GmbH


Real-time Noise Reduction and Speech Dereverberation Using a Small Microphone Array

        Benjamin Cauchi, Ante Jukic, Feifei Xiong, Jan Wellman, Simon Doclo and Stefan Goetze, Fraunhofer IDMT

13:00 - 14:00 Lunch Break

14:00 - 15:30 Oral Session: Efficient Modeling ASR


Robust Online Multi-Channel Speech Recognition

        Markus Kitza1, Jahn Heymann2, Albert Zeyer1, Ralf Schlüter1 and Reinhold Haeb-Umbach2
        1RWTH Aachen, 2Universität Paderborn


Modeling of Phone Features for Phoneme Perception

        Harald Höge, Universität der Bundeswehr München


Language Feature Vectors for Resource Constraint Speech Recognition

        Markus Müller, Sebastian Stüker and Alex Waibel, Karlsruher Institut für Technologie


Uncertainty Decoding Using a Sampling Strategy Based on the Eigenvalue Decomposition

        Christian Huemmer, Philipp Stadter and Walter Kellermann, Universität Erlangen-Nürnberg


Growing a Deep Neural Network Acoustic Model with Singular Value Decomposition

        Kevin Kilgour, Igor Tseyzer, Thai Son Nguyen, Sebastian Stüker and Alex Waibel
        Karlsruher Institut für Technologie


Rank based Decoding for Improved DNN/HMM Hybrid Acoustic Models in the EML Transcription Platform

        Volker Fischer and Siegfried Kunzmann, European Media Laboratory GmbH

15:30 - 15:45 Closing Session

15:45 - 16:00 Coffee Break