---
_id: '17598'
author:
- first_name: Tomohiro
full_name: Nakatani, Tomohiro
last_name: Nakatani
- first_name: Christoph
full_name: Boeddeker, Christoph
id: '40767'
last_name: Boeddeker
- first_name: Keisuke
full_name: Kinoshita, Keisuke
last_name: Kinoshita
- first_name: Rintaro
full_name: Ikeshita, Rintaro
last_name: Ikeshita
- first_name: Marc
full_name: Delcroix, Marc
last_name: Delcroix
- first_name: Reinhold
full_name: Haeb-Umbach, Reinhold
id: '242'
last_name: Haeb-Umbach
citation:
ama: Nakatani T, Boeddeker C, Kinoshita K, Ikeshita R, Delcroix M, Haeb-Umbach R.
Jointly optimal denoising, dereverberation, and source separation. IEEE/ACM
Transactions on Audio, Speech, and Language Processing. Published online 2020:1-1.
doi:10.1109/TASLP.2020.3013118
apa: Nakatani, T., Boeddeker, C., Kinoshita, K., Ikeshita, R., Delcroix, M., &
Haeb-Umbach, R. (2020). Jointly optimal denoising, dereverberation, and source
separation. IEEE/ACM Transactions on Audio, Speech, and Language Processing,
1–1. https://doi.org/10.1109/TASLP.2020.3013118
bibtex: '@article{Nakatani_Boeddeker_Kinoshita_Ikeshita_Delcroix_Haeb-Umbach_2020,
title={Jointly optimal denoising, dereverberation, and source separation}, DOI={10.1109/TASLP.2020.3013118},
journal={IEEE/ACM Transactions on Audio, Speech, and Language Processing}, author={Nakatani,
Tomohiro and Boeddeker, Christoph and Kinoshita, Keisuke and Ikeshita, Rintaro
and Delcroix, Marc and Haeb-Umbach, Reinhold}, year={2020}, pages={1–1} }'
chicago: Nakatani, Tomohiro, Christoph Boeddeker, Keisuke Kinoshita, Rintaro Ikeshita,
Marc Delcroix, and Reinhold Haeb-Umbach. “Jointly Optimal Denoising, Dereverberation,
and Source Separation.” IEEE/ACM Transactions on Audio, Speech, and Language
Processing, 2020, 1–1. https://doi.org/10.1109/TASLP.2020.3013118.
ieee: 'T. Nakatani, C. Boeddeker, K. Kinoshita, R. Ikeshita, M. Delcroix, and R.
Haeb-Umbach, “Jointly optimal denoising, dereverberation, and source separation,”
IEEE/ACM Transactions on Audio, Speech, and Language Processing, pp. 1–1,
2020, doi: 10.1109/TASLP.2020.3013118.'
mla: Nakatani, Tomohiro, et al. “Jointly Optimal Denoising, Dereverberation, and
Source Separation.” IEEE/ACM Transactions on Audio, Speech, and Language Processing,
2020, pp. 1–1, doi:10.1109/TASLP.2020.3013118.
short: T. Nakatani, C. Boeddeker, K. Kinoshita, R. Ikeshita, M. Delcroix, R. Haeb-Umbach,
IEEE/ACM Transactions on Audio, Speech, and Language Processing (2020) 1–1.
date_created: 2020-08-05T06:16:56Z
date_updated: 2022-12-05T12:34:01Z
department:
- _id: '54'
doi: 10.1109/TASLP.2020.3013118
language:
- iso: eng
main_file_link:
- open_access: '1'
url: https://groups.uni-paderborn.de/nt/pubs/2020/journal_2020_boeddeker.pdf
oa: '1'
page: 1-1
publication: IEEE/ACM Transactions on Audio, Speech, and Language Processing
status: public
title: Jointly optimal denoising, dereverberation, and source separation
type: journal_article
user_id: '40767'
year: '2020'
...
---
_id: '20504'
abstract:
- lang: eng
text: 'In recent years time domain speech separation has excelled over frequency
domain separation in single channel scenarios and noise-free environments. In
this paper we dissect the gains of the time-domain audio separation network (TasNet)
approach by gradually replacing components of an utterance-level permutation invariant
training (u-PIT) based separation system in the frequency domain until the TasNet
system is reached, thus blending components of frequency domain approaches with
those of time domain approaches. Some of the intermediate variants achieve comparable
signal-to-distortion ratio (SDR) gains to TasNet, but retain the advantage of
frequency domain processing: compatibility with classic signal processing tools
such as frequency-domain beamforming and the human interpretability of the masks.
Furthermore, we show that the scale invariant signal-to-distortion ratio (si-SDR)
criterion used as loss function in TasNet is related to a logarithmic mean square
error criterion and that it is this criterion which contributes most reliable
to the performance advantage of TasNet. Finally, we critically assess which gains
in a noise-free single channel environment generalize to more realistic reverberant
conditions.'
author:
- first_name: Jens
full_name: Heitkaemper, Jens
id: '27643'
last_name: Heitkaemper
- first_name: Darius
full_name: Jakobeit, Darius
last_name: Jakobeit
- first_name: Christoph
full_name: Boeddeker, Christoph
id: '40767'
last_name: Boeddeker
- first_name: Lukas
full_name: Drude, Lukas
last_name: Drude
- first_name: Reinhold
full_name: Haeb-Umbach, Reinhold
id: '242'
last_name: Haeb-Umbach
citation:
ama: 'Heitkaemper J, Jakobeit D, Boeddeker C, Drude L, Haeb-Umbach R. Demystifying
TasNet: A Dissecting Approach. In: ICASSP 2020 Virtual Barcelona Spain.
; 2020.'
apa: 'Heitkaemper, J., Jakobeit, D., Boeddeker, C., Drude, L., & Haeb-Umbach,
R. (2020). Demystifying TasNet: A Dissecting Approach. ICASSP 2020 Virtual
Barcelona Spain.'
bibtex: '@inproceedings{Heitkaemper_Jakobeit_Boeddeker_Drude_Haeb-Umbach_2020, title={Demystifying
TasNet: A Dissecting Approach}, booktitle={ICASSP 2020 Virtual Barcelona Spain},
author={Heitkaemper, Jens and Jakobeit, Darius and Boeddeker, Christoph and Drude,
Lukas and Haeb-Umbach, Reinhold}, year={2020} }'
chicago: 'Heitkaemper, Jens, Darius Jakobeit, Christoph Boeddeker, Lukas Drude,
and Reinhold Haeb-Umbach. “Demystifying TasNet: A Dissecting Approach.” In ICASSP
2020 Virtual Barcelona Spain, 2020.'
ieee: 'J. Heitkaemper, D. Jakobeit, C. Boeddeker, L. Drude, and R. Haeb-Umbach,
“Demystifying TasNet: A Dissecting Approach,” 2020.'
mla: 'Heitkaemper, Jens, et al. “Demystifying TasNet: A Dissecting Approach.” ICASSP
2020 Virtual Barcelona Spain, 2020.'
short: 'J. Heitkaemper, D. Jakobeit, C. Boeddeker, L. Drude, R. Haeb-Umbach, in:
ICASSP 2020 Virtual Barcelona Spain, 2020.'
date_created: 2020-11-25T14:56:53Z
date_updated: 2022-01-13T08:47:32Z
ddc:
- '000'
department:
- _id: '54'
file:
- access_level: closed
content_type: application/pdf
creator: jensheit
date_created: 2020-12-11T12:36:37Z
date_updated: 2020-12-11T12:36:37Z
file_id: '20699'
file_name: ms.pdf
file_size: 3871374
relation: main_file
success: 1
file_date_updated: 2020-12-11T12:36:37Z
has_accepted_license: '1'
keyword:
- voice activity detection
- speech activity detection
- neural network
- statistical speech processing
language:
- iso: eng
project:
- _id: '52'
name: 'PC2: Computing Resources Provided by the Paderborn Center for Parallel Computing'
publication: ICASSP 2020 Virtual Barcelona Spain
quality_controlled: '1'
status: public
title: 'Demystifying TasNet: A Dissecting Approach'
type: conference
user_id: '40767'
year: '2020'
...
---
_id: '20505'
abstract:
- lang: eng
text: "Speech activity detection (SAD), which often rests on the fact that the noise
is \"more'' stationary than speech, is particularly challenging in non-stationary
environments, because the time variance of the acoustic scene makes it difficult
to discriminate speech from noise. We propose two approaches to SAD, where one
is based on statistical signal processing, while the other utilizes neural networks.
The former employs sophisticated signal processing to track the noise and speech
energies and is meant to support the case for a resource efficient, unsupervised
signal processing approach.\r\nThe latter introduces a recurrent network layer
that operates on short segments of the input speech to do temporal smoothing in
the presence of non-stationary noise. The systems are tested on the Fearless Steps
challenge database, which consists of the transmission data from the Apollo-11
space mission.\r\nThe statistical SAD achieves comparable detection performance
to earlier proposed neural network based SADs, while the neural network based
approach leads to a decision cost function of 1.07% on the evaluation set of the
2020 Fearless Steps Challenge, which sets a new state of the art."
author:
- first_name: Jens
full_name: Heitkaemper, Jens
id: '27643'
last_name: Heitkaemper
- first_name: Joerg
full_name: Schmalenstroeer, Joerg
id: '460'
last_name: Schmalenstroeer
- first_name: Reinhold
full_name: Haeb-Umbach, Reinhold
id: '242'
last_name: Haeb-Umbach
citation:
ama: 'Heitkaemper J, Schmalenstroeer J, Haeb-Umbach R. Statistical and Neural Network
Based Speech Activity Detection in Non-Stationary Acoustic Environments. In: INTERSPEECH
2020 Virtual Shanghai China. ; 2020.'
apa: Heitkaemper, J., Schmalenstroeer, J., & Haeb-Umbach, R. (2020). Statistical
and Neural Network Based Speech Activity Detection in Non-Stationary Acoustic
Environments. INTERSPEECH 2020 Virtual Shanghai China.
bibtex: '@inproceedings{Heitkaemper_Schmalenstroeer_Haeb-Umbach_2020, title={Statistical
and Neural Network Based Speech Activity Detection in Non-Stationary Acoustic
Environments}, booktitle={INTERSPEECH 2020 Virtual Shanghai China}, author={Heitkaemper,
Jens and Schmalenstroeer, Joerg and Haeb-Umbach, Reinhold}, year={2020} }'
chicago: Heitkaemper, Jens, Joerg Schmalenstroeer, and Reinhold Haeb-Umbach. “Statistical
and Neural Network Based Speech Activity Detection in Non-Stationary Acoustic
Environments.” In INTERSPEECH 2020 Virtual Shanghai China, 2020.
ieee: J. Heitkaemper, J. Schmalenstroeer, and R. Haeb-Umbach, “Statistical and Neural
Network Based Speech Activity Detection in Non-Stationary Acoustic Environments,”
2020.
mla: Heitkaemper, Jens, et al. “Statistical and Neural Network Based Speech Activity
Detection in Non-Stationary Acoustic Environments.” INTERSPEECH 2020 Virtual
Shanghai China, 2020.
short: 'J. Heitkaemper, J. Schmalenstroeer, R. Haeb-Umbach, in: INTERSPEECH 2020
Virtual Shanghai China, 2020.'
date_created: 2020-11-25T15:03:19Z
date_updated: 2023-10-26T08:28:49Z
ddc:
- '000'
department:
- _id: '54'
file:
- access_level: closed
content_type: application/pdf
creator: jensheit
date_created: 2020-12-11T12:33:04Z
date_updated: 2020-12-11T12:33:04Z
file_id: '20697'
file_name: ms.pdf
file_size: 998706
relation: main_file
success: 1
file_date_updated: 2020-12-11T12:33:04Z
has_accepted_license: '1'
keyword:
- voice activity detection
- speech activity detection
- neural network
- statistical speech processing
language:
- iso: eng
project:
- _id: '52'
name: Computing Resources Provided by the Paderborn Center for Parallel Computing
publication: INTERSPEECH 2020 Virtual Shanghai China
status: public
title: Statistical and Neural Network Based Speech Activity Detection in Non-Stationary
Acoustic Environments
type: conference
user_id: '460'
year: '2020'
...
---
_id: '20762'
abstract:
- lang: eng
text: The rising interest in single-channel multi-speaker speech separation sparked
development of End-to-End (E2E) approaches to multispeaker speech recognition.
However, up until now, state-of-theart neural network–based time domain source
separation has not yet been combined with E2E speech recognition. We here demonstrate
how to combine a separation module based on a Convolutional Time domain Audio
Separation Network (Conv-TasNet) with an E2E speech recognizer and how to train
such a model jointly by distributing it over multiple GPUs or by approximating
truncated back-propagation for the convolutional front-end. To put this work into
perspective and illustrate the complexity of the design space, we provide a compact
overview of single-channel multi-speaker recognition systems. Our experiments
show a word error rate of 11.0% on WSJ0-2mix and indicate that our joint time
domain model can yield substantial improvements over cascade DNN-HMM and monolithic
E2E frequency domain systems proposed so far.
author:
- first_name: Thilo
full_name: von Neumann, Thilo
id: '49870'
last_name: von Neumann
orcid: https://orcid.org/0000-0002-7717-8670
- first_name: Keisuke
full_name: Kinoshita, Keisuke
last_name: Kinoshita
- first_name: Lukas
full_name: Drude, Lukas
last_name: Drude
- first_name: Christoph
full_name: Boeddeker, Christoph
id: '40767'
last_name: Boeddeker
- first_name: Marc
full_name: Delcroix, Marc
last_name: Delcroix
- first_name: Tomohiro
full_name: Nakatani, Tomohiro
last_name: Nakatani
- first_name: Reinhold
full_name: Haeb-Umbach, Reinhold
id: '242'
last_name: Haeb-Umbach
citation:
ama: 'von Neumann T, Kinoshita K, Drude L, et al. End-to-End Training of Time Domain
Audio Separation and Recognition. In: ICASSP 2020 - 2020 IEEE International
Conference on Acoustics, Speech and Signal Processing (ICASSP). ; 2020:7004-7008.
doi:10.1109/ICASSP40776.2020.9053461'
apa: von Neumann, T., Kinoshita, K., Drude, L., Boeddeker, C., Delcroix, M., Nakatani,
T., & Haeb-Umbach, R. (2020). End-to-End Training of Time Domain Audio Separation
and Recognition. ICASSP 2020 - 2020 IEEE International Conference on Acoustics,
Speech and Signal Processing (ICASSP), 7004–7008. https://doi.org/10.1109/ICASSP40776.2020.9053461
bibtex: '@inproceedings{von Neumann_Kinoshita_Drude_Boeddeker_Delcroix_Nakatani_Haeb-Umbach_2020,
title={End-to-End Training of Time Domain Audio Separation and Recognition}, DOI={10.1109/ICASSP40776.2020.9053461},
booktitle={ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech
and Signal Processing (ICASSP)}, author={von Neumann, Thilo and Kinoshita, Keisuke
and Drude, Lukas and Boeddeker, Christoph and Delcroix, Marc and Nakatani, Tomohiro
and Haeb-Umbach, Reinhold}, year={2020}, pages={7004–7008} }'
chicago: Neumann, Thilo von, Keisuke Kinoshita, Lukas Drude, Christoph Boeddeker,
Marc Delcroix, Tomohiro Nakatani, and Reinhold Haeb-Umbach. “End-to-End Training
of Time Domain Audio Separation and Recognition.” In ICASSP 2020 - 2020 IEEE
International Conference on Acoustics, Speech and Signal Processing (ICASSP),
7004–8, 2020. https://doi.org/10.1109/ICASSP40776.2020.9053461.
ieee: 'T. von Neumann et al., “End-to-End Training of Time Domain Audio Separation
and Recognition,” in ICASSP 2020 - 2020 IEEE International Conference on Acoustics,
Speech and Signal Processing (ICASSP), 2020, pp. 7004–7008, doi: 10.1109/ICASSP40776.2020.9053461.'
mla: von Neumann, Thilo, et al. “End-to-End Training of Time Domain Audio Separation
and Recognition.” ICASSP 2020 - 2020 IEEE International Conference on Acoustics,
Speech and Signal Processing (ICASSP), 2020, pp. 7004–08, doi:10.1109/ICASSP40776.2020.9053461.
short: 'T. von Neumann, K. Kinoshita, L. Drude, C. Boeddeker, M. Delcroix, T. Nakatani,
R. Haeb-Umbach, in: ICASSP 2020 - 2020 IEEE International Conference on Acoustics,
Speech and Signal Processing (ICASSP), 2020, pp. 7004–7008.'
date_created: 2020-12-16T14:07:54Z
date_updated: 2023-11-15T12:17:45Z
ddc:
- '000'
department:
- _id: '54'
doi: 10.1109/ICASSP40776.2020.9053461
file:
- access_level: open_access
content_type: application/pdf
creator: huesera
date_created: 2020-12-16T14:09:48Z
date_updated: 2020-12-16T14:09:48Z
file_id: '20763'
file_name: ICASSP_2020_vonNeumann_Paper.pdf
file_size: 192529
relation: main_file
file_date_updated: 2020-12-16T14:09:48Z
has_accepted_license: '1'
language:
- iso: eng
oa: '1'
page: 7004-7008
project:
- _id: '52'
name: Computing Resources Provided by the Paderborn Center for Parallel Computing
publication: ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech
and Signal Processing (ICASSP)
quality_controlled: '1'
status: public
title: End-to-End Training of Time Domain Audio Separation and Recognition
type: conference
user_id: '49870'
year: '2020'
...
---
_id: '20764'
abstract:
- lang: eng
text: 'Most approaches to multi-talker overlapped speech separation and recognition
assume that the number of simultaneously active speakers is given, but in realistic
situations, it is typically unknown. To cope with this, we extend an iterative
speech extraction system with mechanisms to count the number of sources and combine
it with a single-talker speech recognizer to form the first end-to-end multi-talker
automatic speech recognition system for an unknown number of active speakers.
Our experiments show very promising performance in counting accuracy, source separation
and speech recognition on simulated clean mixtures from WSJ0-2mix and WSJ0-3mix.
Among others, we set a new state-of-the-art word error rate on the WSJ0-2mix database.
Furthermore, our system generalizes well to a larger number of speakers than it
ever saw during training, as shown in experiments with the WSJ0-4mix database. '
author:
- first_name: Thilo
full_name: von Neumann, Thilo
id: '49870'
last_name: von Neumann
orcid: https://orcid.org/0000-0002-7717-8670
- first_name: Christoph
full_name: Boeddeker, Christoph
id: '40767'
last_name: Boeddeker
- first_name: Lukas
full_name: Drude, Lukas
last_name: Drude
- first_name: Keisuke
full_name: Kinoshita, Keisuke
last_name: Kinoshita
- first_name: Marc
full_name: Delcroix, Marc
last_name: Delcroix
- first_name: Tomohiro
full_name: Nakatani, Tomohiro
last_name: Nakatani
- first_name: Reinhold
full_name: Haeb-Umbach, Reinhold
id: '242'
last_name: Haeb-Umbach
citation:
ama: 'von Neumann T, Boeddeker C, Drude L, et al. Multi-Talker ASR for an Unknown
Number of Sources: Joint Training of Source Counting, Separation and ASR. In:
Proc. Interspeech 2020. ; 2020:3097-3101. doi:10.21437/Interspeech.2020-2519'
apa: 'von Neumann, T., Boeddeker, C., Drude, L., Kinoshita, K., Delcroix, M., Nakatani,
T., & Haeb-Umbach, R. (2020). Multi-Talker ASR for an Unknown Number of Sources:
Joint Training of Source Counting, Separation and ASR. Proc. Interspeech 2020,
3097–3101. https://doi.org/10.21437/Interspeech.2020-2519'
bibtex: '@inproceedings{von Neumann_Boeddeker_Drude_Kinoshita_Delcroix_Nakatani_Haeb-Umbach_2020,
title={Multi-Talker ASR for an Unknown Number of Sources: Joint Training of Source
Counting, Separation and ASR}, DOI={10.21437/Interspeech.2020-2519},
booktitle={Proc. Interspeech 2020}, author={von Neumann, Thilo and Boeddeker,
Christoph and Drude, Lukas and Kinoshita, Keisuke and Delcroix, Marc and Nakatani,
Tomohiro and Haeb-Umbach, Reinhold}, year={2020}, pages={3097–3101} }'
chicago: 'Neumann, Thilo von, Christoph Boeddeker, Lukas Drude, Keisuke Kinoshita,
Marc Delcroix, Tomohiro Nakatani, and Reinhold Haeb-Umbach. “Multi-Talker ASR
for an Unknown Number of Sources: Joint Training of Source Counting, Separation
and ASR.” In Proc. Interspeech 2020, 3097–3101, 2020. https://doi.org/10.21437/Interspeech.2020-2519.'
ieee: 'T. von Neumann et al., “Multi-Talker ASR for an Unknown Number of
Sources: Joint Training of Source Counting, Separation and ASR,” in Proc. Interspeech
2020, 2020, pp. 3097–3101, doi: 10.21437/Interspeech.2020-2519.'
mla: 'von Neumann, Thilo, et al. “Multi-Talker ASR for an Unknown Number of Sources:
Joint Training of Source Counting, Separation and ASR.” Proc. Interspeech 2020,
2020, pp. 3097–101, doi:10.21437/Interspeech.2020-2519.'
short: 'T. von Neumann, C. Boeddeker, L. Drude, K. Kinoshita, M. Delcroix, T. Nakatani,
R. Haeb-Umbach, in: Proc. Interspeech 2020, 2020, pp. 3097–3101.'
date_created: 2020-12-16T14:12:45Z
date_updated: 2023-11-15T12:17:57Z
ddc:
- '000'
department:
- _id: '54'
doi: 10.21437/Interspeech.2020-2519
file:
- access_level: open_access
content_type: application/pdf
creator: huesera
date_created: 2020-12-16T14:14:14Z
date_updated: 2020-12-16T14:14:14Z
file_id: '20765'
file_name: INTERSPEECH_2020_vonNeumann_Paper.pdf
file_size: 267893
relation: main_file
file_date_updated: 2020-12-16T14:14:14Z
has_accepted_license: '1'
language:
- iso: eng
oa: '1'
page: 3097-3101
project:
- _id: '52'
name: Computing Resources Provided by the Paderborn Center for Parallel Computing
publication: Proc. Interspeech 2020
quality_controlled: '1'
status: public
title: 'Multi-Talker ASR for an Unknown Number of Sources: Joint Training of Source
Counting, Separation and ASR'
type: conference
user_id: '49870'
year: '2020'
...
---
_id: '18651'
abstract:
- lang: eng
text: 'We present an approach to deep neural network based (DNN-based) distance
estimation in reverberant rooms for supporting geometry calibration tasks in wireless
acoustic sensor networks. Signal diffuseness information from acoustic signals
is aggregated via the coherent-to-diffuse power ratio to obtain a distance-related
feature, which is mapped to a source-to-microphone distance estimate by means
of a DNN. This information is then combined with direction-of-arrival estimates
from compact microphone arrays to infer the geometry of the sensor network. Unlike
many other approaches to geometry calibration, the proposed scheme does only require
that the sampling clocks of the sensor nodes are roughly synchronized. In simulations
we show that the proposed DNN-based distance estimator generalizes to unseen acoustic
environments and that precise estimates of the sensor node positions are obtained. '
author:
- first_name: Tobias
full_name: Gburrek, Tobias
id: '44006'
last_name: Gburrek
- first_name: Joerg
full_name: Schmalenstroeer, Joerg
id: '460'
last_name: Schmalenstroeer
- first_name: Andreas
full_name: Brendel, Andreas
last_name: Brendel
- first_name: Walter
full_name: Kellermann, Walter
last_name: Kellermann
- first_name: Reinhold
full_name: Haeb-Umbach, Reinhold
id: '242'
last_name: Haeb-Umbach
citation:
ama: 'Gburrek T, Schmalenstroeer J, Brendel A, Kellermann W, Haeb-Umbach R. Deep
Neural Network based Distance Estimation for Geometry Calibration in Acoustic
Sensor Network. In: European Signal Processing Conference (EUSIPCO). ;
2020.'
apa: Gburrek, T., Schmalenstroeer, J., Brendel, A., Kellermann, W., & Haeb-Umbach,
R. (2020). Deep Neural Network based Distance Estimation for Geometry Calibration
in Acoustic Sensor Network. European Signal Processing Conference (EUSIPCO).
bibtex: '@inproceedings{Gburrek_Schmalenstroeer_Brendel_Kellermann_Haeb-Umbach_2020,
title={Deep Neural Network based Distance Estimation for Geometry Calibration
in Acoustic Sensor Network}, booktitle={European Signal Processing Conference
(EUSIPCO)}, author={Gburrek, Tobias and Schmalenstroeer, Joerg and Brendel, Andreas
and Kellermann, Walter and Haeb-Umbach, Reinhold}, year={2020} }'
chicago: Gburrek, Tobias, Joerg Schmalenstroeer, Andreas Brendel, Walter Kellermann,
and Reinhold Haeb-Umbach. “Deep Neural Network Based Distance Estimation for Geometry
Calibration in Acoustic Sensor Network.” In European Signal Processing Conference
(EUSIPCO), 2020.
ieee: T. Gburrek, J. Schmalenstroeer, A. Brendel, W. Kellermann, and R. Haeb-Umbach,
“Deep Neural Network based Distance Estimation for Geometry Calibration in Acoustic
Sensor Network,” 2020.
mla: Gburrek, Tobias, et al. “Deep Neural Network Based Distance Estimation for
Geometry Calibration in Acoustic Sensor Network.” European Signal Processing
Conference (EUSIPCO), 2020.
short: 'T. Gburrek, J. Schmalenstroeer, A. Brendel, W. Kellermann, R. Haeb-Umbach,
in: European Signal Processing Conference (EUSIPCO), 2020.'
date_created: 2020-08-31T07:20:57Z
date_updated: 2023-11-17T06:23:39Z
ddc:
- '004'
department:
- _id: '54'
file:
- access_level: open_access
content_type: application/pdf
creator: tgburrek
date_created: 2023-11-17T06:21:40Z
date_updated: 2023-11-17T06:21:40Z
file_id: '48987'
file_name: Gburrek2020.pdf
file_size: 292159
relation: main_file
file_date_updated: 2023-11-17T06:21:40Z
has_accepted_license: '1'
language:
- iso: eng
oa: '1'
publication: European Signal Processing Conference (EUSIPCO)
quality_controlled: '1'
status: public
title: Deep Neural Network based Distance Estimation for Geometry Calibration in Acoustic
Sensor Network
type: conference
user_id: '44006'
year: '2020'
...
---
_id: '20766'
abstract:
- lang: eng
text: Recently, the source separation performance was greatly improved by time-domain
audio source separation based on dual-path recurrent neural network (DPRNN). DPRNN
is a simple but effective model for a long sequential data. While DPRNN is quite
efficient in modeling a sequential data of the length of an utterance, i.e., about
5 to 10 second data, it is harder to apply it to longer sequences such as whole
conversations consisting of multiple utterances. It is simply because, in such
a case, the number of time steps consumed by its internal module called inter-chunk
RNN becomes extremely large. To mitigate this problem, this paper proposes a multi-path
RNN (MPRNN), a generalized version of DPRNN, that models the input data in a hierarchical
manner. In the MPRNN framework, the input data is represented at several (>_ 3)
time-resolutions, each of which is modeled by a specific RNN sub-module. For example,
the RNN sub-module that deals with the finest resolution may model temporal relationship
only within a phoneme, while the RNN sub-module handling the most coarse resolution
may capture only the relationship between utterances such as speaker information.
We perform experiments using simulated dialogue-like mixtures and show that MPRNN
has greater model capacity, and it outperforms the current state-of-the-art DPRNN
framework especially in online processing scenarios.
author:
- first_name: Keisuke
full_name: Kinoshita, Keisuke
last_name: Kinoshita
- first_name: Thilo
full_name: von Neumann, Thilo
id: '49870'
last_name: von Neumann
orcid: https://orcid.org/0000-0002-7717-8670
- first_name: Marc
full_name: Delcroix, Marc
last_name: Delcroix
- first_name: Tomohiro
full_name: Nakatani, Tomohiro
last_name: Nakatani
- first_name: Reinhold
full_name: Haeb-Umbach, Reinhold
id: '242'
last_name: Haeb-Umbach
citation:
ama: 'Kinoshita K, von Neumann T, Delcroix M, Nakatani T, Haeb-Umbach R. Multi-Path
RNN for Hierarchical Modeling of Long Sequential Data and its Application to Speaker
Stream Separation. In: Proc. Interspeech 2020. ; 2020:2652-2656. doi:10.21437/Interspeech.2020-2388'
apa: Kinoshita, K., von Neumann, T., Delcroix, M., Nakatani, T., & Haeb-Umbach,
R. (2020). Multi-Path RNN for Hierarchical Modeling of Long Sequential Data and
its Application to Speaker Stream Separation. Proc. Interspeech 2020, 2652–2656.
https://doi.org/10.21437/Interspeech.2020-2388
bibtex: '@inproceedings{Kinoshita_von Neumann_Delcroix_Nakatani_Haeb-Umbach_2020,
title={Multi-Path RNN for Hierarchical Modeling of Long Sequential Data and its
Application to Speaker Stream Separation}, DOI={10.21437/Interspeech.2020-2388},
booktitle={Proc. Interspeech 2020}, author={Kinoshita, Keisuke and von Neumann,
Thilo and Delcroix, Marc and Nakatani, Tomohiro and Haeb-Umbach, Reinhold}, year={2020},
pages={2652–2656} }'
chicago: Kinoshita, Keisuke, Thilo von Neumann, Marc Delcroix, Tomohiro Nakatani,
and Reinhold Haeb-Umbach. “Multi-Path RNN for Hierarchical Modeling of Long Sequential
Data and Its Application to Speaker Stream Separation.” In Proc. Interspeech
2020, 2652–56, 2020. https://doi.org/10.21437/Interspeech.2020-2388.
ieee: 'K. Kinoshita, T. von Neumann, M. Delcroix, T. Nakatani, and R. Haeb-Umbach,
“Multi-Path RNN for Hierarchical Modeling of Long Sequential Data and its Application
to Speaker Stream Separation,” in Proc. Interspeech 2020, 2020, pp. 2652–2656,
doi: 10.21437/Interspeech.2020-2388.'
mla: Kinoshita, Keisuke, et al. “Multi-Path RNN for Hierarchical Modeling of Long
Sequential Data and Its Application to Speaker Stream Separation.” Proc. Interspeech
2020, 2020, pp. 2652–56, doi:10.21437/Interspeech.2020-2388.
short: 'K. Kinoshita, T. von Neumann, M. Delcroix, T. Nakatani, R. Haeb-Umbach,
in: Proc. Interspeech 2020, 2020, pp. 2652–2656.'
date_created: 2020-12-16T14:15:24Z
date_updated: 2023-11-15T12:14:25Z
ddc:
- '000'
department:
- _id: '54'
doi: 10.21437/Interspeech.2020-2388
file:
- access_level: open_access
content_type: application/pdf
creator: huesera
date_created: 2020-12-16T14:16:32Z
date_updated: 2020-12-16T14:16:32Z
file_id: '20767'
file_name: INTERSPEECH_2020_vonNeumann1_Paper.pdf
file_size: 1725219
relation: main_file
file_date_updated: 2020-12-16T14:16:32Z
has_accepted_license: '1'
language:
- iso: eng
oa: '1'
page: 2652-2656
publication: Proc. Interspeech 2020
quality_controlled: '1'
status: public
title: Multi-Path RNN for Hierarchical Modeling of Long Sequential Data and its Application
to Speaker Stream Separation
type: conference
user_id: '49870'
year: '2020'
...
---
_id: '20753'
abstract:
- lang: eng
text: 'In this paper we present our system for the detection and classification
of acoustic scenes and events (DCASE) 2020 Challenge Task 4: Sound event detection
and separation in domestic environments. We introduce two new models: the forward-backward
convolutional recurrent neural network (FBCRNN) and the tag-conditioned convolutional
neural network (CNN). The FBCRNN employs two recurrent neural network (RNN) classifiers
sharing the same CNN for preprocessing. With one RNN processing a recording in
forward direction and the other in backward direction, the two networks are trained
to jointly predict audio tags, i.e., weak labels, at each time step within a recording,
given that at each time step they have jointly processed the whole recording.
The proposed training encourages the classifiers to tag events as soon as possible.
Therefore, after training, the networks can be applied to shorter audio segments
of, e.g., 200ms, allowing sound event detection (SED). Further, we propose a tag-conditioned
CNN to complement SED. It is trained to predict strong labels while using (predicted)
tags, i.e., weak labels, as additional input. For training pseudo strong labels
from a FBCRNN ensemble are used. The presented system scored the fourth and third
place in the systems and teams rankings, respectively. Subsequent improvements
allow our system to even outperform the challenge baseline and winner systems
in average by, respectively, 18.0% and 2.2% event-based F1-score on the validation
set. Source code is publicly available at https://github.com/fgnt/pb_sed.'
author:
- first_name: Janek
full_name: Ebbers, Janek
id: '34851'
last_name: Ebbers
- first_name: Reinhold
full_name: Haeb-Umbach, Reinhold
id: '242'
last_name: Haeb-Umbach
citation:
ama: 'Ebbers J, Haeb-Umbach R. Forward-Backward Convolutional Recurrent Neural Networks
and Tag-Conditioned Convolutional Neural Networks for Weakly Labeled Semi-Supervised
Sound Event Detection. In: Proceedings of the Detection and Classification
of Acoustic Scenes and Events 2020 Workshop (DCASE2020). ; 2020.'
apa: Ebbers, J., & Haeb-Umbach, R. (2020). Forward-Backward Convolutional Recurrent
Neural Networks and Tag-Conditioned Convolutional Neural Networks for Weakly Labeled
Semi-Supervised Sound Event Detection. Proceedings of the Detection and Classification
of Acoustic Scenes and Events 2020 Workshop (DCASE2020).
bibtex: '@inproceedings{Ebbers_Haeb-Umbach_2020, title={Forward-Backward Convolutional
Recurrent Neural Networks and Tag-Conditioned Convolutional Neural Networks for
Weakly Labeled Semi-Supervised Sound Event Detection}, booktitle={Proceedings
of the Detection and Classification of Acoustic Scenes and Events 2020 Workshop
(DCASE2020)}, author={Ebbers, Janek and Haeb-Umbach, Reinhold}, year={2020} }'
chicago: Ebbers, Janek, and Reinhold Haeb-Umbach. “Forward-Backward Convolutional
Recurrent Neural Networks and Tag-Conditioned Convolutional Neural Networks for
Weakly Labeled Semi-Supervised Sound Event Detection.” In Proceedings of the
Detection and Classification of Acoustic Scenes and Events 2020 Workshop (DCASE2020),
2020.
ieee: J. Ebbers and R. Haeb-Umbach, “Forward-Backward Convolutional Recurrent Neural
Networks and Tag-Conditioned Convolutional Neural Networks for Weakly Labeled
Semi-Supervised Sound Event Detection,” 2020.
mla: Ebbers, Janek, and Reinhold Haeb-Umbach. “Forward-Backward Convolutional Recurrent
Neural Networks and Tag-Conditioned Convolutional Neural Networks for Weakly Labeled
Semi-Supervised Sound Event Detection.” Proceedings of the Detection and Classification
of Acoustic Scenes and Events 2020 Workshop (DCASE2020), 2020.
short: 'J. Ebbers, R. Haeb-Umbach, in: Proceedings of the Detection and Classification
of Acoustic Scenes and Events 2020 Workshop (DCASE2020), 2020.'
date_created: 2020-12-16T08:55:27Z
date_updated: 2023-11-22T08:27:32Z
ddc:
- '000'
department:
- _id: '54'
file:
- access_level: open_access
content_type: application/pdf
creator: huesera
date_created: 2020-12-16T08:57:22Z
date_updated: 2020-12-16T08:57:22Z
file_id: '20754'
file_name: DCASE2020Workshop_Ebbers_Paper.pdf
file_size: 108326
relation: main_file
file_date_updated: 2020-12-16T08:57:22Z
has_accepted_license: '1'
language:
- iso: eng
oa: '1'
project:
- _id: '52'
name: 'PC2: Computing Resources Provided by the Paderborn Center for Parallel Computing'
publication: Proceedings of the Detection and Classification of Acoustic Scenes and
Events 2020 Workshop (DCASE2020)
quality_controlled: '1'
status: public
title: Forward-Backward Convolutional Recurrent Neural Networks and Tag-Conditioned
Convolutional Neural Networks for Weakly Labeled Semi-Supervised Sound Event Detection
type: conference
user_id: '34851'
year: '2020'
...
---
_id: '20695'
author:
- first_name: Christoph
full_name: Boeddeker, Christoph
id: '40767'
last_name: Boeddeker
- first_name: Tomohiro
full_name: Nakatani, Tomohiro
last_name: Nakatani
- first_name: Keisuke
full_name: Kinoshita, Keisuke
last_name: Kinoshita
- first_name: Reinhold
full_name: Haeb-Umbach, Reinhold
id: '242'
last_name: Haeb-Umbach
citation:
ama: 'Boeddeker C, Nakatani T, Kinoshita K, Haeb-Umbach R. Jointly Optimal Dereverberation
and Beamforming. In: ICASSP 2020 - 2020 IEEE International Conference on Acoustics,
Speech and Signal Processing (ICASSP). ; 2020. doi:10.1109/icassp40776.2020.9054393'
apa: Boeddeker, C., Nakatani, T., Kinoshita, K., & Haeb-Umbach, R. (2020). Jointly
Optimal Dereverberation and Beamforming. ICASSP 2020 - 2020 IEEE International
Conference on Acoustics, Speech and Signal Processing (ICASSP). https://doi.org/10.1109/icassp40776.2020.9054393
bibtex: '@inproceedings{Boeddeker_Nakatani_Kinoshita_Haeb-Umbach_2020, title={Jointly
Optimal Dereverberation and Beamforming}, DOI={10.1109/icassp40776.2020.9054393},
booktitle={ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech
and Signal Processing (ICASSP)}, author={Boeddeker, Christoph and Nakatani, Tomohiro
and Kinoshita, Keisuke and Haeb-Umbach, Reinhold}, year={2020} }'
chicago: Boeddeker, Christoph, Tomohiro Nakatani, Keisuke Kinoshita, and Reinhold
Haeb-Umbach. “Jointly Optimal Dereverberation and Beamforming.” In ICASSP 2020
- 2020 IEEE International Conference on Acoustics, Speech and Signal Processing
(ICASSP), 2020. https://doi.org/10.1109/icassp40776.2020.9054393.
ieee: 'C. Boeddeker, T. Nakatani, K. Kinoshita, and R. Haeb-Umbach, “Jointly Optimal
Dereverberation and Beamforming,” 2020, doi: 10.1109/icassp40776.2020.9054393.'
mla: Boeddeker, Christoph, et al. “Jointly Optimal Dereverberation and Beamforming.”
ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal
Processing (ICASSP), 2020, doi:10.1109/icassp40776.2020.9054393.
short: 'C. Boeddeker, T. Nakatani, K. Kinoshita, R. Haeb-Umbach, in: ICASSP 2020
- 2020 IEEE International Conference on Acoustics, Speech and Signal Processing
(ICASSP), 2020.'
date_created: 2020-12-11T12:28:49Z
date_updated: 2024-11-14T09:17:32Z
ddc:
- '000'
department:
- _id: '54'
doi: 10.1109/icassp40776.2020.9054393
file:
- access_level: open_access
content_type: application/pdf
creator: cbj
date_created: 2020-12-11T12:32:44Z
date_updated: 2020-12-11T12:32:44Z
file_id: '20698'
file_name: convBF.pdf
file_size: 200127
relation: main_file
file_date_updated: 2020-12-11T12:32:44Z
has_accepted_license: '1'
language:
- iso: eng
oa: '1'
project:
- _id: '52'
name: Computing Resources Provided by the Paderborn Center for Parallel Computing
publication: ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech
and Signal Processing (ICASSP)
publication_identifier:
isbn:
- '9781509066315'
publication_status: published
status: public
title: Jointly Optimal Dereverberation and Beamforming
type: conference
user_id: '40767'
year: '2020'
...
---
_id: '17762'
abstract:
- lang: eng
text: 'Abstract Wenn akustische Signalverarbeitung mit automatisiertem Lernen verknüpft
wird: Nachrichtentechniker arbeiten mit mehreren Mikrofonen und tiefen neuronalen
Netzen an besserer Spracherkennung unter widrigsten Bedingungen. Von solchen Sensornetzwerken
könnten langfristig auch digitale Sprachassistenten profitieren.'
author:
- first_name: Reinhold
full_name: Haeb-Umbach, Reinhold
id: '242'
last_name: Haeb-Umbach
citation:
ama: Haeb-Umbach R. Lektionen für Alexa \& Co?! forschung. 2019;44(1):12-15.
doi:10.1002/fors.201970104
apa: Haeb-Umbach, R. (2019). Lektionen für Alexa \& Co?! Forschung, 44(1),
12–15. https://doi.org/10.1002/fors.201970104
bibtex: '@article{Haeb-Umbach_2019, title={Lektionen für Alexa \& Co?!}, volume={44},
DOI={10.1002/fors.201970104},
number={1}, journal={forschung}, author={Haeb-Umbach, Reinhold}, year={2019},
pages={12–15} }'
chicago: 'Haeb-Umbach, Reinhold. “Lektionen Für Alexa \& Co?!” Forschung
44, no. 1 (2019): 12–15. https://doi.org/10.1002/fors.201970104.'
ieee: R. Haeb-Umbach, “Lektionen für Alexa \& Co?!,” forschung, vol.
44, no. 1, pp. 12–15, 2019.
mla: Haeb-Umbach, Reinhold. “Lektionen Für Alexa \& Co?!” Forschung,
vol. 44, no. 1, 2019, pp. 12–15, doi:10.1002/fors.201970104.
short: R. Haeb-Umbach, Forschung 44 (2019) 12–15.
date_created: 2020-08-10T09:51:09Z
date_updated: 2022-01-06T06:53:19Z
department:
- _id: '54'
doi: 10.1002/fors.201970104
intvolume: ' 44'
issue: '1'
language:
- iso: eng
page: 12-15
publication: forschung
status: public
title: Lektionen für Alexa \& Co?!
type: journal_article
user_id: '44006'
volume: 44
year: '2019'
...
---
_id: '19446'
abstract:
- lang: eng
text: 'We present a multi-channel database of overlapping speech for training, evaluation,
and detailed analysis of source separation and extraction algorithms: SMS-WSJ
-- Spatialized Multi-Speaker Wall Street Journal. It consists of artificially
mixed speech taken from the WSJ database, but unlike earlier databases we consider
all WSJ0+1 utterances and take care of strictly separating the speaker sets present
in the training, validation and test sets. When spatializing the data we ensure
a high degree of randomness w.r.t. room size, array center and rotation, as well
as speaker position. Furthermore, this paper offers a critical assessment of recently
proposed measures of source separation performance. Alongside the code to generate
the database we provide a source separation baseline and a Kaldi recipe with competitive
word error rates to provide common ground for evaluation.'
author:
- first_name: Lukas
full_name: Drude, Lukas
last_name: Drude
- first_name: Jens
full_name: Heitkaemper, Jens
id: '27643'
last_name: Heitkaemper
- first_name: Christoph
full_name: Boeddeker, Christoph
id: '40767'
last_name: Boeddeker
- first_name: Reinhold
full_name: Haeb-Umbach, Reinhold
id: '242'
last_name: Haeb-Umbach
citation:
ama: 'Drude L, Heitkaemper J, Boeddeker C, Haeb-Umbach R. SMS-WSJ: Database, performance
measures, and baseline recipe for multi-channel source separation and recognition.
ArXiv e-prints. 2019.'
apa: 'Drude, L., Heitkaemper, J., Boeddeker, C., & Haeb-Umbach, R. (2019). SMS-WSJ:
Database, performance measures, and baseline recipe for multi-channel source separation
and recognition. ArXiv E-Prints.'
bibtex: '@article{Drude_Heitkaemper_Boeddeker_Haeb-Umbach_2019, title={SMS-WSJ:
Database, performance measures, and baseline recipe for multi-channel source separation
and recognition}, journal={ArXiv e-prints}, author={Drude, Lukas and Heitkaemper,
Jens and Boeddeker, Christoph and Haeb-Umbach, Reinhold}, year={2019} }'
chicago: 'Drude, Lukas, Jens Heitkaemper, Christoph Boeddeker, and Reinhold Haeb-Umbach.
“SMS-WSJ: Database, Performance Measures, and Baseline Recipe for Multi-Channel
Source Separation and Recognition.” ArXiv E-Prints, 2019.'
ieee: 'L. Drude, J. Heitkaemper, C. Boeddeker, and R. Haeb-Umbach, “SMS-WSJ: Database,
performance measures, and baseline recipe for multi-channel source separation
and recognition,” ArXiv e-prints, 2019.'
mla: 'Drude, Lukas, et al. “SMS-WSJ: Database, Performance Measures, and Baseline
Recipe for Multi-Channel Source Separation and Recognition.” ArXiv E-Prints,
2019.'
short: L. Drude, J. Heitkaemper, C. Boeddeker, R. Haeb-Umbach, ArXiv E-Prints (2019).
date_created: 2020-09-16T07:59:46Z
date_updated: 2022-01-06T06:54:04Z
ddc:
- '000'
department:
- _id: '54'
file:
- access_level: open_access
content_type: application/pdf
creator: huesera
date_created: 2020-09-16T08:00:56Z
date_updated: 2020-12-11T12:22:31Z
file_id: '19448'
file_name: ArXiv_2019_Drude.pdf
file_size: 288594
relation: main_file
file_date_updated: 2020-12-11T12:22:31Z
has_accepted_license: '1'
language:
- iso: eng
oa: '1'
project:
- _id: '52'
name: Computing Resources Provided by the Paderborn Center for Parallel Computing
publication: ArXiv e-prints
status: public
title: 'SMS-WSJ: Database, performance measures, and baseline recipe for multi-channel
source separation and recognition'
type: journal_article
user_id: '40767'
year: '2019'
...
---
_id: '11965'
abstract:
- lang: eng
text: 'We present an unsupervised training approach for a neural network-based mask
estimator in an acoustic beamforming application. The network is trained to maximize
a likelihood criterion derived from a spatial mixture model of the observations.
It is trained from scratch without requiring any parallel data consisting of degraded
input and clean training targets. Thus, training can be carried out on real recordings
of noisy speech rather than simulated ones. In contrast to previous work on unsupervised
training of neural mask estimators, our approach avoids the need for a possibly
pre-trained teacher model entirely. We demonstrate the effectiveness of our approach
by speech recognition experiments on two different datasets: one mainly deteriorated
by noise (CHiME 4) and one by reverberation (REVERB). The results show that the
performance of the proposed system is on par with a supervised system using oracle
target masks for training and with a system trained using a model-based teacher.'
author:
- first_name: Lukas
full_name: Drude, Lukas
id: '11213'
last_name: Drude
- first_name: Jahn
full_name: Heymann, Jahn
id: '9168'
last_name: Heymann
- first_name: Reinhold
full_name: Haeb-Umbach, Reinhold
id: '242'
last_name: Haeb-Umbach
citation:
ama: 'Drude L, Heymann J, Haeb-Umbach R. Unsupervised training of neural mask-based
beamforming. In: INTERSPEECH 2019, Graz, Austria. ; 2019.'
apa: Drude, L., Heymann, J., & Haeb-Umbach, R. (2019). Unsupervised training
of neural mask-based beamforming. In INTERSPEECH 2019, Graz, Austria.
bibtex: '@inproceedings{Drude_Heymann_Haeb-Umbach_2019, title={Unsupervised training
of neural mask-based beamforming}, booktitle={INTERSPEECH 2019, Graz, Austria},
author={Drude, Lukas and Heymann, Jahn and Haeb-Umbach, Reinhold}, year={2019}
}'
chicago: Drude, Lukas, Jahn Heymann, and Reinhold Haeb-Umbach. “Unsupervised Training
of Neural Mask-Based Beamforming.” In INTERSPEECH 2019, Graz, Austria,
2019.
ieee: L. Drude, J. Heymann, and R. Haeb-Umbach, “Unsupervised training of neural
mask-based beamforming,” in INTERSPEECH 2019, Graz, Austria, 2019.
mla: Drude, Lukas, et al. “Unsupervised Training of Neural Mask-Based Beamforming.”
INTERSPEECH 2019, Graz, Austria, 2019.
short: 'L. Drude, J. Heymann, R. Haeb-Umbach, in: INTERSPEECH 2019, Graz, Austria,
2019.'
date_created: 2019-07-18T09:11:39Z
date_updated: 2022-01-06T06:51:14Z
ddc:
- '000'
department:
- _id: '54'
file:
- access_level: open_access
content_type: application/pdf
creator: huesera
date_created: 2019-08-13T06:36:44Z
date_updated: 2019-08-13T06:41:35Z
file_id: '12914'
file_name: INTERSPEECH_2019_Drude_Paper.pdf
file_size: 223413
relation: main_file
file_date_updated: 2019-08-13T06:41:35Z
has_accepted_license: '1'
language:
- iso: eng
oa: '1'
project:
- _id: '52'
name: Computing Resources Provided by the Paderborn Center for Parallel Computing
publication: INTERSPEECH 2019, Graz, Austria
status: public
title: Unsupervised training of neural mask-based beamforming
type: conference
user_id: '59789'
year: '2019'
...
---
_id: '12874'
abstract:
- lang: eng
text: We propose a training scheme to train neural network-based source separation
algorithms from scratch when parallel clean data is unavailable. In particular,
we demonstrate that an unsupervised spatial clustering algorithm is sufficient
to guide the training of a deep clustering system. We argue that previous work
on deep clustering requires strong supervision and elaborate on why this is a
limitation. We demonstrate that (a) the single-channel deep clustering system
trained according to the proposed scheme alone is able to achieve a similar performance
as the multi-channel teacher in terms of word error rates and (b) initializing
the spatial clustering approach with the deep clustering result yields a relative
word error rate reduction of 26% over the unsupervised teacher.
author:
- first_name: Lukas
full_name: Drude, Lukas
id: '11213'
last_name: Drude
- first_name: Daniel
full_name: Hasenklever, Daniel
last_name: Hasenklever
- first_name: Reinhold
full_name: Haeb-Umbach, Reinhold
id: '242'
last_name: Haeb-Umbach
citation:
ama: 'Drude L, Hasenklever D, Haeb-Umbach R. Unsupervised Training of a Deep Clustering
Model for Multichannel Blind Source Separation. In: ICASSP 2019, Brighton,
UK. ; 2019.'
apa: Drude, L., Hasenklever, D., & Haeb-Umbach, R. (2019). Unsupervised Training
of a Deep Clustering Model for Multichannel Blind Source Separation. In ICASSP
2019, Brighton, UK.
bibtex: '@inproceedings{Drude_Hasenklever_Haeb-Umbach_2019, title={Unsupervised
Training of a Deep Clustering Model for Multichannel Blind Source Separation},
booktitle={ICASSP 2019, Brighton, UK}, author={Drude, Lukas and Hasenklever, Daniel
and Haeb-Umbach, Reinhold}, year={2019} }'
chicago: Drude, Lukas, Daniel Hasenklever, and Reinhold Haeb-Umbach. “Unsupervised
Training of a Deep Clustering Model for Multichannel Blind Source Separation.”
In ICASSP 2019, Brighton, UK, 2019.
ieee: L. Drude, D. Hasenklever, and R. Haeb-Umbach, “Unsupervised Training of a
Deep Clustering Model for Multichannel Blind Source Separation,” in ICASSP
2019, Brighton, UK, 2019.
mla: Drude, Lukas, et al. “Unsupervised Training of a Deep Clustering Model for
Multichannel Blind Source Separation.” ICASSP 2019, Brighton, UK, 2019.
short: 'L. Drude, D. Hasenklever, R. Haeb-Umbach, in: ICASSP 2019, Brighton, UK,
2019.'
date_created: 2019-07-23T07:37:54Z
date_updated: 2022-01-06T06:51:21Z
ddc:
- '000'
department:
- _id: '54'
file:
- access_level: open_access
content_type: application/pdf
creator: huesera
date_created: 2019-08-14T07:19:13Z
date_updated: 2019-08-14T07:19:13Z
file_id: '12925'
file_name: ICASSP_2019_Drude_Paper.pdf
file_size: 368225
relation: main_file
file_date_updated: 2019-08-14T07:19:13Z
has_accepted_license: '1'
language:
- iso: eng
oa: '1'
project:
- _id: '52'
name: Computing Resources Provided by the Paderborn Center for Parallel Computing
publication: ICASSP 2019, Brighton, UK
status: public
title: Unsupervised Training of a Deep Clustering Model for Multichannel Blind Source
Separation
type: conference
user_id: '59789'
year: '2019'
...
---
_id: '12875'
abstract:
- lang: eng
text: Signal dereverberation using the Weighted Prediction Error (WPE) method has
been proven to be an effective means to raise the accuracy of far-field speech
recognition. First proposed as an iterative algorithm, follow-up works have reformulated
it as a recursive least squares algorithm and therefore enabled its use in online
applications. For this algorithm, the estimation of the power spectral density
(PSD) of the anechoic signal plays an important role and strongly influences its
performance. Recently, we showed that using a neural network PSD estimator leads
to improved performance for online automatic speech recognition. This, however,
comes at a price. To train the network, we require parallel data, i.e., utterances
simultaneously available in clean and reverberated form. Here we propose to overcome
this limitation by training the network jointly with the acoustic model of the
speech recognizer. To be specific, the gradients computed from the cross-entropy
loss between the target senone sequence and the acoustic model network output
is backpropagated through the complex-valued dereverberation filter estimation
to the neural network for PSD estimation. Evaluation on two databases demonstrates
improved performance for on-line processing scenarios while imposing fewer requirements
on the available training data and thus widening the range of applications.
author:
- first_name: Jahn
full_name: Heymann, Jahn
id: '9168'
last_name: Heymann
- first_name: Lukas
full_name: Drude, Lukas
id: '11213'
last_name: Drude
- first_name: Reinhold
full_name: Haeb-Umbach, Reinhold
id: '242'
last_name: Haeb-Umbach
- first_name: Keisuke
full_name: Kinoshita, Keisuke
last_name: Kinoshita
- first_name: Tomohiro
full_name: Nakatani, Tomohiro
last_name: Nakatani
citation:
ama: 'Heymann J, Drude L, Haeb-Umbach R, Kinoshita K, Nakatani T. Joint Optimization
of Neural Network-based WPE Dereverberation and Acoustic Model for Robust Online
ASR. In: ICASSP 2019, Brighton, UK. ; 2019.'
apa: Heymann, J., Drude, L., Haeb-Umbach, R., Kinoshita, K., & Nakatani, T.
(2019). Joint Optimization of Neural Network-based WPE Dereverberation and Acoustic
Model for Robust Online ASR. In ICASSP 2019, Brighton, UK.
bibtex: '@inproceedings{Heymann_Drude_Haeb-Umbach_Kinoshita_Nakatani_2019, title={Joint
Optimization of Neural Network-based WPE Dereverberation and Acoustic Model for
Robust Online ASR}, booktitle={ICASSP 2019, Brighton, UK}, author={Heymann, Jahn
and Drude, Lukas and Haeb-Umbach, Reinhold and Kinoshita, Keisuke and Nakatani,
Tomohiro}, year={2019} }'
chicago: Heymann, Jahn, Lukas Drude, Reinhold Haeb-Umbach, Keisuke Kinoshita, and
Tomohiro Nakatani. “Joint Optimization of Neural Network-Based WPE Dereverberation
and Acoustic Model for Robust Online ASR.” In ICASSP 2019, Brighton, UK,
2019.
ieee: J. Heymann, L. Drude, R. Haeb-Umbach, K. Kinoshita, and T. Nakatani, “Joint
Optimization of Neural Network-based WPE Dereverberation and Acoustic Model for
Robust Online ASR,” in ICASSP 2019, Brighton, UK, 2019.
mla: Heymann, Jahn, et al. “Joint Optimization of Neural Network-Based WPE Dereverberation
and Acoustic Model for Robust Online ASR.” ICASSP 2019, Brighton, UK, 2019.
short: 'J. Heymann, L. Drude, R. Haeb-Umbach, K. Kinoshita, T. Nakatani, in: ICASSP
2019, Brighton, UK, 2019.'
date_created: 2019-07-23T07:42:26Z
date_updated: 2022-01-06T06:51:22Z
ddc:
- '000'
department:
- _id: '54'
file:
- access_level: open_access
content_type: application/pdf
creator: huesera
date_created: 2019-12-17T07:28:06Z
date_updated: 2019-12-17T07:28:06Z
file_id: '15334'
file_name: ICASSP_2019_Heymann_Paper.pdf
file_size: 199109
relation: main_file
file_date_updated: 2019-12-17T07:28:06Z
has_accepted_license: '1'
language:
- iso: eng
oa: '1'
project:
- _id: '52'
name: Computing Resources Provided by the Paderborn Center for Parallel Computing
publication: ICASSP 2019, Brighton, UK
status: public
title: Joint Optimization of Neural Network-based WPE Dereverberation and Acoustic
Model for Robust Online ASR
type: conference
user_id: '59789'
year: '2019'
...
---
_id: '12876'
abstract:
- lang: eng
text: In this paper, we present libDirectional, a MATLAB library for directional
statistics and directional estimation. It supports a variety of commonly used
distributions on the unit circle, such as the von Mises, wrapped normal, and wrapped
Cauchy distributions. Furthermore, various distributions on higher-dimensional
manifolds such as the unit hypersphere and the hypertorus are available. Based
on these distributions, several recursive filtering algorithms in libDirectional
allow estimation on these manifolds. The functionality is implemented in a clear,
well-documented, and object-oriented structure that is both easy to use and easy
to extend.
author:
- first_name: Gerhard
full_name: Kurz, Gerhard
last_name: Kurz
- first_name: Igor
full_name: Gilitschenski, Igor
last_name: Gilitschenski
- first_name: Florian
full_name: Pfaff, Florian
last_name: Pfaff
- first_name: Lukas
full_name: Drude, Lukas
id: '11213'
last_name: Drude
- first_name: Uwe D.
full_name: Hanebeck, Uwe D.
last_name: Hanebeck
- first_name: Reinhold
full_name: Haeb-Umbach, Reinhold
id: '242'
last_name: Haeb-Umbach
- first_name: Roland Y.
full_name: Siegwart, Roland Y.
last_name: Siegwart
citation:
ama: 'Kurz G, Gilitschenski I, Pfaff F, et al. Directional Statistics and Filtering
Using libDirectional. In: Journal of Statistical Software 89(4). ; 2019.'
apa: Kurz, G., Gilitschenski, I., Pfaff, F., Drude, L., Hanebeck, U. D., Haeb-Umbach,
R., & Siegwart, R. Y. (2019). Directional Statistics and Filtering Using libDirectional.
In Journal of Statistical Software 89(4).
bibtex: '@inproceedings{Kurz_Gilitschenski_Pfaff_Drude_Hanebeck_Haeb-Umbach_Siegwart_2019,
title={Directional Statistics and Filtering Using libDirectional}, booktitle={Journal
of Statistical Software 89(4)}, author={Kurz, Gerhard and Gilitschenski, Igor
and Pfaff, Florian and Drude, Lukas and Hanebeck, Uwe D. and Haeb-Umbach, Reinhold
and Siegwart, Roland Y.}, year={2019} }'
chicago: Kurz, Gerhard, Igor Gilitschenski, Florian Pfaff, Lukas Drude, Uwe D. Hanebeck,
Reinhold Haeb-Umbach, and Roland Y. Siegwart. “Directional Statistics and Filtering
Using LibDirectional.” In Journal of Statistical Software 89(4), 2019.
ieee: G. Kurz et al., “Directional Statistics and Filtering Using libDirectional,”
in Journal of Statistical Software 89(4), 2019.
mla: Kurz, Gerhard, et al. “Directional Statistics and Filtering Using LibDirectional.”
Journal of Statistical Software 89(4), 2019.
short: 'G. Kurz, I. Gilitschenski, F. Pfaff, L. Drude, U.D. Hanebeck, R. Haeb-Umbach,
R.Y. Siegwart, in: Journal of Statistical Software 89(4), 2019.'
date_created: 2019-07-23T07:44:59Z
date_updated: 2022-01-06T06:51:22Z
ddc:
- '000'
department:
- _id: '54'
file:
- access_level: open_access
content_type: application/pdf
creator: huesera
date_created: 2019-08-14T07:16:05Z
date_updated: 2019-08-14T07:16:05Z
file_id: '12923'
file_name: JournalofStatisticalSoftware_2019_Drude_Paper.pdf
file_size: 1522964
relation: main_file
file_date_updated: 2019-08-14T07:16:05Z
has_accepted_license: '1'
language:
- iso: eng
oa: '1'
publication: Journal of Statistical Software 89(4)
status: public
title: Directional Statistics and Filtering Using libDirectional
type: conference
user_id: '59789'
year: '2019'
...
---
_id: '12890'
abstract:
- lang: eng
text: 'We formulate a generic framework for blind source separation (BSS), which
allows integrating data-driven spectro-temporal methods, such as deep clustering
and deep attractor networks, with physically motivated probabilistic spatial methods,
such as complex angular central Gaussian mixture models. The integrated model
exploits the complementary strengths of the two approaches to BSS: the strong
modeling power of neural networks, which, however, is based on supervised learning,
and the ease of unsupervised learning of the spatial mixture models whose few
parameters can be estimated on as little as a single segment of a real mixture
of speech. Experiments are carried out on both artificially mixed speech and true
recordings of speech mixtures. The experiments verify that the integrated models
consistently outperform the individual components. We further extend the models
to cope with noisy, reverberant speech and introduce a cross-domain teacher–student
training where the mixture model serves as the teacher to provide training targets
for the student neural network.'
author:
- first_name: Lukas
full_name: Drude, Lukas
id: '11213'
last_name: Drude
- first_name: Reinhold
full_name: Haeb-Umbach, Reinhold
id: '242'
last_name: Haeb-Umbach
citation:
ama: Drude L, Haeb-Umbach R. Integration of Neural Networks and Probabilistic Spatial
Models for Acoustic Blind Source Separation. IEEE Journal of Selected Topics
in Signal Processing. 2019. doi:10.1109/JSTSP.2019.2912565
apa: Drude, L., & Haeb-Umbach, R. (2019). Integration of Neural Networks and
Probabilistic Spatial Models for Acoustic Blind Source Separation. IEEE Journal
of Selected Topics in Signal Processing. https://doi.org/10.1109/JSTSP.2019.2912565
bibtex: '@article{Drude_Haeb-Umbach_2019, title={Integration of Neural Networks
and Probabilistic Spatial Models for Acoustic Blind Source Separation}, DOI={10.1109/JSTSP.2019.2912565},
journal={IEEE Journal of Selected Topics in Signal Processing}, author={Drude,
Lukas and Haeb-Umbach, Reinhold}, year={2019} }'
chicago: Drude, Lukas, and Reinhold Haeb-Umbach. “Integration of Neural Networks
and Probabilistic Spatial Models for Acoustic Blind Source Separation.” IEEE
Journal of Selected Topics in Signal Processing, 2019. https://doi.org/10.1109/JSTSP.2019.2912565.
ieee: L. Drude and R. Haeb-Umbach, “Integration of Neural Networks and Probabilistic
Spatial Models for Acoustic Blind Source Separation,” IEEE Journal of Selected
Topics in Signal Processing, 2019.
mla: Drude, Lukas, and Reinhold Haeb-Umbach. “Integration of Neural Networks and
Probabilistic Spatial Models for Acoustic Blind Source Separation.” IEEE Journal
of Selected Topics in Signal Processing, 2019, doi:10.1109/JSTSP.2019.2912565.
short: L. Drude, R. Haeb-Umbach, IEEE Journal of Selected Topics in Signal Processing
(2019).
date_created: 2019-07-26T08:38:46Z
date_updated: 2022-01-06T06:51:23Z
ddc:
- '050'
department:
- _id: '54'
doi: 10.1109/JSTSP.2019.2912565
file:
- access_level: open_access
content_type: application/pdf
creator: huesera
date_created: 2019-08-07T07:12:21Z
date_updated: 2019-08-14T07:11:22Z
file_id: '12903'
file_name: IEEE Jounal_2019_Drude_Paper.pdf
file_size: 967424
relation: main_file
file_date_updated: 2019-08-14T07:11:22Z
has_accepted_license: '1'
language:
- iso: eng
oa: '1'
project:
- _id: '52'
name: Computing Resources Provided by the Paderborn Center for Parallel Computing
publication: IEEE Journal of Selected Topics in Signal Processing
publication_identifier:
eissn:
- 1941-0484
status: public
title: Integration of Neural Networks and Probabilistic Spatial Models for Acoustic
Blind Source Separation
type: journal_article
user_id: '11213'
year: '2019'
...
---
_id: '15816'
abstract:
- lang: eng
text: 'Despite the strong modeling power of neural network acoustic models, speech
enhancement has been shown to deliver additional word error rate improvements
if multi-channel data is available. However, there has been a longstanding debate
whether enhancement should also be carried out on the ASR training data. In an
extensive experimental evaluation on the acoustically very challenging CHiME-5
dinner party data we show that: (i) cleaning up the training data can lead to
substantial error rate reductions, and (ii) enhancement in training is advisable
as long as enhancement in test is at least as strong as in training. This approach
stands in contrast and delivers larger gains than the common strategy reported
in the literature to augment the training database with additional artificially
degraded speech. Together with an acoustic model topology consisting of initial
CNN layers followed by factorized TDNN layers we achieve with 41.6% and 43.2%
WER on the DEV and EVAL test sets, respectively, a new single-system state-of-the-art
result on the CHiME-5 data. This is a 8% relative improvement compared to the
best word error rate published so far for a speech recognizer without system combination.'
author:
- first_name: Catalin
full_name: Zorila, Catalin
last_name: Zorila
- first_name: Christoph
full_name: Boeddeker, Christoph
id: '40767'
last_name: Boeddeker
- first_name: Rama
full_name: Doddipatla, Rama
last_name: Doddipatla
- first_name: Reinhold
full_name: Haeb-Umbach, Reinhold
id: '242'
last_name: Haeb-Umbach
citation:
ama: 'Zorila C, Boeddeker C, Doddipatla R, Haeb-Umbach R. An Investigation Into
the Effectiveness of Enhancement in ASR Training and Test for Chime-5 Dinner Party
Transcription. In: ASRU 2019, Sentosa, Singapore. ; 2019.'
apa: Zorila, C., Boeddeker, C., Doddipatla, R., & Haeb-Umbach, R. (2019). An
Investigation Into the Effectiveness of Enhancement in ASR Training and Test for
Chime-5 Dinner Party Transcription. In ASRU 2019, Sentosa, Singapore.
bibtex: '@inproceedings{Zorila_Boeddeker_Doddipatla_Haeb-Umbach_2019, title={An
Investigation Into the Effectiveness of Enhancement in ASR Training and Test for
Chime-5 Dinner Party Transcription}, booktitle={ASRU 2019, Sentosa, Singapore},
author={Zorila, Catalin and Boeddeker, Christoph and Doddipatla, Rama and Haeb-Umbach,
Reinhold}, year={2019} }'
chicago: Zorila, Catalin, Christoph Boeddeker, Rama Doddipatla, and Reinhold Haeb-Umbach.
“An Investigation Into the Effectiveness of Enhancement in ASR Training and Test
for Chime-5 Dinner Party Transcription.” In ASRU 2019, Sentosa, Singapore,
2019.
ieee: C. Zorila, C. Boeddeker, R. Doddipatla, and R. Haeb-Umbach, “An Investigation
Into the Effectiveness of Enhancement in ASR Training and Test for Chime-5 Dinner
Party Transcription,” in ASRU 2019, Sentosa, Singapore, 2019.
mla: Zorila, Catalin, et al. “An Investigation Into the Effectiveness of Enhancement
in ASR Training and Test for Chime-5 Dinner Party Transcription.” ASRU 2019,
Sentosa, Singapore, 2019.
short: 'C. Zorila, C. Boeddeker, R. Doddipatla, R. Haeb-Umbach, in: ASRU 2019, Sentosa,
Singapore, 2019.'
date_created: 2020-02-06T07:35:08Z
date_updated: 2022-01-06T06:52:37Z
ddc:
- '000'
department:
- _id: '54'
file:
- access_level: open_access
content_type: application/pdf
creator: huesera
date_created: 2020-02-06T07:42:42Z
date_updated: 2020-02-06T07:42:42Z
file_id: '15817'
file_name: ASRU_2019_Boeddeker_Paper.pdf
file_size: 200256
relation: main_file
- access_level: open_access
content_type: application/pdf
creator: huesera
date_created: 2020-02-06T07:42:55Z
date_updated: 2020-02-06T07:42:55Z
file_id: '15818'
file_name: ASRU_2019_Boeddeker_Poster.pdf
file_size: 123963
relation: main_file
file_date_updated: 2020-02-06T07:42:55Z
has_accepted_license: '1'
language:
- iso: eng
oa: '1'
project:
- _id: '52'
name: Computing Resources Provided by the Paderborn Center for Parallel Computing
publication: ASRU 2019, Sentosa, Singapore
status: public
title: An Investigation Into the Effectiveness of Enhancement in ASR Training and
Test for Chime-5 Dinner Party Transcription
type: conference
user_id: '59789'
year: '2019'
...
---
_id: '14822'
abstract:
- lang: eng
text: Multi-talker speech and moving speakers still pose a significant challenge
to automatic speech recognition systems. Assuming an enrollment utterance of the
target speakeris available, the so-called SpeakerBeam concept has been recently
proposed to extract the target speaker from a speech mixture. If multi-channel
input is available, spatial properties of the speaker can be exploited to support
the source extraction. In this contribution we investigate different approaches
to exploit such spatial information. In particular, we are interested in the question,
how useful this information is if the target speaker changes his/her position.
To this end, we present a SpeakerBeam-based source extraction network that is
adapted to work on moving speakers by recursively updating the beamformer coefficients.
Experimental results are presented on two data sets, one with articially created
room impulse responses, and one with real room impulse responses and noise recorded
in a conference room. Interestingly, spatial features turn out to be advantageous
even if the speaker position changes.
author:
- first_name: Jens
full_name: Heitkaemper, Jens
id: '27643'
last_name: Heitkaemper
- first_name: Thomas
full_name: Feher, Thomas
last_name: Feher
- first_name: Michael
full_name: Freitag, Michael
last_name: Freitag
- first_name: Reinhold
full_name: Haeb-Umbach, Reinhold
id: '242'
last_name: Haeb-Umbach
citation:
ama: 'Heitkaemper J, Feher T, Freitag M, Haeb-Umbach R. A Study on Online Source
Extraction in the Presence of Changing Speaker Positions. In: International
Conference on Statistical Language and Speech Processing 2019, Ljubljana, Slovenia.
; 2019.'
apa: Heitkaemper, J., Feher, T., Freitag, M., & Haeb-Umbach, R. (2019). A Study
on Online Source Extraction in the Presence of Changing Speaker Positions. In
International Conference on Statistical Language and Speech Processing 2019,
Ljubljana, Slovenia.
bibtex: '@inproceedings{Heitkaemper_Feher_Freitag_Haeb-Umbach_2019, title={A Study
on Online Source Extraction in the Presence of Changing Speaker Positions}, booktitle={International
Conference on Statistical Language and Speech Processing 2019, Ljubljana, Slovenia},
author={Heitkaemper, Jens and Feher, Thomas and Freitag, Michael and Haeb-Umbach,
Reinhold}, year={2019} }'
chicago: Heitkaemper, Jens, Thomas Feher, Michael Freitag, and Reinhold Haeb-Umbach.
“A Study on Online Source Extraction in the Presence of Changing Speaker Positions.”
In International Conference on Statistical Language and Speech Processing 2019,
Ljubljana, Slovenia, 2019.
ieee: J. Heitkaemper, T. Feher, M. Freitag, and R. Haeb-Umbach, “A Study on Online
Source Extraction in the Presence of Changing Speaker Positions,” in International
Conference on Statistical Language and Speech Processing 2019, Ljubljana, Slovenia,
2019.
mla: Heitkaemper, Jens, et al. “A Study on Online Source Extraction in the Presence
of Changing Speaker Positions.” International Conference on Statistical Language
and Speech Processing 2019, Ljubljana, Slovenia, 2019.
short: 'J. Heitkaemper, T. Feher, M. Freitag, R. Haeb-Umbach, in: International
Conference on Statistical Language and Speech Processing 2019, Ljubljana, Slovenia,
2019.'
date_created: 2019-11-06T09:43:03Z
date_updated: 2022-01-06T06:52:06Z
ddc:
- '006'
department:
- _id: '54'
file:
- access_level: open_access
content_type: application/pdf
creator: huesera
date_created: 2019-11-06T10:02:26Z
date_updated: 2019-11-08T07:47:12Z
file_id: '14823'
file_name: SLSP_2019_Heitkaemper_Paper.pdf
file_size: 578595
relation: main_file
file_date_updated: 2019-11-08T07:47:12Z
has_accepted_license: '1'
language:
- iso: eng
oa: '1'
project:
- _id: '52'
name: Computing Resources Provided by the Paderborn Center for Parallel Computing
publication: International Conference on Statistical Language and Speech Processing
2019, Ljubljana, Slovenia
status: public
title: A Study on Online Source Extraction in the Presence of Changing Speaker Positions
type: conference
user_id: '59789'
year: '2019'
...
---
_id: '14824'
abstract:
- lang: eng
text: This paper deals with multi-channel speech recognition in scenarios with multiple
speakers. Recently, the spectral characteristics of a target speaker, extracted
from an adaptation utterance, have been used to guide a neural network mask estimator
to focus on that speaker. In this work we present two variants of speakeraware
neural networks, which exploit both spectral and spatial information to allow
better discrimination between target and interfering speakers. Thus, we introduce
either a spatial preprocessing prior to the mask estimation or a spatial plus
spectral speaker characterization block whose output is directly fed into the
neural mask estimator. The target speaker’s spectral and spatial signature is
extracted from an adaptation utterance recorded at the beginning of a session.
We further adapt the architecture for low-latency processing by means of block-online
beamforming that recursively updates the signal statistics. Experimental results
show that the additional spatial information clearly improves source extraction,
in particular in the same-gender case, and that our proposal achieves state-of-the-art
performance in terms of distortion reduction and recognition accuracy.
author:
- first_name: Juan M.
full_name: Martin-Donas, Juan M.
last_name: Martin-Donas
- first_name: Jens
full_name: Heitkaemper, Jens
id: '27643'
last_name: Heitkaemper
- first_name: Reinhold
full_name: Haeb-Umbach, Reinhold
id: '242'
last_name: Haeb-Umbach
- first_name: Angel M.
full_name: Gomez, Angel M.
last_name: Gomez
- first_name: Antonio M.
full_name: Peinado, Antonio M.
last_name: Peinado
citation:
ama: 'Martin-Donas JM, Heitkaemper J, Haeb-Umbach R, Gomez AM, Peinado AM. Multi-Channel
Block-Online Source Extraction based on Utterance Adaptation. In: INTERSPEECH
2019, Graz, Austria. ; 2019.'
apa: Martin-Donas, J. M., Heitkaemper, J., Haeb-Umbach, R., Gomez, A. M., &
Peinado, A. M. (2019). Multi-Channel Block-Online Source Extraction based on Utterance
Adaptation. In INTERSPEECH 2019, Graz, Austria.
bibtex: '@inproceedings{Martin-Donas_Heitkaemper_Haeb-Umbach_Gomez_Peinado_2019,
title={Multi-Channel Block-Online Source Extraction based on Utterance Adaptation},
booktitle={INTERSPEECH 2019, Graz, Austria}, author={Martin-Donas, Juan M. and
Heitkaemper, Jens and Haeb-Umbach, Reinhold and Gomez, Angel M. and Peinado, Antonio
M.}, year={2019} }'
chicago: Martin-Donas, Juan M., Jens Heitkaemper, Reinhold Haeb-Umbach, Angel M.
Gomez, and Antonio M. Peinado. “Multi-Channel Block-Online Source Extraction Based
on Utterance Adaptation.” In INTERSPEECH 2019, Graz, Austria, 2019.
ieee: J. M. Martin-Donas, J. Heitkaemper, R. Haeb-Umbach, A. M. Gomez, and A. M.
Peinado, “Multi-Channel Block-Online Source Extraction based on Utterance Adaptation,”
in INTERSPEECH 2019, Graz, Austria, 2019.
mla: Martin-Donas, Juan M., et al. “Multi-Channel Block-Online Source Extraction
Based on Utterance Adaptation.” INTERSPEECH 2019, Graz, Austria, 2019.
short: 'J.M. Martin-Donas, J. Heitkaemper, R. Haeb-Umbach, A.M. Gomez, A.M. Peinado,
in: INTERSPEECH 2019, Graz, Austria, 2019.'
date_created: 2019-11-06T10:04:49Z
date_updated: 2022-01-06T06:52:07Z
ddc:
- '000'
department:
- _id: '54'
file:
- access_level: open_access
content_type: application/pdf
creator: huesera
date_created: 2019-11-06T10:07:15Z
date_updated: 2019-11-08T07:46:37Z
file_id: '14825'
file_name: INTERSPEECH_2019_Heitkaemper_Paper.pdf
file_size: 225689
relation: main_file
file_date_updated: 2019-11-08T07:46:37Z
has_accepted_license: '1'
language:
- iso: eng
oa: '1'
project:
- _id: '52'
name: Computing Resources Provided by the Paderborn Center for Parallel Computing
publication: INTERSPEECH 2019, Graz, Austria
status: public
title: Multi-Channel Block-Online Source Extraction based on Utterance Adaptation
type: conference
user_id: '59789'
year: '2019'
...
---
_id: '14826'
abstract:
- lang: eng
text: In this paper, we present Hitachi and Paderborn University’s joint effort
for automatic speech recognition (ASR) in a dinner party scenario. The main challenges
of ASR systems for dinner party recordings obtained by multiple microphone arrays
are (1) heavy speech overlaps, (2) severe noise and reverberation, (3) very natural
onversational content, and possibly (4) insufficient training data. As an example
of a dinner party scenario, we have chosen the data presented during the CHiME-5
speech recognition challenge, where the baseline ASR had a 73.3% word error rate
(WER), and even the best performing system at the CHiME-5 challenge had a 46.1%
WER. We extensively investigated a combination of the guided source separation-based
speech enhancement technique and an already proposed strong ASR backend and found
that a tight combination of these techniques provided substantial accuracy improvements.
Our final system achieved WERs of 39.94% and 41.64% for the development and evaluation
data, respectively, both of which are the best published results for the dataset.
We also investigated with additional training data on the official small data
in the CHiME-5 corpus to assess the intrinsic difficulty of this ASR task.
author:
- first_name: Naoyuki
full_name: Kanda, Naoyuki
last_name: Kanda
- first_name: Christoph
full_name: Boeddeker, Christoph
id: '40767'
last_name: Boeddeker
- first_name: Jens
full_name: Heitkaemper, Jens
id: '27643'
last_name: Heitkaemper
- first_name: Yusuke
full_name: Fujita, Yusuke
last_name: Fujita
- first_name: Shota
full_name: Horiguchi, Shota
last_name: Horiguchi
- first_name: Reinhold
full_name: Haeb-Umbach, Reinhold
id: '242'
last_name: Haeb-Umbach
citation:
ama: 'Kanda N, Boeddeker C, Heitkaemper J, Fujita Y, Horiguchi S, Haeb-Umbach R.
Guided Source Separation Meets a Strong ASR Backend: Hitachi/Paderborn University
Joint Investigation for Dinner Party ASR. In: INTERSPEECH 2019, Graz, Austria.
; 2019.'
apa: 'Kanda, N., Boeddeker, C., Heitkaemper, J., Fujita, Y., Horiguchi, S., &
Haeb-Umbach, R. (2019). Guided Source Separation Meets a Strong ASR Backend: Hitachi/Paderborn
University Joint Investigation for Dinner Party ASR. In INTERSPEECH 2019, Graz,
Austria.'
bibtex: '@inproceedings{Kanda_Boeddeker_Heitkaemper_Fujita_Horiguchi_Haeb-Umbach_2019,
title={Guided Source Separation Meets a Strong ASR Backend: Hitachi/Paderborn
University Joint Investigation for Dinner Party ASR}, booktitle={INTERSPEECH 2019,
Graz, Austria}, author={Kanda, Naoyuki and Boeddeker, Christoph and Heitkaemper,
Jens and Fujita, Yusuke and Horiguchi, Shota and Haeb-Umbach, Reinhold}, year={2019}
}'
chicago: 'Kanda, Naoyuki, Christoph Boeddeker, Jens Heitkaemper, Yusuke Fujita,
Shota Horiguchi, and Reinhold Haeb-Umbach. “Guided Source Separation Meets a Strong
ASR Backend: Hitachi/Paderborn University Joint Investigation for Dinner Party
ASR.” In INTERSPEECH 2019, Graz, Austria, 2019.'
ieee: 'N. Kanda, C. Boeddeker, J. Heitkaemper, Y. Fujita, S. Horiguchi, and R. Haeb-Umbach,
“Guided Source Separation Meets a Strong ASR Backend: Hitachi/Paderborn University
Joint Investigation for Dinner Party ASR,” in INTERSPEECH 2019, Graz, Austria,
2019.'
mla: 'Kanda, Naoyuki, et al. “Guided Source Separation Meets a Strong ASR Backend:
Hitachi/Paderborn University Joint Investigation for Dinner Party ASR.” INTERSPEECH
2019, Graz, Austria, 2019.'
short: 'N. Kanda, C. Boeddeker, J. Heitkaemper, Y. Fujita, S. Horiguchi, R. Haeb-Umbach,
in: INTERSPEECH 2019, Graz, Austria, 2019.'
date_created: 2019-11-06T10:08:49Z
date_updated: 2022-01-06T06:52:07Z
ddc:
- '000'
department:
- _id: '54'
file:
- access_level: open_access
content_type: application/pdf
creator: huesera
date_created: 2019-11-06T10:10:23Z
date_updated: 2019-11-08T07:45:15Z
file_id: '14827'
file_name: INTERSPEECH_2019_Boeddeker_Paper.pdf
file_size: 216202
relation: main_file
file_date_updated: 2019-11-08T07:45:15Z
has_accepted_license: '1'
language:
- iso: eng
oa: '1'
project:
- _id: '52'
name: Computing Resources Provided by the Paderborn Center for Parallel Computing
publication: INTERSPEECH 2019, Graz, Austria
status: public
title: 'Guided Source Separation Meets a Strong ASR Backend: Hitachi/Paderborn University
Joint Investigation for Dinner Party ASR'
type: conference
user_id: '59789'
year: '2019'
...