Environment-aware ideal binary mask estimation using monaural cues

Tobias May; Torsten Dau

doi:10.1109/WASPAA.2013.6701821

Environment-aware ideal binary mask estimation using monaural cues

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

Abstract

We present a monaural approach to speech segregation that estimates the ideal binary mask (IBM) by combining amplitude modulation spectrogram (AMS) features, pitch-based features and speech presence probability (SPP) features derived from noise statistics. To maintain a high mask estimation accuracy in the presence of various background noises, the system employs environment-specific segregation models and automatically selects the appropriate model for a given input signal. Furthermore, instead of classifying each timefrequency (T-F) unit independently, the a posteriori probabilities of speech and noise presence are evaluated by considering adjacent TF units. The proposed system achieves high classification accuracy.

Original language	English
Title of host publication	2013 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics
Number of pages	4
Publisher	IEEE
Publication date	2013
DOIs	https://doi.org/10.1109/WASPAA.2013.6701821
Publication status	Published - 2013
Event	2013 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics - Mohonk Mountain House , New Paltz, United States Duration: 20 Oct 2013 → 23 Oct 2013

Conference

Conference	2013 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics
Location	Mohonk Mountain House
Country/Territory	United States
City	New Paltz
Period	20/10/2013 → 23/10/2013

Keywords

Ideal binary mask estimation
Speech segregation
Background noise classification

Access to Document

10.1109/WASPAA.2013.6701821

OpenUrl availability

Check availability in DTU Findit

Cite this

@inproceedings{b86610e486394795a952294c0b50c457,

title = "Environment-aware ideal binary mask estimation using monaural cues",

abstract = "We present a monaural approach to speech segregation that estimates the ideal binary mask (IBM) by combining amplitude modulation spectrogram (AMS) features, pitch-based features and speech presence probability (SPP) features derived from noise statistics. To maintain a high mask estimation accuracy in the presence of various background noises, the system employs environment-specific segregation models and automatically selects the appropriate model for a given input signal. Furthermore, instead of classifying each timefrequency (T-F) unit independently, the a posteriori probabilities of speech and noise presence are evaluated by considering adjacent TF units. The proposed system achieves high classification accuracy.",

keywords = "Ideal binary mask estimation, Speech segregation, Background noise classification",

author = "Tobias May and Torsten Dau",

year = "2013",

doi = "10.1109/WASPAA.2013.6701821",

language = "English",

booktitle = "2013 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics",

publisher = "IEEE",

address = "United States",

note = "2013 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics, WASPAA 2013 ; Conference date: 20-10-2013 Through 23-10-2013",

}

TY - GEN

T1 - Environment-aware ideal binary mask estimation using monaural cues

AU - May, Tobias

AU - Dau, Torsten

PY - 2013

Y1 - 2013

N2 - We present a monaural approach to speech segregation that estimates the ideal binary mask (IBM) by combining amplitude modulation spectrogram (AMS) features, pitch-based features and speech presence probability (SPP) features derived from noise statistics. To maintain a high mask estimation accuracy in the presence of various background noises, the system employs environment-specific segregation models and automatically selects the appropriate model for a given input signal. Furthermore, instead of classifying each timefrequency (T-F) unit independently, the a posteriori probabilities of speech and noise presence are evaluated by considering adjacent TF units. The proposed system achieves high classification accuracy.

AB - We present a monaural approach to speech segregation that estimates the ideal binary mask (IBM) by combining amplitude modulation spectrogram (AMS) features, pitch-based features and speech presence probability (SPP) features derived from noise statistics. To maintain a high mask estimation accuracy in the presence of various background noises, the system employs environment-specific segregation models and automatically selects the appropriate model for a given input signal. Furthermore, instead of classifying each timefrequency (T-F) unit independently, the a posteriori probabilities of speech and noise presence are evaluated by considering adjacent TF units. The proposed system achieves high classification accuracy.

KW - Ideal binary mask estimation

KW - Speech segregation

KW - Background noise classification

U2 - 10.1109/WASPAA.2013.6701821

DO - 10.1109/WASPAA.2013.6701821

M3 - Article in proceedings

BT - 2013 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics

PB - IEEE

T2 - 2013 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics

Y2 - 20 October 2013 through 23 October 2013

ER -

Environment-aware ideal binary mask estimation using monaural cues

Abstract

Conference

Keywords

Access to Document

OpenUrl availability

Fingerprint

Cite this