Blind Speech Separation

By Shoji Makino, Te-Won Lee, Hiroshi Sawada. 2007

400 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Signal Processing

Citations in the corpus

Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech Separation (2018, background): “arability of the target and interfering sources. This aspect, however, limits their applicability when all the sources are speech signals and therefore have similar statistical properties [35], [36], [37], [38]. Beyond ICA methods, several recent studies have investigated the feasibility of time-dom”
The DIRHA-ENGLISH corpus and related tasks for distant-speech recognition in domestic environments (2015, background): “…considerable progresses were made at multi-microphone front-end processing level in order to feed ASR with an enhanced speech input [4, 5, 6, 7, 8], the performance loss observed from close-talking to distant-speech remains quite critical, even when the most advanced DNN-based backend…”
The Conversation: Deep Audio-Visual Speech Enhancement (2018, background): “ous works have proposed methods to isolate multi-talker simultaneous speech. The majority of these are based on methods that only use the audio, e.g. by using voice characteristics of a known speaker [3,4,5,6,7]. Compared to audio-only methods, we not only separate the voices but also properly assig”
Batch-normalized joint training for DNN-based distant speech recognition (2016, background): “…most recent DSR systems [9] must rely on a combination of several interconnected technologies, including for instance speech enhancement [10], speech separation [11], acoustic event detection and classiﬁcation [12, 13], speaker identiﬁcation [14], speaker localization [15, 16], just to name a few.”
Comparing Measures of Sparsity (2008, methodology): “For example, one method of source separation is to transform the signal to a domain in which it is sparse (e.g. time-frequency or wavelet) where the separation can be performed by a partition of the transformed signal space due to the sparsity of the representation [20], [21].”
Underdetermined Convolutive Blind Source Separation via Frequency Bin-Wise Clustering and Permutation Alignment (2011, methodology): “HE technique for estimating individual source components from their mixtures at multiple sensors is known as blind source separation (BSS) [1]–[5].”
Multichannel music separation with deep neural networks (2016, neutral): “It has many interesting applications, including music editing/remixing, upmixing, music information retrieval, and karaoke [1]–[3].”
An Overview of Noise-Robust Automatic Speech Recognition (2014, background): “The Aurora 3 task consists of noisy speech data recorded inside cars as part of the SpeechDatCar project [27].”
Realistic Multi-Microphone Data Simulation for Distant Speech Recognition (2016, background): “Despite the recent substantial progress in various related ﬁelds, including spatial ﬁltering [1, 2], microphone selection [3], source separation [4], speech dereverberation [5], speaker localization [6], acoustic event detection [7] as well as acoustic modeling [8, 9, 10, 11, 12], DSR still…”
The Diverse Environments Multi-channel Acoustic Noise Database (DEMAND): A database of multichannel environmental noise recordings (2013, neutral): “As a result, there is continued interest in developing methods to disentangle some sound of interest from that background noise [1, 2, 3].”
The PASCAL CHiME speech separation and recognition challenge (2013, neutral): “The topics of blind and then semi-blind source separation have emerged as research problems in their own right with their own associated conferences and evaluations (Makino et al., 2007; Vincent et al., 2012).”
Contaminated speech training methods for robust DNN-HMM distant speech recognition (2017, background): “ 288121 DIRHA. More details can be found in http://dirha.fbk.eu/. Considerable progresses were fostered by substantial advances in spatial ﬁltering [6, 7], microphone selection [8], source separation [9], speech dereverberation [10] as well as speaker localization [11], acoustic event detection [12]”
A network of deep neural networks for Distant Speech Recognition (2017, neutral): “To counteract such adversities, most DSR systems must rely on a combination of several interconnected technologies [6], including methods for speech enhancement [7], speech separation [8], acoustic event classiﬁcation [9,10], speaker identiﬁcation [11], just to name a few.”
Equalization matching of speech recordings in real-world environments (2016, background): “Separation unit Source separation is an active field of audio research [11] [12] [13] [14] [15].”
The second ‘chime’ speech separation and recognition challenge: Datasets, tasks and baselines (2013, background): “Indeed, ASR systems can be surprisingly sensitive to speaker location and it is well known that systems optimized for small vocabulary read speech often fail to scale to larger vocabulary spontaneous speech.”