FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W86620541> ?p ?o ?g. }

• W86620541 abstract "In this paper we address the two speaker segregation problem in a single channel paradigm using sinusoidal residual modeling. An appropriate selection of the number of sine waves, window length and hysteresis threshold, is done so as to model and synthesize the underlying signal corresponding to the speaker with the lower pitch period, using an amplitude only sine wave synthesis. The sinusoidal residual is then computed after restimating the phases with known amplitudes, by minimizing a criterion function. This residual corresponds to the the speaker with the higher pitch period. But such a residual consists of harmonic components of the speaker with the lower pitch period. We therefore estimate a binary mask from the spectrograms of the synthesized signal and the residual using a min-max technique to further improve the quality of the segregated speech. This segregation technique is then integrated into a co-channel speaker identification system, at various target to interference ratios. Reasonable improvements in identification performance are noted from these experiments. I. INTRODUCTION Recovering individual speech signals from a combination of two or more sources is a becoming a central problem in speech processing. Several approaches (1), have been tried to solve this problem ranging from the use of spatial information (2), to incorporating visual information (3), along with speech. But single channel speech segregation without the prior knowledge of the speech sources is challenging. In this paper we attempt to segregate the individual speakers from a mixture of two speakers collected over a single microphone. In Section II and III, we describe the sinusoidal residual modeling technique (4), and a formulation of the two speaker segregation problem respectively. The additive bank of sine wave synthesis of the estimated amplitudes and frequencies using this model results in a signal that corresponds to the source with the lower pitch period. But it contains some background information corresponding to the second speaker with the higher pitch period. The sinusoidal residual computed using a synthesis after restimating the phases results in the source with the higher pitch period. We attempt to illustrate and justify the reasons for the sinusoidal residual to contain information of the source with a higher pitch period in Section III-A. From the synthesized signal and the residual, we derive a mask using the min-max technique. This mask is applied on the synthesized signal to further refine the quality of the segregated sources. The computation of the mask and subsequent results are discussed in Section IV-A. This method is then integrated into a co-channel speaker identification system in Section V. The limitations of the technique and conclusions are discussed in in Section VI. II. SINUSOIDAL MODELING Sinusoidal modeling is based on the model suggested by Quatieri and McAulay (4), where a speech signal can be represented by a sum of amplitude-frequency modulated sine waves. The speech signal x(n) can be expressed as a sum of time varying frequencies, amplitudes and phases as" @default.
• W86620541 created "2016-06-24" @default.
• W86620541 creator A5023986320 @default.
• W86620541 creator A5085503354 @default.
• W86620541 date "2009-01-01" @default.
• W86620541 modified "2023-09-26" @default.
• W86620541 title "Single Channel Speaker Segregation using Sinusoidal Residual Modeling" @default.
• W86620541 cites W1560013842 @default.
• W86620541 cites W2015143272 @default.
• W86620541 cites W2070230103 @default.
• W86620541 cites W2099128937 @default.
• W86620541 cites W2108384452 @default.
• W86620541 cites W2119599673 @default.
• W86620541 cites W2164764235 @default.
• W86620541 cites W2170491071 @default.
• W86620541 cites W2619993508 @default.
• W86620541 cites W3127686677 @default.
• W86620541 hasPublicationYear "2009" @default.
• W86620541 type Work @default.
• W86620541 sameAs 86620541 @default.
• W86620541 citedByCount "0" @default.
• W86620541 crossrefType "journal-article" @default.
• W86620541 hasAuthorship W86620541A5023986320 @default.
• W86620541 hasAuthorship W86620541A5085503354 @default.
• W86620541 hasConcept C11413529 @default.
• W86620541 hasConcept C121332964 @default.
• W86620541 hasConcept C127162648 @default.
• W86620541 hasConcept C155512373 @default.
• W86620541 hasConcept C157138929 @default.
• W86620541 hasConcept C165801399 @default.
• W86620541 hasConcept C180205008 @default.
• W86620541 hasConcept C199360897 @default.
• W86620541 hasConcept C24890656 @default.
• W86620541 hasConcept C2778263558 @default.
• W86620541 hasConcept C2779843651 @default.
• W86620541 hasConcept C28490314 @default.
• W86620541 hasConcept C32022120 @default.
• W86620541 hasConcept C41008148 @default.
• W86620541 hasConcept C45273575 @default.
• W86620541 hasConcept C61328038 @default.
• W86620541 hasConcept C62520636 @default.
• W86620541 hasConcept C66907618 @default.
• W86620541 hasConcept C76155785 @default.
• W86620541 hasConceptScore W86620541C11413529 @default.
• W86620541 hasConceptScore W86620541C121332964 @default.
• W86620541 hasConceptScore W86620541C127162648 @default.
• W86620541 hasConceptScore W86620541C155512373 @default.
• W86620541 hasConceptScore W86620541C157138929 @default.
• W86620541 hasConceptScore W86620541C165801399 @default.
• W86620541 hasConceptScore W86620541C180205008 @default.
• W86620541 hasConceptScore W86620541C199360897 @default.
• W86620541 hasConceptScore W86620541C24890656 @default.
• W86620541 hasConceptScore W86620541C2778263558 @default.
• W86620541 hasConceptScore W86620541C2779843651 @default.
• W86620541 hasConceptScore W86620541C28490314 @default.
• W86620541 hasConceptScore W86620541C32022120 @default.
• W86620541 hasConceptScore W86620541C41008148 @default.
• W86620541 hasConceptScore W86620541C45273575 @default.
• W86620541 hasConceptScore W86620541C61328038 @default.
• W86620541 hasConceptScore W86620541C62520636 @default.
• W86620541 hasConceptScore W86620541C66907618 @default.
• W86620541 hasConceptScore W86620541C76155785 @default.
• W86620541 hasLocation W866205411 @default.
• W86620541 hasOpenAccess W86620541 @default.
• W86620541 hasPrimaryLocation W866205411 @default.
• W86620541 hasRelatedWork W165783309 @default.
• W86620541 hasRelatedWork W1826745750 @default.
• W86620541 hasRelatedWork W1925561251 @default.
• W86620541 hasRelatedWork W1986225192 @default.
• W86620541 hasRelatedWork W1989364713 @default.
• W86620541 hasRelatedWork W2032074738 @default.
• W86620541 hasRelatedWork W2078135751 @default.
• W86620541 hasRelatedWork W2100443089 @default.
• W86620541 hasRelatedWork W2470502990 @default.
• W86620541 hasRelatedWork W2955753320 @default.
• W86620541 hasRelatedWork W8706387 @default.
• W86620541 hasRelatedWork W2117790975 @default.
• W86620541 hasRelatedWork W2240798263 @default.
• W86620541 hasRelatedWork W2266658284 @default.
• W86620541 hasRelatedWork W2400346490 @default.
• W86620541 hasRelatedWork W2727341167 @default.
• W86620541 hasRelatedWork W2761247573 @default.
• W86620541 hasRelatedWork W2841934980 @default.
• W86620541 hasRelatedWork W2938303226 @default.
• W86620541 hasRelatedWork W2945717710 @default.
• W86620541 isParatext "false" @default.
• W86620541 isRetracted "false" @default.
• W86620541 magId "86620541" @default.
• W86620541 workType "article" @default.