{"id":209770,"metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00209770","sets":["1164:5159:10515:10530"]},"path":["10530"],"owner":"44499","recid":"209770","title":["Speech Enhancement in the Presence of Background Music Considering Speech and Music Characteristics"],"pubdate":{"attribute_name":"公開日","attribute_value":"2021-02-24"},"_buckets":{"deposit":"50ca3793-21ac-43f8-8f26-2051181edee4"},"_deposit":{"id":"209770","pid":{"type":"depid","value":"209770","revision_id":0},"owners":[44499],"status":"published","created_by":44499},"item_title":"Speech Enhancement in the Presence of Background Music Considering Speech and Music Characteristics","author_link":["529698","529699","529701","529702","529704","529705","529700","529703"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"Speech Enhancement in the Presence of Background Music Considering Speech and Music Characteristics"},{"subitem_title":"Speech Enhancement in the Presence of Background Music Considering Speech and Music Characteristics","subitem_title_language":"en"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"SLP1","subitem_subject_scheme":"Other"}]},"item_type_id":"4","publish_date":"2021-02-24","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"Graduate School of Informatics, Kyoto University"},{"subitem_text_value":"Graduate School of Informatics, Kyoto University"},{"subitem_text_value":"Graduate School of Informatics, Kyoto University"},{"subitem_text_value":"Graduate School of Informatics, Kyoto University"}]},"item_4_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"Graduate School of Informatics, Kyoto University","subitem_text_language":"en"},{"subitem_text_value":"Graduate School of Informatics, Kyoto University","subitem_text_language":"en"},{"subitem_text_value":"Graduate School of Informatics, Kyoto University","subitem_text_language":"en"},{"subitem_text_value":"Graduate School of Informatics, Kyoto University","subitem_text_language":"en"}]},"item_language":{"attribute_name":"言語","attribute_value_mlt":[{"subitem_language":"eng"}]},"item_publisher":{"attribute_name":"出版者","attribute_value_mlt":[{"subitem_publisher":"情報処理学会","subitem_publisher_language":"ja"}]},"publish_status":"0","weko_shared_id":-1,"item_file_price":{"attribute_name":"Billing file","attribute_type":"file","attribute_value_mlt":[{"url":{"url":"https://ipsj.ixsq.nii.ac.jp/record/209770/files/IPSJ-SLP21136032.pdf","label":"IPSJ-SLP21136032.pdf"},"date":[{"dateType":"Available","dateValue":"2023-02-24"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-SLP21136032.pdf","filesize":[{"value":"14.6 MB"}],"mimetype":"application/pdf","priceinfo":[{"tax":["include_tax"],"price":"660","billingrole":"5"},{"tax":["include_tax"],"price":"330","billingrole":"6"},{"tax":["include_tax"],"price":"0","billingrole":"22"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_date","version_id":"ff5ef0ee-36b9-42d0-b280-5b1277861b41","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2021 by the Information Processing Society of Japan"}]},"item_4_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Jeongwoo, Woo"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Masato, Mimura"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Kazuyoshi, Yoshii"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Tatsuya, Kawahara"}],"nameIdentifiers":[{}]}]},"item_4_creator_6":{"attribute_name":"著者名(英)","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Jeongwoo, Woo","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Masato, Mimura","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Kazuyoshi, Yoshii","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Tatsuya, Kawahara","creatorNameLang":"en"}],"nameIdentifiers":[{}]}]},"item_4_source_id_9":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AN10442647","subitem_source_identifier_type":"NCID"}]},"item_4_textarea_12":{"attribute_name":"Notice","attribute_value_mlt":[{"subitem_textarea_value":"SIG Technical Reports are nonrefereed and hence may later appear in any journals, conferences, symposia, etc."}]},"item_resource_type":{"attribute_name":"資源タイプ","attribute_value_mlt":[{"resourceuri":"http://purl.org/coar/resource_type/c_18gh","resourcetype":"technical report"}]},"item_4_source_id_11":{"attribute_name":"ISSN","attribute_value_mlt":[{"subitem_source_identifier":"2188-8663","subitem_source_identifier_type":"ISSN"}]},"item_4_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"Speech enhancement in the presence of background music is not so different from noise reduction if music is treated as just noise. However, music has definite characteristics which are made by human beings, unlike noise which can be any. In order to consider characteristics of background music instead of noise reduction, we introduce a generative adversarial network (GAN). We combine two multi-scale discriminators for speech and music with Conv-TasNet modified for speech enhancement. We train it jointly with SI-SDR and the GAN objective. Experimental evaluations through speech recognition demonstrate that the proposed model is improved from the baseline model. It is notable that the more music interference is large, the more the proposed method is effective. Comparing the spectrogram of enhanced speech by the proposed and baseline model demonstrate that the baseline model tends to cut off noise excessively, in contrast the proposed model reconstructs more faithfully.","subitem_description_type":"Other"}]},"item_4_description_8":{"attribute_name":"論文抄録(英)","attribute_value_mlt":[{"subitem_description":"Speech enhancement in the presence of background music is not so different from noise reduction if music is treated as just noise. However, music has definite characteristics which are made by human beings, unlike noise which can be any. In order to consider characteristics of background music instead of noise reduction, we introduce a generative adversarial network (GAN). We combine two multi-scale discriminators for speech and music with Conv-TasNet modified for speech enhancement. We train it jointly with SI-SDR and the GAN objective. Experimental evaluations through speech recognition demonstrate that the proposed model is improved from the baseline model. It is notable that the more music interference is large, the more the proposed method is effective. Comparing the spectrogram of enhanced speech by the proposed and baseline model demonstrate that the baseline model tends to cut off noise excessively, in contrast the proposed model reconstructs more faithfully.","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"5","bibliographic_titles":[{"bibliographic_title":"研究報告音声言語情報処理（SLP）"}],"bibliographicPageStart":"1","bibliographicIssueDates":{"bibliographicIssueDate":"2021-02-24","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"32","bibliographicVolumeNumber":"2021-SLP-136"}]},"relation_version_is_last":true,"weko_creator_id":"44499"},"updated":"2025-01-19T18:23:48.105623+00:00","created":"2025-01-19T01:11:03.493744+00:00","links":{}}