Interspeech2019&サテライト論文読み会（音声生成系） #interspeech2019jp

by gaku kotani

Slide 1

Slide 1 text

l5IF645$4ZTUFNGPS#MJ[[BSE$IBMMFOHFz খ୩ַ  ౦ژେֶᜊ౻ݚ% *OUFSTQFFDIαςϥΠτ࿦จಡΈճɹ࿦จ঺հ

Slide 2

Slide 2 text

ࣗݾ঺հ ⾣খ୩ַʢͨ͜ʹ͕͘ʣ ɹὊ౦ژେֶᜊ౻େีݚڀࣨ%ʢֶ͔࢜Βʣ ɹὊ޷͖ͳ੠༏͸Ֆᖒ߳ࡊ ⾣ݚڀ෼໺ ɹὊ੠࣭ม׵ ɹὊ554΋͍ͨ͠ ɹɹՖᖒ߳ࡊͷ੠ʹͳΓ͍ͨͱ͍͏ΑΓɺՖᖒ߳ࡊͷ੠Λฉ͖͍ͨ 2

Slide 3

Slide 3 text

঺հ͢Δ࿦จ ⾣:+JBOHFUBM l5IF645$4ZTUFNGPS#MJ[[BSE$IBMMFOHF z  ɹJO#MJ[[BSE$IBMMFOHF8PSLTIPQ ɹ#MJ[[BSE$IBMMFOHF554ͷίϯϖʢࠓ೥ͷλεΫ͸.BOEBSJO554ʣ ɹ#$Ͱ࠷΋݁Ռ͕ྑ͔ͬͨνʔϜͷγεςϜઆ໌࿦จ ⾣બΜͩಈػ  ɹᜊ౻ݚ΋65PLZPνʔϜͱͯ͠ࢀՃͨ͠ͷͰҰ൪ͷγεςϜ͕ؾʹͳΔ ɹ554ͷษڧ͠Α͏ 3 ޡղ΍ݸਓͷҙݟΛؚΜͰ͍·͢ ਖ਼֬ͳ৘ใ͸࿦จΛࢀর͍ͯͩ͘͠͞

Slide 4

Slide 4 text

#MJ[[BSEDIBMMFOHF ⾣5BTL ɹὊ#VJMEBWPJDFGSPNUIFQSPWJEFEEBUB TVJUBCMFGPSFYQSFTTJWF554  ɹɹGSPNQMBJOUFYUJOQVUz ɹὊλʔήοτ͸தࠃޠΛ஻Δ͓͡͞ΜʢZPVUVCFSΈ͍ͨͳʜ ʣ ɹὊ࣌ؒ෼ͷσʔλʢ෼ºVUUFSTʣ ɹɹL)[TBNQMJOH NQ ɹɹσʔλ͕ͪΐͬͱ͋ΕɿಡΜͰͳ͍ʗॻ͍ͯͳ͍ʗςΩετԻ੠ϖΞ͕ؒҧͬͯΔ ⾣645$͸Ϛϯύϫʔͱ֎෦σʔλͰղܾʜ ɹὊ΋͸΍ຊਓͳԻ੠߹੒ʢαϯϓϧ୳ͯ͠΋ແ͔ͬͨͰ͢͢Έ·ͤΜʜʣ 4 ࠓఱੋ෕౔ࢫɼૹ۳େՈҰ࿽ྱ෺ɻզૌᬐ౨ԋჴဗཱ৽࿕؀ྃʬၕཕՈ඀ʭɼ  ࡞ູ৽ҰقతʬՈఉഎܠ੠ʭӁ඼ɼ䓟ՄҎࡏʮಘ౸ʯ"QQཬ໔ٮਵऔɻʢଓ͘ʣ

Slide 5

Slide 5 text

'SBNFXPSLPG645$TZTUFNGPS#$ ⾣NPEVMFT ɹὊ#&35CBTFEGSPOUFOEQSFEJDUJPO ɹὊ%VSBUJPONPEFMJOH ɹὊ("/CBTFENVMUJUBTLBDPVTUJDNPEFMJOH ɹὊ8BWF/FUCBTFEOFVSBMWPDPEFS 5 'MPXDIBSUPGUIFTZTUFN

Slide 6

Slide 6 text

'SBNFXPSLPG645$TZTUFNGPS#$ ⾣1SFQSPDFTT  ɹὊ.BOVBMBOOPUBUJPOT  ɹɹ1JOZJO QSPTPEJDXPSECPVOEBSZ   ɹɹQSPTPEJDQISBTFCPVOEBSZ GPDVTQPTJUJPO ɹὊ&YUSBDUJPOPGGSBNFMFWFMBDPVTUJDGFBUVSFT ɹɹ.FMDFQTUSVN ' 767 ɹɹ)..BMJHONFOUGPSQIPOFNFCPVOEBSJFT  6 ਓؒίετ͕͍͢͝ 'MPXDIBSUPGUIFTZTUFN

Slide 7

Slide 7 text

'SPOUFOENPEFM 7 ⾣5IFNBJOQSPDFEVSF ɹὊ5FYUQSPDFTTJOH ɹɹ4QFDJBMNBSLTDPOWFSTJPO ɹὊ(SBQIFNFUPQIPOFNFDPOWFSTJPO ɹɹ$IJOFTFDIBSBDUFSUPQJOZJO ɹὊ1SPTPEZQSFEJDUJPO ɹɹ#SFBLJOHBTFOUFODFJOUPQISBTFTGPDVTPOXPSET .VMUJUBTLNPEFMCBTFEPO#&35

Slide 8

Slide 8 text

'SPOUFOENPEFM 8 ⾣5IFNBJOQSPDFEVSF ɹὊ5FYUQSPDFTTJOH ɹɹ4QFDJBMNBSLTDPOWFSTJPO ɹὊ(SBQIFNFUPQIPOFNFDPOWFSTJPO ɹɹ$IJOFTFDIBSBDUFSUPQJOZJO ɹὊ1SPTPEZQSFEJDUJPO ɹɹ#SFBLJOHBTFOUFODFJOUPQISBTFTGPDVTPOXPSET JF"SBCJDOVNFSBMTUPWBMVFTPSTUSJOHT lzPSlUXP ٢ ٥ lzPSlpWFz

Slide 9

Slide 9 text

'SPOUFOENPEFM 9 ⾣5IFNBJOQSPDFEVSF ɹὊ5FYUQSPDFTTJOH ɹɹ4QFDJBMNBSLTDPOWFSTJPO ɹὊ(SBQIFNFUPQIPOFNFDPOWFSTJPO ɹɹ$IJOFTFDIBSBDUFSUPQJOZJO ɹὊ1SPTPEZQSFEJDUJPO ɹɹ#SFBLJOHBTFOUFODFJOUPQISBTFTGPDVTPOXPSET "TJOHMFDIBSIBWFTFWFSBMEJ⒎FSFOUQSPOVOT  BDDPSEJOHUPVTBHFTNFBOJOHTJOTFOUFODFT 单 lE⒑OzPSlTIÁOzPSlDIÂOz

Slide 10

Slide 10 text

'SPOUFOENPEFM 10 ⾣5IFNBJOQSPDFEVSF ɹὊ5FYUQSPDFTTJOH ɹɹ4QFDJBMNBSLTDPOWFSTJPO ɹὊ(SBQIFNFUPQIPOFNFDPOWFSTJPO ɹɹ$IJOFTFDIBSBDUFSUPQJOZJO ɹὊ1SPTPEZQSFEJDUJPO ɹɹ#SFBLJOHBTFOUFODFJOUPQISBTFTGPDVTPOXPSET 5IFTUSFOHUIPGUIFCSFBL  1SPTPEJDQISBTF - NBKPSCSFBL 1SPTPEJDXPSE - NJOPSCSFBL 'JHVSF&YBNQMFPGNBOEBSJOQSPTPEZTUSVDUVSF JO $IBSBDUFS -FYJDPOXPSE 1SPTPEJDXPSE 1SPTPEJDQISBTF *OUPOBUJPOBMQISBTF

Slide 11

Slide 11 text

'SPOUFOENPEFM 11 ⾣5IFNBJOQSPDFEVSF ɹὊ5FYUQSPDFTTJOH ɹɹ4QFDJBMNBSLTDPOWFSTJPO ɹὊ(SBQIFNFUPQIPOFNFDPOWFSTJPO ɹɹ$IJOFTFDIBSBDUFSUPQJOZJO ɹὊ1SPTPEZQSFEJDUJPO ɹɹ#SFBLJOHBTFOUFODFJOUPQISBTFTGPDVTPOXPSET 'PDVTXPSETFNQIBTJ[FEXPSET UPESBXBUUFOUJPOUPQBSUJDVMBSXPSET

Slide 12

Slide 12 text

'SPOUFOENPEFM ⾣#&35CBTFEGSPOUFOENPEFM ɹὊPVUQVUT ɹɹ1PMZ@PVUQJOZJOQSFEJDUJPO ɹɹ-@PVUNJOPSCSFBLPSOPU ɹɹ-@PVUNBKPSCSFBLPSOPU ɹɹ4QDM@PVUTQFDJBMNBSLTDPOWFSTJPO ɹɹ'PDVT@PVUGPDVTPSOPU ɹὊ$IJOFTFDPSQVTJOOPWFMBOEOFXTpFMETUPQSFUSBJO 12 #&35CBTFEGSPOUFOENPEFM ۩ମతͳྔ͸࿦จʹࡌ͍ͬͯͳ͍ˍൃදͰฉ͍ͨ͸͕ͣͩ๨Ε·ͨ͠ʜ

Slide 13

Slide 13 text

%VSBUJPONPEFMJOH ⾣"VUPSFHSFTTJWFNPEFMGPSRVBOUJpFEEVSBUJPOT  ɹὊ2VBOUJGZJOHJOMPHEPNBJOTPGUNBYDSPTTFOUSPQZ ɹὊ$BVTBMFODPEFSPVUQVUTUIFQSFWJPVTFNCFEEJOHT ɹɹEVSBUJPODPOUFYUDP[MJUUMFDPSSFMBUJPOXJUIEVSBUJPOT ɹὊ5IFDPOUFYUJOQVUJODMVEFT ɹɹQIPOFNFJEFOUJpDBUJPOT UPOFT   ɹɹ5P#* UPLFOTBOECSFBLJOEJDFT TUSFTTqBHT BOETPPO ɹὊ1SFUSBJOFEPOBMBSHFNVMUJTQFBLFSDPSQVT 13

Slide 14

Slide 14 text

Slide 15

Slide 15 text

%VSBUJPONPEFMJOH ⾣"VUPSFHSFTTJWFNPEFMGPSRVBOUJpFEEVSBUJPOT  ɹὊ2VBOUJGZJOHJOMPHEPNBJOTPGUNBYDSPTTFOUSPQZ ɹὊ$BVTBMFODPEFSPVUQVUTUIFQSFWJPVTFNCFEEJOHT ɹɹEVSBUJPODPOUFYUDP[MJUUMFDPSSFMBUJPOJOEVSBUJPOWBMVFT ɹὊ5IFDPOUFYUJOQVUJODMVEFT ɹɹQIPOFNFJEFOUJpDBUJPOT UPOFT   ɹɹ5P#* UPLFOTBOECSFBLJOEJDFT TUSFTTqBHT BOETPPO ɹὊ1SFUSBJOFEPOBMBSHFNVMUJTQFBLFSDPSQVT 15

Slide 16

Slide 16 text

Slide 17

Slide 17 text

Slide 18

Slide 18 text

"DPVTUJDNPEFMJOH 18 ⾣-45.3//BDPVTUJDNPEFMXJUI("/ ɹὊ5IFPWFSTNPPUIJOHQSPCMFN("/ ɹɹ%JTDSJNJOBUJOHJOTQFDUSBMEPNBJO POMZ41& ɹɹ-FBTUTRVBSF("/ -4("/ ɹὊ1SFEJDUFEGFBUVSFT' EJN.$&1 #"1

Slide 19

Slide 19 text

7PDPEJOH ⾣8BWF/FUCBTFEOFVSBMWPDPEJOH ɹὊ1SFUSBJOFEXJUIBNVMUJTQFBLFSEBUBTFU ɹɹ4QFBLFSFNCFEEJOHWFDUPSTBSFVTFEEVSJOHPOMZQSFUSBJOJOH ɹὊ$PNCJOBUJPOPG8BWF/FUT L)[BOEL)[WFST ɹɹL)[8BWF/FUHFOFSBUFTCFUUFSMPXGSFRVFODZIBSNPOJD ɹɹ$PODBUFOBUJOHUIFJSBNQMJUVEFT ɹὊ"TJOHMFWBSJBODFCPVOEFE(BVTTJBOEJTUSJCVUJPONPEFMJOHSBXXBWFGPSN ɹɹJOUSPEVDFEJO$MBSJ/FU<1JOH > ɹɹSFMJFWFTUIFRVBOUJ[BUJPOOPJTF 19

Slide 20

Slide 20 text

7PDPEJOH ⾣8BWF/FUCBTFEOFVSBMWPDPEJOH ɹὊ$PNCJOBUJPOPG8BWF/FUT L)[BOEL)[WFST ɹɹL)[8BWF/FUHFOFSBUFTCFUUFSMPXGSFRVFODZIBSNPOJD ɹɹ$PODBUFOBUJOHUIFJSBNQMJUVEFT 20 ࿦จʹઆ໌͕΄ͱΜͲͳ͍͕ɺߴଳҬͷҐ૬͸ϊΠζͰຒΊͯ  (SJ⒏O-JNʢతͳ͜ͱʣΛͨ͠ͱฉ͍ͨؾ͕͢Δʢ͏Ζ֮͑ʜʣ

Slide 21

Slide 21 text

3FTVMUTPGFWBMVBUJPO 21 722 722 722 722 722 722 722 722 722 722 722 722 722 722 722 722 722 722 722 722 722 722 722 722 722 722 n A M I W S Y Z E C G T H D V J N X K Q F O R B U L P 1 2 3 4 5 Mean Opinion Scores (All Listeners) System Score 361 361 361 361 361 361 361 361 361 361 361 361 361 361 361 361 361 361 361 361 361 361 361 361 361 361 n A M I W S Y Z E C G T H D V J N X K Q F O R B U L P 1 2 3 4 5 Similarity Scores Comparing to Original Speaker (All Listeners) System Score /BUVSBMOFTTTDPSFTPGTVCNJUUFETZTUFN 4JNJMBSJUZTDPSFTPGTVCNJUUFETZTUFN

Slide 22

Slide 22 text

4VNNBSZ ⾣645$TZTUFNGPS#MJ[[BSE$IBMMFOHF ɹὊ#&35CBTFEGSPOUFOEQSFEJDUJPO ɹὊ"VUPSFHSFTTJWFEVSBUJPONPEFM ɹὊ("/CBTFENVMUJUBTLBDPVTUJDNPEFM ɹὊ8BWF/FUCBTFEOFVSBMWPDPEFS ⾣5IFWFSZIJHI.04TDPSFJTBDIJFWFEJOIJTUPSZPG#MJ[[BSE$IBMMFOHF ⾣ײ૝ ɹὊʢԻ੠αϯϓϧ͸༻ҙͰ͖ͳ͔͕ͬͨɺݱ஍Ͱฉ͍ͨײ͡ʣ·ΔͰຊਓ͔ͷΑ͏ͳ  ɹɹԻ੠͕ੜ੒Ͱ͖͍ͯͯײಈ͕͋ͬͨ 22