Upgrade to Pro — share decks privately, control downloads, hide ads and more …

メディアが運用すべき持続可能なVTuberをつくる技術

Hirosaji
September 17, 2019

 メディアが運用すべき持続可能なVTuberをつくる技術

PyCon JP 2019 (2019/09/17) @Hirosaji
https://pycon.jp/2019/

Hirosaji

September 17, 2019
Tweet

More Decks by Hirosaji

Other Decks in Programming

Transcript

  1. 1Z$PO+1

    ϝσΟΞ͕ӡ༻͢΂͖
    ࣋ଓՄೳͳ75VCFSΛͭ͘Δٕज़
    ೔ຊܦࡁ৽ฉࣾ
    Ճ౻ᚠ໵ !IJSPTBKJ

    View full-size slide

  2. ͜ͷϓϨθϯͷ໨త
    ͜ͷϓϨθϯͰ͸ɺօ͞Μͷண૝ͷܹࢗʹͳΔΑ͏ͳ࣮૷ྫΛ঺հ͠·͢ɻ
    75VCFSΛ୊ࡐͱͨ͠։ൃͷண૝Λಘͯ΋Β͍ɺ
    ͱࢥͬͯ΋Β͏͜ͱɻ
    ʮฐࣾͰ΋75VCFSΛಋೖͯ͠ΈΑ͏͔ͳʯ
    ௌߨલ ௌߨத ௌߨޙ

    View full-size slide

  3. ͜ͷϓϨθϯΛൃද͢Δਓ
    ࢓ࣄͰ͸ͳ͘ɺझຯͰऔΓ૊Ή75VCFS։ൃऀͷ਎ͱ͓ͯ͠࿩͠͠·͢ɻ
    ɾ)JSPZB,BUP !)JSPTBKJ

    ɾ৽ଔ̎೥໨ͷ8FCΤϯδχΞ
    ɾ75VCFS։ൃ͸झຯ
    ɾ೔ܦ7JTVBM%BUBͷ։ൃ୲౰
    ɾϑϩϯτʗόοΫʗ%FW0QTʗݚڀ։ൃͳͲ
    Ұॠ͚ͩ
    $.ʹग़·ͨ͠

    View full-size slide

  4. ͳͥϝσΟΞ͕75VCFSΛʁ
    75VCFSӡ༻Ͱى͜Δ՝୊ΛղܾͰ͖ΔϦιʔε
    ͕͋Δ͔Β
    ɾಈը੍࡞ίετ
    ɾதͷਓͷଐਓੑ
    ՝୊
    "35*$-&4
    ʢςΩετσʔλʣ
    +063/"-*454
    ʢ৭ʑͳઐ໳Ոʣ
    Ϧιʔε
    ղܾ
    ʮϝσΟΞ͕75VCFSʁʯͱࢥ͏͔΋͠Ε·ͤΜ͕ɺҙ֎ͱ༗ޮͳΜͰ͢ɻ

    View full-size slide

  5. ࠓճ঺հ͢Δ75VCFS૾
    ࠓճ͸ɺ্هͷ̎ͭͷ75VCFS૾Λ঺հ͠·͢ɻ
    ࣗಈੜ੒ܕ75VCFS

    ଟॏਓ֨ܕ75VCFS

    View full-size slide

  6. ࠓճ঺հ͢Δ75VCFS૾
    ࣌ؒͷ౎߹্ɺ۩ମతͳٕज़ղઆΛ͢Δͷ͸̎ͷ75VCFS͚ͩͰ͢ɻ
    ٕज़ղઆ͸
    ͚ͬͪͩ͜
    ࣗಈੜ੒ܕ75VCFS

    ଟॏਓ֨ܕ75VCFS

    View full-size slide

  7. ࣗಈੜ੒ܕ75VCFS
    ̍

    View full-size slide

  8. ࣗಈੜ੒ܕ75VCFSͱ͸
    ςΩετσʔλΛ
    ༩͑Δ͚ͩͰࣗಈੜ੒Ͱ͖Δ75VCFSɻ
    */165 065165
    $0.165&3 7*%&0
    "35*$-&4
    ʢςΩετσʔλʣ
    ·ͣ͸ɺಈը੍࡞ίετֻ͕͔Βͳ͍ࣗಈੜ੒ܕ75VCFSΛ঺հ͠·͢ɻ

    View full-size slide

  9. Ͳ͏΍ͬͯࣗಈੜ੒͢Δͷʁ
    5FYUUP(FTUVSF 55(

    ίʔύεͰɺϞσϧͷϞʔγϣϯΛ
    ؅ཧ͢Δ࣌ܥྻσʔλΛ࡞Δɻ
    5FYUUP4QFFDI 554

    Ի੠߹੒ϞσϧͰςΩετσʔλΛ
    Ի੠σʔλʹม׵͢Δɻ

    7*%&0
    /-1GPS5FYU1SFQSPDFTTJOH
    ੜͷςΩετσʔλΛม׵ɾΫϨϯδϯάɾߏ଄Խ͢Δɻ
    ·ͨɺඞཁʹԠͯ͡ɺޙͷ55(ʹ࢖͏ίʔύε΋࡞੒͢Δɻ
    "35*$-&4
    ʢςΩετσʔλʣ
    ( )
    ࣗવݴޠ΍Ի੠߹੒ͳͲɺػցֶशͰࣗಈੜ੒γεςϜΛߏங͠·͢ɻ

    View full-size slide

  10. ࣗಈੜ੒ܕ75VCFSɿ۩ମྫᶃ
    ࣾ಺ࣄྫὊ"*Ωϟελʔʮ޷࣮ΤϦΧʯ
    ೔ܦͷࣗಈੜ੒γεςϜͰ࡞੒ͨ͠75VCFSಈը͕ͪ͜Βɻ

    View full-size slide

  11. ࣾ಺ࣄྫὊ"*Ωϟελʔʮ޷࣮ΤϦΧʯ
    ࣗಈੜ੒ܕ75VCFSɿ۩ମྫᶃ
    /-1GPS5FYU1SFQSPDFTTJOH
    ɾॻ͖ݴ༿Λ࿩͠ݴ༿ʹม׵
    ɾ௕͍هࣄΛࣈఔ౓ʹཁ໿
    ɾԻಡΈɺ܇ಡΈͳͲͷಡΈํʹ΋ରԠ
    ޱޠม׵΍จষཁ໿ͳͲɺςΩετͷՃ޻ʹಛʹྗ͕ೖ͍ͬͯ·͢ɻ

    View full-size slide

  12. ࣗಈੜ੒ܕ75VCFSɿ۩ମྫᶄ
    ·ͨɺͱʹ͔͘αΫοͱ࣮૷͍ͨ͠ํ͸4VNFSJBO͕Φεεϝɻ
    ࠷΋ָͳ࣮૷ྫὊ "NB[PO4VNFSJBOΛ࢖͏

    View full-size slide

  13. ࠷΋ָͳ࣮૷ྫὊ "NB[PO4VNFSJBOΛ࢖͏
    ˜65+6$-
    ࣗಈੜ੒ܕ75VCFSɿ۩ମྫᶄ
    5FYUUP(FTUVSF 55(

    ɾ(FTUVSF.BQͰϞʔγϣϯ؅ཧ
    ɾ৚݅෇͖ͰϞσϧ΋௥ՃՄೳ
    ɾΧελϜϞʔγϣϯΛ௥ՃՄೳ
    ʢ.#ҎԼͷGCY PCKʹݶΔʣ
    5FYUUP4QFFDI 554

    ɾ"NB[PO1PMMZ͕ར༻Մೳ
    ɾԻ੠߹੒Ϟσϧͷ௥Ճ͸ෆՄ
    ɾӳޠ͸΋ͪΖΜ೔ຊޠʹ΋ରԠ
    ʢ೥݄ݱࡏʣ
    ೖ໳͢Δ͚ͩͳΒ෼͘Β͍ͰɺϊϯίʔσΟϯάͰग़དྷ·͢ɻ

    View full-size slide

  14. ଟॏਓ֨ܕ75VCFS
    ̎

    View full-size slide

  15. +063/"-*454
    ʢ৭ʑͳઐ໳Ոʣ
    */165 065165
    $0.165&3 7*%&0
    ଟॏਓ֨ܕ75VCFSͱ͸
    ͲΜͳzதͷਓzͰ΋
    ಉҰͷΩϟϥΫλʔͱͯ͠ৼΔ෣͑Δ75VCFSɻ
    ࣍ʹɺzதͷਓzͷଐਓੑʹͱΒΘΕͳ͍ଟॏਓ֨ܕ75VCFSΛ঺հ͠·͢ɻ

    View full-size slide

  16. +063/"-*4
    ʢ৭ʑͳઐ໳Ոʣ
    Ͳ͏΍ͬͯಉҰԽ͢Δͷʁ

    7*%&0
    .PUJPO$BQUVSF
    zதͷਓzͷද৘΍ମͷಈ͖ΛϞσϧʹಉظͤ͞Δɻ
    ΩϟϥΫλʔͱͯ͠ͷݟͨ໨Λදݱ͢ΔͨΊʹඞཁɻ
    7PJDF$POWFSTJPO
    zதͷਓzͷ੠ΛಛఆͷԻ੠ʹม׵͢Δɻ
    ΩϟϥΫλʔͱͯ͠ͷ੠Λදݱ͢ΔͨΊʹඞཁɻ
    ϞʔγϣϯΩϟϓνϟͱ੠࣭ม׵Ͱɺݟͨ໨ͱ੠Λग़ྗ͠ɺ߹੒͠·͢ɻ

    View full-size slide

  17. +063/"-*4
    ʢ৭ʑͳઐ໳Ոʣ
    Ͳ͏΍ͬͯಉҰԽ͢Δͷʁ

    7*%&0
    .PUJPO$BQUVSF
    zதͷਓzͷද৘΍ମͷಈ͖ΛϞσϧʹಉظͤ͞Δɻ
    ΩϟϥΫλʔͱͯ͠ͷݟͨ໨Λදݱ͢ΔͨΊʹඞཁɻ
    7PJDF$POWFSTJPO
    zதͷਓzͷ੠ΛಛఆͷԻ੠ʹม׵͢Δɻ
    ΩϟϥΫλʔͱͯ͠ͷ੠Λදݱ͢ΔͨΊʹඞཁɻ
    ࠓճ͸੠࣭ม׵ʹয఺Λ͋ͯͯɺͦͷجૅ͔Βݚڀಈ޲·Ͱղઆ͠·͢ɻ

    View full-size slide

  18. ੠࣭ม׵ͷجૅ
    7PJDF$POWFSTJPOʢ੠࣭ม׵ʣͷ֓ཁ
    "OBMZTJT
    ෼ੳ
    $POWFSTJPO
    ม׵
    4ZOUIFTJT
    ߹੒
    ೖྗͨ͠੠Λɺಛఆͷ੠ʹม׵͢Δٕज़ɻ
    ෼ੳɾม׵ɾ߹੒ͷ̏ͭͷஈ֊Ͱม׵͢Δɻ
    ੠࣭ม׵Ͱ͸جຊɺ̏ͭͷஈ֊ʹ෼͚ͯԻ੠Λม׵͠·͢ɻʢ&&΋੝Μʣ

    View full-size slide

  19. ੠࣭ม׵ʹΑ͘ར༻͞ΕΔʮԻڹಛ௃ྔʯ
    "OBMZTJT
    ෼ੳ
    $POWFSTJPO
    ม׵
    4ZOUIFTJT
    ߹੒

    جຊप೾਺'
    ϝϧέϓετϥϜ
    ඇपظੑࢦඪ

    جຊप೾਺'
    ϝϧέϓετϥϜ
    ඇपظੑࢦඪ
    Ի੠͔ΒʮԻڹಛ௃ྔʯʹநग़ͯ͠ѻ͏ɻ
    Իͷߴ͞΍ɺ੠ͷ͔͢Ε۩߹ͱ͍ͬͨࢦඪΛ࢖͏ɻ
    ʢجຊप೾਺'


    ʢඇपظੑࢦඪ

    Ի੠͔ΒԻڹಛ௃ྔΛநग़͠ɺม׵΍ม׵Ϟσϧͷߏஙʹར༻͠·͢ɻ
    ੠࣭ม׵ͷجૅ

    View full-size slide

  20. "OBMZTJT
    ෼ੳ
    $POWFSTJPO
    ม׵
    4ZOUIFTJT
    ߹੒

    جຊप೾਺'
    ϝϧέϓετϥϜ
    ඇपظੑࢦඪ

    جຊप೾਺'
    ϝϧέϓετϥϜ
    ඇपظੑࢦඪ
    ಛʹಛ௃తͳͷ͕ϝϧέϓετϥϜ .$&1
    ɻ
    ੠ͷߏ଄Λදݱͨ͠ɺॏཁͳԻڹಛ௃ྔͰ͋Δɻ
    தͰ΋.$&1͸ಛʹॏཁɻػցֶशͷֶशର৅ͱͳΔͷ͸ɺ͍ͭ΋.$&1ɻ
    ʮϝϧέϓετϥϜʯ͸ಛʹॏཁ
    ੠࣭ม׵ͷجૅ

    View full-size slide

  21. "OBMZTJT
    ෼ੳ
    $POWFSTJPO
    ม׵
    4ZOUIFTJT
    ߹੒

    جຊप೾਺'
    ϝϧέϓετϥϜ
    ඇपظੑࢦඪ

    جຊप೾਺'
    ϝϧέϓετϥϜ
    ඇपظੑࢦඪ
    ϝϧέϓετϥϜ
    .$&1

    ͱ͸ɺ
    ϝϧई౓Ͱεέʔϧͨ͠έϓετϥϜͷ͜ͱɻ
    ʢˢߴ͍Իͷҧ͍ʹಷײͳɺਓͷௌ֮Λߟྀͨ͠ई౓ʣ
    㱠ϝϧप೾਺έϓετϥϜ܎਺ʢ.'$$ʣ
    ʢ ʣ
    ʮϝϧέϓετϥϜʯͱ͸ʁ
    .$&1͸ɺϝϧई౓Ͱਓؒͷௌ֮Λߟྀͨ͠ʮέϓετϥϜʯͰ͢ɻ
    ੠࣭ม׵ͷجૅ

    View full-size slide

  22. ʮέϓετϥϜʯͱ͸ʁ
    प೾਺<)[>
    ର਺ৼ෯
    έϓετϥϜ܎਺ʢ࣍ݩ਺ʣ
    ର਺ৼ෯
    ࣌ؒ
    ৼ෯
    ର਺ৼ෯
    प೾਺<)[>
    ର਺ৼ෯
    प೾਺<)[>
    έϓετϥϜ
    ର਺ύϫʔεϖΫτϧ
    Ի੠೾ܗ
    εϖΫτϧแབྷ εϖΫτϧඍࡉߏ଄
    ௿߲࣍Λ''5 ߴ߲࣍Λ''5
    J''5
    ''5
    ί
    Ϩ
    έϓετϥϜΛ࢖͑͹ɺ௿࣍ݩͰʮεϖΫτϧแབྷʗඍࡉߏ଄ʯΛදݱՄೳɻ
    ੠࣭ม׵ͷجૅ

    View full-size slide

  23. ʮέϓετϥϜʯͱ͸ʁ
    प೾਺<)[>
    ର਺ৼ෯
    έϓετϥϜ܎਺ʢ࣍ݩ਺ʣ
    ର਺ৼ෯
    ࣌ؒ
    ৼ෯
    ର਺ৼ෯
    प೾਺<)[>
    ର਺ৼ෯
    प೾਺<)[>
    έϓετϥϜ
    ର਺ύϫʔεϖΫτϧ
    Ի੠೾ܗ
    εϖΫτϧแབྷ εϖΫτϧඍࡉߏ଄
    ௿߲࣍Λ''5 ߴ߲࣍Λ''5
    J''5
    ''5
    ί
    Ϩ
    έϓετϥϜΛ࢖͑͹ɺ௿࣍ݩͰʮεϖΫτϧแབྷʗඍࡉߏ଄ʯΛදݱՄೳɻ
    ੠࣭ม׵ͷجૅ
    ௿࣍ݩʢগͳ͍৘ใʣͰ
    දݱՄೳɻ

    View full-size slide

  24. ʮεϖΫτϧแབྷʗඍࡉߏ଄ʯͱ͸ʁ
    ੠ಓ੒෼Λੜ੒
    ޱ΍৶ͷܗɺઉͷҐஔʹΑͬͯ
    ܾ·ΔԻͷ৘ใɻ
    ʢԻૉͳͲʣ
    ഏ͔ΒૹΔۭؾͷྔ΍ɺ޴ͷకΊ
    ۩߹ʹΑܾͬͯ·ΔԻͷ৘ใɻ
    ʢԻͷେ͖͞΍ߴ͞ͳͲʣ
    ੠ଳ੒෼Λੜ੒
    ෼ੳ
    εϖΫτϧแབྷ
    εϖΫτϧඍࡉߏ଄
    εϖΫτϧแབྷʗඍࡉߏ଄͸ɺ੠ಓ΍੠ଳͰ࡞ΒΕΔԻ੠ͷ੒෼ͷ͜ͱɻ
    ੠࣭ม׵ͷجૅ

    View full-size slide

  25. ͭ·ΓɺʮϝϧέϓετϥϜʯͱ͸ʁ
    ੠࣭ม׵ͷجૅ
    ਓؒͷௌ֮Λߟྀͨ͠ɺ
    ੠ಓ੒෼ʮεϖΫτϧแབྷʯ΍
    ੠ଳ੒෼ʮεϖΫτϧඍࡉߏ଄ʯΛ
    ௿࣍ݩʹදݱͰ͖Δ
    Իڹಛ௃ྔɻ
    ੠࣭ม׵Ͱ͸͜ͷ.$&1ͷɺಛʹεϖΫτϧแབྷͷ৘ใ͕ॏཁͱͳΓ·͢ɻ

    View full-size slide

  26. Ͱ͸ɺ͜͜·Ͱͷ࿩Λ࣮૷ͯ͠Έ·͢ɻ
    ੠࣭ม׵ͷجૅ
    Ի੠ͷ෼ੳʗ߹੒ͷ࣮૷ʢʣ
    ඞཁͳϥΠϒϥϦͷ
    ಡΈࠐΈ
    Ի੠σʔλಡΈࠐΈ
    ڞ௨ͷύϥϝʔλΛ
    ઃఆ

    View full-size slide

  27. ੠࣭ม׵ͷجૅ
    Ի੠ͷ෼ੳʗ߹੒ͷ࣮૷ʢʣ
    803-%ɺ415,Ͱ
    Իڹಛ௃ྔΛநग़
    εϖΫτϧแབྷΛ
    .$&1͔Βࢉग़
    Ի੠೾ܗΛ߹੒
    ԻྔΛௐ੔
    ͜ͷαϯϓϧίʔυ͸ɺ੠࣭ม׵ϞσϧΛؒʹڬΉ͜ͱΛ૝ఆ͍ͯ͠·͢ɻ
    Ի੠σʔλॻ͖ग़͠

    View full-size slide

  28. ੠࣭ม׵ͷجૅ
    Ի੠ͷ෼ੳʗ߹੒ͷ࣮૷ʢʣ
    803-%ɺ415,Ͱ
    Իڹಛ௃ྔΛநग़
    εϖΫτϧแབྷΛ
    .$&1͔Βࢉग़
    Ի੠೾ܗΛ߹੒
    ԻྔΛௐ੔
    Ի੠σʔλॻ͖ग़͠
    ೖྗʗग़ྗԻ੠ͷൺֱσϞɻແม׵Ͱ͕͢ɺ߹੒ޙͷԻ੠͸΄΅ྼԽͳ͠ɻ
    ೖྗʗग़ྗԻ੠ͷൺֱ

    View full-size slide

  29. ม׵ϞσϧΛߏங͢Δݚڀಈ޲
    ੠࣭ม׵Ϟσϧͷ୅දతͳֶशख๏
    ଓ͍ͯɺԻڹಛ௃ྔΛֶशͤ͞ΔϞσϧʹ͍ͭͯɻ୅දతͳͷ͕ίνϥɻ
    "OBMZTJT
    ෼ੳ
    4ZOUIFTJT
    ߹੒

    جຊप೾਺'
    ϝϧέϓετϥϜ
    ඇपظੑࢦඪ

    جຊप೾਺'
    ϝϧέϓετϥϜ
    ඇपظੑࢦඪ
    ઢܗม׵
    ແม׵
    ৭ʑ
    ػցֶशʹΑΔਪఆ͸ɺ
    ϝϧέϓετϥϜͷΈͷ৔߹͕ଟ͍ɻ
    $POWFSTJPO
    ม׵

    View full-size slide

  30. ม׵ϞσϧΛߏங͢Δݚڀಈ޲
    ӡ༻ʹͷͤ΍ֶ͍͢शख๏ͷಛ௃
    σʔλ͕ूΊ΍͘͢ɺϞσϧߏஙͷख͕ؒগͳ͍ಛ௃Λ΋ͭख๏΋͋Γ·͢ɻ
    ϊϯύϥϨϧσʔλ
    ಉҰͷจষΛಡΈ্͛ͨ̎छྨͷϖΞԻ੠͕ʮύϥϨϧσʔλʯͰ͋Δ
    ͷʹରͯ͠ɺҟͳΔ̎छྨͷϖΞԻ੠ͷ͜ͱΛࢦ͢ɻ
    ଟରଟͷม׵
    ࿩ऀ"͔Β࿩ऀ#΁ͷม׵Λֶश͢Δɺ̍ର̍ͷม׵ख๏ʹରͯ͠ɺ
    ༻ҙͨ͢͠΂ͯͷ࿩ऀͷ૊Έ߹ΘͤΛҰ౓ʹֶश͢Δม׵ख๏ͷ͜ͱɻ
    ࣮૷σϞ
    ʢ4UBS("/7$ʣ

    View full-size slide

  31. ·ͱΊ
    ɹϝσΟΞͷϦιʔεͰӡ༻͢Δ̎ͭͷ75VCFS૾ΛఏҊͨ͠

    75VCFSΛλϨϯτͰͳ͘χϡʔεഔମͱͯ͠ଊ͑Δ͜ͱͰɺ
    ػցֶशʹΑͬͯ75VCFSΛʮ࣮૷ʯͰ͖Δɻ
    ɹ੠࣭ม׵ͷجૅʙ࠷ۙͷݚڀಈ޲·Ͱɺͬ͘͟Γղઆͨ͠

    Իڹֶͷલఏ஌ࣝͱɺ࠷௿ݶͷ࣮૷ํ๏ΛΧόʔͨ͠ɻ
    ֶशϞσϧͷߏஙํ๏ͷղઆ͸ɺ·ͨผͷػձʹɻ
    ࠓճ͸̎ͭͷςʔϚͰ঺հ͠·ͨ͠ɻߋʹ۩ମతͳ͓࿩͸·ͨผͷػձʹɻ

    View full-size slide

  32. ͪΐͬͱ͚ͩએ఻
    ຊ೔ͷൃද಺༰ͷ
    ৄ͍͠ղઆ͸ɺ
    དྷिͷٕज़ॻయ̓Ͱ
    ൦෍༧ఆͰ͢ɻ

    ʙ
    ஑ାαϯγϟΠϯγςΟ'
    లࣔϗʔϧ$͍$
    ಉ༷ͷ಺༰Λগ͠ৄ͘͠ॻ͍ͨهࣄΛࡌͤͨຊΛɺདྷि൦෍͠·͢ɻͥͻɻ

    View full-size slide

  33. 5IBOLZPV
    Ҏ্ɺ͝੩ௌ͋Γ͕ͱ͏͍͟͝·ͨ͠ɻ
    SFMBUFEEPDT
    ɾAIχϡʔεΩϟελʔʮ޷࣮ΤϦΧʯ։ൃऀΠϯλϏϡʔ
    https://note.mu/nikkei_jisedai/n/n4fc1cd3fd9b5
    ɾԻ੠෼ੳ߹੒αϯϓϧίʔυ
    https://github.com/Hirosaji/toy-codes/blob/master/handle_wav/forPyCon/
    analyze_and_synthesize_wav.py
    ɾhujinsen / StarGAN Voice Conversion
    https://github.com/hujinsen/StarGAN-Voice-Conversion
    ɾLearn Amazon Sumerian
    https://docs.sumerian.amazonaws.com/

    View full-size slide

  34. 5IBOLZPV
    Ҏ্ɺ͝੩ௌ͋Γ͕ͱ͏͍͟͝·ͨ͠ɻ
    SFGFSFODF
    ɾӋాཅҰɼେ઒ໜथɼ໦୩ढ़հɿԻڹֶೖ໳ϖσΟΞ, ίϩφࣾʢ2017ʣ
    ɾ৿੎কխɿԻ੠෼ੳ߹੒, ίϩφࣾʢ2018ʣ
    ɾߥ઒ ཮ɼߴಓ ৻೭հɼԐ౉ ༸ɿϦΞϧλΠϜDNNԻ੠ม׵ͷ࣮૷ͱσʔλ
    ɹ֦ு๏ʹΑΔԻ࣭վળ๏, ೔ຊԻڹֶձ2019೥य़قݚڀൃදձߨԋ࿦จू,
    ɹ1-10-4.ʢ2019ʣ
    ɾToda, T., Black, A. W. and Tokuda, K.: Voice conversion based on
    ɹmaximum-likelihood estimation of spectral parameter trajectory,
    IEEE T. ASLP.ʢ2007ʣ

    View full-size slide

  35. 5IBOLZPV
    Ҏ্ɺ͝੩ௌ͋Γ͕ͱ͏͍͟͝·ͨ͠ɻ
    SFGFSFODF
    ɾTakuhiro, K. and Hirokazu, K.: Parallel-Data-Free Voice Conversion
    ɹUsing Cycle-Consistent Adversarial Networks, arXiv.ʢ2017ʣ
    ɾFuming, F., Junichi Y., Isao, E., and Jaime, L. T.: High-quality
    ɹnonparallel voice conversion based on cycle-consistent adversarial
    network, ICASSP.ʢ2018ʣ
    ɾKameoka, H., Kaneko, T., Tanaka, K., Hojo, N.ɿStarGAN-VC: Non-parallel
    ɹmany-to-many voice conversion using star generative adversarial networks.,
    In 2018 IEEE Spoken Language Technology Workshop (SLT), pp. 266-273,
    IEEE.ʢ2018ʣ

    View full-size slide