Tokyo BISH Bash #02 音声情報処理と音声変換技術入門

Slide 1

Slide 1 text

Ի੠ม׵ٕज़ೖ໳ ʙԻ੠৘ใॲཧೖ໳Λఴ͑ͯʙ Tokyo BISH Bash #02 2020/06/30

Slide 2

Slide 2 text

໨࣍ • Ի੠৘ใॲཧೖ໳ʢʙ5෼ʣ • Ի੠ม׵ٕज़ೖ໳ʢ20෼ʣ • ศརͳύοέʔδɾιϑτ΢ΣΞͷ঺հʢ࢒Γ࣌ؒʣ ߴߍଔۀఔ౓ͷʮ೾ಈʯͷ஌ࣝ͸લఏͱ͠·͢; ৼ෯ɺप೾਺ͳͲ ຊൃදͰ͸ԻΛྲྀ͠·͢ͷͰɺ͓खݩͷPCͷԻྔʹ͝஫ҙΛʂ 3

Slide 3

Slide 3 text

Ի੠৘ใॲཧೖ໳ Ի੠ͷʮ৘ใʯͬͯԿͧʁ

Slide 4

Slide 4 text

Ի੠ͱ͸Կ͔ʁ • ۭؒΛ఻ΘΔʮ೾ʯ • ۭؾ෼ࢠͷૈີ͕఻ΘΔʢഏ → ޴ → ޱ→ۭؒʣ • ίϛϡχέʔγϣϯͷʮϝσΟΞʯ • ݴޠ΍ײ৘ɺ࿩ऀੑͳͲΛ఻͑Δഔମ • จԽɾܳज़ʹ͓͚Δʮ෇ՃՁ஋ʯ • ਓͷ৺Λಈ͔͢Վএ 5 Tokyo BISH Bash ! Tokyo BISH Bash ̇ Tokyo BISH Bash ͕ྑ͍Μͩͬͯ͞

Slide 5

Slide 5 text

Ի੠ʹؚ·ΕΔ̏ͭͷ৘ใ 6 ݴޠ৘ใ ύϥݴޠ৘ใ ඇݴޠ৘ใ ςΩετԽ͕Մೳ • ςΩετԽ͕ෆՄೳ • ࿩ऀ͕ҙਤతʹੜ੒ • ײ৘΍ଶ౓ͳͲ • ςΩετԽ͕ෆՄೳ • ࿩ऀ͕ҙਤ͠ͳ͍ • ೥ྸ΍ੑผɺ࿩ऀੑ Ի੠ تͼͷ ײ৘ ʮ"͞Μʯ ͱ͍͏৘ใ "͞Μ

Slide 6

Slide 6 text

Ի੠৘ใॲཧͱ͸Կ͔ʁ Ի੠ʹؚ·ΕΔʮ̏ͭͷ৘ใʯΛՃ޻/ಉఆ/ॲཧ͢Δٕज़ • Ի੠߹੒ɺԻ੠ೝࣝɺ࿩ऀೝࣝɺײ৘ೝࣝɺԻ੠ม׵ɺͳͲ ⇒ ຊߨԋ͸ʮԻ੠ม׵ʯʢಛʹ࿩ऀม׵ʣʹϑΥʔΧε • Ի੠ม׵ɿݴޠ৘ใΛอ࣋ͭͭ͠࿩ऀͷಛੑΛมߋ͢Δٕज़ 7 ࿩ऀͷมߋ ࿩͠ํɾײ৘ͷมߋ ຊߨԋͷ ओ୊͸ ͬͪ͜ˠ ඇݴޠ৘ใͷΈΛมߋ ύϥݴޠ৘ใͷΈΛมߋ

Slide 7

Slide 7 text

Ի੠৘ใॲཧͱ͸Կ͔ʁ Ի੠ʹؚ·ΕΔʮ̏ͭͷ৘ใʯΛՃ޻/ಉఆ/ॲཧ͢Δٕज़ • Ի੠߹੒ɺԻ੠ೝࣝɺ࿩ऀೝࣝɺײ৘ೝࣝɺԻ੠ม׵ɺͳͲ ⇒ ຊߨԋ͸ʮԻ੠ม׵ʯʢಛʹ࿩ऀม׵ʣʹϑΥʔΧε • Ի੠ม׵ɿݴޠ৘ใΛอ࣋ͭͭ͠࿩ऀͷಛੑΛมߋ͢Δٕज़ 8 ม׵ख๏໊ ݴޠ৘ใ ύϥݴޠ ৘ใ ඇݴޠ ৘ใ Իӆม׵ ྫ Bˠ J มߋ ݻఆ ݻఆ ײ৘ม׵ ྫ ت ˠ ౖ ݻఆ มߋ ݻఆ ࿩ऀม׵ ྫ ঁੑ"ˠ உੑ# ݻఆ ݻఆ มߋ

Slide 8

Slide 8 text

Ի੠ม׵ٕज़ͷجૅ ಛʹ࿩ऀʢ੠࣭ʣม׵ʹϑΥʔΧε

Slide 9

Slide 9 text

͸͡Ίʹ • όʔνϟϧYouTuberʢVTuberʣͷོ੝ 10 %%ΩϟϥΫλʔ όʔνϟϧYouTuber

Slide 10

Slide 10 text

͸͡Ίʹ • உੑʢঁੑʣ ͕ঁੑʢஉੑʣΩϟϥΫλʔʹ ⇒ ϘΠενΣϯδϟʔͷར༻ 11 ϘΠενΣϯδϟʔ ϘΠενΣϯδϟʔͷ࢓૊Έ JTԿʁ

Slide 11

Slide 11 text

Ի੠ͷੜ੒աఔʢͬ͘͟Γͱʣ • ࣮͸؅ָثͱྨࣅ 12 ઉ΍ޱͷܗΛมܗ ੠ଳͷ։ดʹΑΔ ۭؾͷৼಈ ϛοΫεʂ Ի৭ Իߴ ੠ʹͳͬͯग़ͯ͘Δ ์ࣹ ʮ͜Μʹͪ͸ʯ https://ahcweb01.naist.jp/lecture/2018/sp/ material/20181108_slide.pdf ࢀߟ ͜Μʹͪ͸

Slide 12

Slide 12 text

Ի੠ͷੜ੒աఔʢप೾਺෼ੳͷ؍఺͔Βʣ • प೾਺ಛੑදݱ 13 ύϫʔ प೾਺ Ի৭ͷ෇༩ʢڞ໐ʣ ͳΊΒ͔ ύϫʔ प೾਺ Իߴͷੜ੒ʢԻݯʣ ΪβΪβ มௐ ύϫʔ प೾਺ प೾਺ಛੑ ˞ͳΊΒ͔੒෼ɿεϖΫτϧแབྷ ˞ΪβΪβ੒෼ɿඍࡉߏ଄ ˞มௐɿ৞ΈࠐΈૢ࡞

Slide 13

Slide 13 text

σϞϯετϨʔγϣϯ XJUI1ZUIPO • ೾ܗͷϦΞϧλΠϜϓϩοτʮ͍͋͏͓͑ʯ 14 ֤฼ԻͰҟͳΔ೾ܗ

Slide 14

Slide 14 text

σϞϯετϨʔγϣϯ XJUI1ZUIPO • FFTεϖΫτϧͷϦΞϧλΠϜϓϩοτʮ͍͋͏͓͑ʯ 15 εϖΫτϧแབྷʢͳΊΒ͔ʣ ʴඍࡉߏ଄ʢΪβΪβʣ

Slide 15

Slide 15 text

σϞϯετϨʔγϣϯ XJUI1ZUIPO • εϖΫτϧแབྷͷϦΞϧλΠϜϓϩοτʮ͍͋͏͓͑ʯ 16 εϖΫτϧแབྷʢͳΊΒ͔ʣ

Slide 16

Slide 16 text

Ի੠ੜ੒աఔͷϞσϧԽʢʣ • Իݯ৴߸ΛΠϯύϧεྻͱന৭ࡶԻͰදݱ • Πϯύϧεྻ͕༗੠Իʢ੠ଳৼಈΛ൐͏ʣʹରԠ • ന৭ࡶԻ͕ແ੠Իʢ੠ଳৼಈΛ൐Θͳ͍ʣʹରԠ 17 https://ahcweb01.naist.jp/lecture/2018/sp/material/20181108_slide.pdf Πϯύϧεྻ ˠ༗੠ԻʹରԠ ന৭ࡶԻ ˠແ੠ԻʹରԠ Իྔ஫ҙʂ

Slide 17

Slide 17 text

Ի੠ੜ੒աఔͷϞσϧԽʢʣ • εϖΫτϧแབྷ͸Իڹ؅ͷ࿈݁Ͱදݱ • ͷͲͷܗͱ੠ͷ௨Γಓʢ੠ಓʣΛදݱ • ͷͲͷܗͷݸਓࠩ = εϖΫτϧแབྷͷݸਓࠩ ⇒ʮ੠৭ʯ 18 ʮ؅ʯͷ࿈݁ ੠ଳଆ ޱ৶ଆ प೾਺ ύϫʔ εϖΫτϧแབྷ

Slide 18

Slide 18 text

Ի੠ੜ੒աఔͷϞσϧԽʢʣ • Իݯ৴߸ΛԻڹ؅ʹ௨͢⇒ Ի੠ੜ੒ʢϑΟϧλϦϯάʂʣ 19 ༗੠Ի ແ੠Ի Իݯ Իݯ৴߸ εϖΫτϧแབྷ Ի੠ੜ੒ ˎ ʹ Իڹ؅ʢεϖΫτϧแབྷʣ ߹੒೾ܗ Իݯ͸HTS slide ΑΓ http://hts.sp.nitech.ac.jp/?Download ݩͷԻ੠

Slide 19

Slide 19 text

Ի੠ಛ௃ྔͷऔಘʢʣ • Ի੠ͷੜ੒աఔʹج͍ͮͯಛ௃ྔΛநग़͠࠶߹੒ ⇒ ϘίʔμʔʢVocoderʣͷద༻ 20 ੠৭ ʢεϖΫτϧแབྷʣ ੠ͷߴ͞ ʢجຊप೾਺ʣ ͔͢Ε۩߹ ʢඇपظੑࢦඪʣ Ի੠෼ੳ Ի੠ੜ੒ ʮ͜Μʹͪ͸ʯ ʮ͜Μʹͪ͸ʯ

Slide 20

Slide 20 text

Ի੠ಛ௃ྔͷऔಘʢʣ • Ի੠Λ෼ੳ ⇒ ಛ௃ྔͷ࣌ܥྻΛऔಘ 21 ൣғΛগͣͭͣ͠Βͯ͠෼ੳ ˞͜Ε͸εϖΫτϧแབྷ

Slide 21

Slide 21 text

Ի੠ಛ௃ྔͷऔಘʢʣ • Ի੠Λ෼ੳ ⇒ ಛ௃ྔͷ࣌ܥྻΛऔಘ 22 ˞͜Ε͸جຊप೾਺ ൣғΛগͣͭͣ͠Βͯ͠෼ੳ

Slide 22

Slide 22 text

Ի੠෼ੳ͔ΒԻ੠ੜ੒·ͰͷྲྀΕ • 23 Ի੠෼ੳ εϖΫτϧ แབྷ جຊप೾਺ ඇपظੑ ࢦඪ Ի੠ੜ੒ ᶃԻ੠Λ෼ੳͯ͠ ಛ௃ྔϕΫτϧͷ ܥྻΛखʹೖΕΔ ʜ ʜ ෼ੳ୯Ґ ʹϑϨʔϜ ᶅੜ੒͞Εͨ೾ܗΛ ॏͶ߹ΘͤΔ ᶄ֤ϑϨʔϜͷಛ௃ྔΛ Ϙίʔμʔʹ௨͢ ᶆݩͷԻ੠೾ܗ͕ ࠶ߏ੒͞ΕΔ ʜ ෼ੳ ੜ੒

Slide 23

Slide 23 text

Ի੠ม׵ʢϘΠενΣϯδϟʔʣͷߟ͑ํ 24 Ի੠෼ੳ εϖΫτϧ แབྷ جຊप೾਺ ඇपظੑࢦඪ Ի੠ੜ੒ • ৴߸ॲཧͰม׵ • ػցֶशͰม׵ Ի੠ಛ௃ྔͷม׵ Ի੠ม׵ ಛ௃ྔؒͷࣸ૾

Slide 24

Slide 24 text

ύϥϝʔλ੍ޚʹΑΔԻ੠Ճ޻ʢʣ • ෼ղͨ͠Ի੠ಛ௃ྔͷՃ޻ 25 ੠৭ ʢεϖΫτϧแབྷʣ ੠ͷߴ͞ ʢجຊप೾਺ʣ ͔͢Ε۩߹ ʢඇपظੑࢦඪʣ Ի੠෼ੳ Ի੠ಛ௃ྔʢύϥϝʔλʣ Ի੠ੜ੒ Ͳ͔͜ͷ#͞Μ ʮ͜Μʹͪ͸ʯ ม׵ɾՃ޻ॲཧ "͞Μ ʮ͜Μʹͪ͸ʯ

Slide 25

Slide 25 text

ύϥϝʔλ੍ޚʹΑΔԻ੠Ճ޻ʢʣ • جຊप೾਺ͷมߋ 26 ࣌ؒ ੠ଳԻݯʢύϧεܥྻʣ ִؒΛڱΊΔ ִؒΛ޿͛Δ ݩͷ੠ ߴ͍੠ ௿͍੠ ੠͸ ߴ͘௿͘ ͳΔ͕ ΍΍ෆࣗવ͕͞࢒Δ

Slide 26

Slide 26 text

ύϥϝʔλ੍ޚʹΑΔԻ੠Ճ޻ʢʣ • ϑΥϧϚϯτγϑτ ⇒ ੠ಓͷ৳ॖ ʹରԠ 27 प೾਺ ύϫʔ ࠨʹγϑτ ӈʹγϑτ ͜΋ͬͨʁ੠ ෺ཧతͳղऍɿ • ੠ಓ͕୹͍ 㱺 ࢠڙঁੑ • ੠ಓ͕௕͍ 㱺 େਓஉੑ ੠ Ӊ஦ਓʁ੠ ੠ಓͷ୹ॖ ੠ಓͷ৳௕ ݩͷ੠

Slide 27

Slide 27 text

ύϥϝʔλ੍ޚʹΑΔԻ੠Ճ޻ʢʣ • प೾਺ˍϑΥϧϚϯτͷมߋ ⇒ ΑΓࣗવͳ੠Λ߹੒ 28 ݩͷ੠ ࠨγϑτ ˍ ڱִ͍ؒ ӈγϑτ ˍ ޿ִ͍ؒ ࢠڙঁੑͷ੠ ໺ଠ͍੠

Slide 28

Slide 28 text

σϞϯετϨʔγϣϯ ݚڀࣨͷֶੜ͞Μ͕࡞੒ͨ͠ΞϓϦ • εϥΠμʔͰύϥϝʔλ੍ޚ • ҎԼΛϕʔεͱͯ͠࡞੒ • $൛ 803-% ˞̍ • +6$&˞̎ 29 ˞̍ɿߴ඼࣭Ϙίʔμʔͷ̍ͭ ˞̎ɿΦʔσΟΦܥΞϓϦͷ ։ൃʹదͨ͠$ϑϨʔϜϫʔΫ εϥΠμʔͰ ੍ޚ εϖΫτϧ දࣔ Ӷҙ։ൃதͰ͢ʂ

Slide 29

Slide 29 text

Ի੠ม׵ٕज़ͷൃల ͍͔ͭ͘ͷൃలͷʮํ޲ੑʯΛڍ͛ΔʹཹΊ·͢ ໢ཏతͰ΋ͳ͘ɺશͯΛ঺հ͖͠Ε·ͤΜ͕ɺ͝༰͍ࣻͩ͘͞

Slide 30

Slide 30 text

౷ܭత੠࣭ม׵ • ౷ܭϞσϧΛར༻ͯ͠Ի੠ಛ௃ྔΛม׵ ⇒ ಛ௃ྔͷରԠؔ܎ʢࣸ૾ʣΛେྔͷσʔλͰֶश͓ͯ͘͠ඞཁ 31 Ի੠෼ੳ Ի੠ੜ੒ #͞Μʮ͜Μʹͪ͸ʯ ౷ܭϞσϧ ʹΑΔม׵ "͞Μʮ͜Μʹͪ͸ʯ εϖΫτϧแབྷ جຊप೾਺ ඇपظੑࢦඪ • Ψ΢εࠞ߹Ϟσϧ • χϡʔϥϧωοτ ൃ࿩಺༰Λอͪͭͭ ࿩ऀͷ੠࣭Λม׵ https://www.slideshare.net/KentaroTachibana1/ss-94259438 ࢀߟ ϘίʔμʔʹΑΔ Ի੠ੜ੒

Slide 31

Slide 31 text

ൃలͷํ޲ੑʢʣ • ͦͷ̍ɿϘίʔμʔͷԻ੠඼࣭Λྑͯ͘͠ੑೳ޲্ ⇒ χϡʔϥϧϘίʔμʔͷొ৔ʢ2017೥Ҏ߱ʣ ࢀߟʰԻ੠೾ܗ௚઀ੜ੒ϞσϧʮχϡʔϥϧϘίʔμʯͷൺֱʱ(2019.10) ͳͲ • ͦͷ̎ɿϘίʔμʔΛ࢖Θͣʹ੠࣭ม׵ʂ ࠩ෼εϖΫτϧิਖ਼ʹجͮ͘੠࣭ม׵ ⇒ ϘίʔμʔʹΑΔԻ࣭ͷྼԽΛܰݮՄೳɺϦΞϧλΠϜࢤ޲ 32 https://www.slideshare.net/Takuma_OKAMOTO/ss-180895505 • ཁ๬ɿߴ඼࣭ɺߴ଎ͳ߹੒ɺֶश͕༰қɺলϝϞϦԽɺɺɺ ⇒ Ի੠ม׵γεςϜʹ΋૊ΈࠐΜͰੑೳ޲্Λ໨ࢦ͢

Slide 32

Slide 32 text

ൃలͷํ޲ੑʢʣ • ͦͷ̎ɿԻ੠ͷ௥Ճऩ࿥ʹ͔͔ΔίετΛԼ͛Δʂ • ͦͷ̏ɿ੠࣭͚ͩͰͳ͘ײ৘΋ม׵ʂ • ͦͷ̐ɿ࿩͠੠͚ͩͰͳ͘Վ੠ͷ੠࣭΋ม׵ʂ • ͦͷ̑ɿϦΞϧλΠϜ͔ͭߴ඼࣭ͳ੠࣭ม׵ʂ ʮχϡʔϥϧωοτશ੝ʯͷ࣌୅ • χϡʔϥϧωοτͷೖྗଆͱग़ྗଆʹԿΛ࣋ͬͯ͘Δ͔͕ΧΪ • σʔλ͑͞༻ҙͰ͖Ε͹ࣸ૾ֶ͕शͰ͖Δ͔΋ʢͰ͖Δͱ͸ݴͬͯͳ͍ʣ • ৴߸ॲཧʹجͮ͘ख๏͸͜Ε͔Β΋݈ࡏʢͷ͸ͣʣ 33

Slide 33

Slide 33 text

Ի੠৘ใॲཧ΍Ի੠ม׵ʹ ศརͳ (Python) ύοέʔδ΍ ιϑτ΢ΣΞͨͪ ࢖༻ײͳͲࢲݟΛؚΈ·͢

Slide 34

Slide 34 text

TPYQZTPY • ίϚϯυϥΠϯ͔ΒϑΥʔϚοτม׵ͳͲΛ͓खܰʹ • ϑΥʔϚοτม׵ʢwav to mp3 ͳͲʣ • ݁߹΍ϛοΫεɺτϦϛϯά ΋Մೳ • όονॲཧ΋ָʢγΣϧεΫϦϓτͳͲʣ • Pythonϥούʔ pysox΋͋Δ • Πϯετʔϧ • brew install sox ͳͲ • pip install sox ← pysox ͷΠϯετʔϧ͸͜Ε 35

Slide 35

Slide 35 text

MJCSPTBʢ͜Ε͸ຊ౰ʹΦεεϝʣ • Ի੠/Իָͷ෼ੳʹศརͳϞδϡʔϧ͕ଗͬͨύοέʔδ • ެࣜϚχϡΞϧɾνϡʔτϦΞϧͷॆ࣮΋ॿ͔Δ • ݸਓతʹΑ͘࢖͏ػೳ • ೾ܗදࣔɺεϖΫτϩάϥϜදࣔ • Ի੠ಛ௃ྔநग़ʢର਺ϝϧεϖΫτϩάϥϜʣ • Πϯετʔϧ pip install librosa • ެࣜϖʔδ https://librosa.org/librosa/index.html 36

Slide 36

Slide 36 text

1Z8PSME • Ի੠ͷ෼ੳ࠶߹੒Λߦ͏Ϙίʔμʔͷύοέʔδ • Ի੠Λʮ੠৭ɾ੠ͷߴ͞ɾ੠ͷ͔͢Εʯͷ֤੒෼ʹ෼ղ͠࠶߹੒ • C++൛ͷPythonϥούʔ • Ի੠ͷಛ௃ྔநग़ʹ΋࢖͑ͯศར ⇒ PySPTKʢޙड़ʣΑΓ΋඼࣭ͷΑ͍εϖΫτϧแབྷ • Πϯετʔϧ pip install pyworld • ެࣜϖʔδ https://github.com/JeremyCCHsu/Python-Wrapper-for-World-Vocoder 37

Slide 37

Slide 37 text

1Z"VEJP • ετϦʔϜ࿥Ի / ࠶ੜʹศརͳύοέʔδ • ϦΞϧλΠϜͷԻ੠ೖྗɾԻ੠ग़ྗʹ࢖͑Δ • ϦΞϧλΠϜԻ੠ม׵ with PythonͳͲ΋Մೳ • Πϯετʔϧ • pip install pyaudio ཁportaudio (e.g., brew install portaudio) 38

Slide 38

Slide 38 text

1Z"VEJPͱ1Z8PSMEͷ૊Έ߹Θͤ • ؆қ൛ͷϘΠενΣϯδϟʔ • ؆қϘΠενΣϯδϟʔͷεΫϦϓτΛվྑɿPyQt5ͷεϥΠμʔʹΑΓ ϐονͱϑΥϧϚϯτΛϦΞϧλΠϜௐ੔͢ΔػೳΛ௥Ճʢฐϒϩάʣ • banibiku • Zoom഑৴޲͚ʹ̎࣍ݩΩϟϥʹͳΓ͖Δ͜ͱΛ໨ࢦͨ͠ϓϩδΣΫτ • scripts/voice_converter.py ͕ྑ͍ײ͡ͷϘΠενΣϯδϟʔ → ฐϒϩάͷαϯϓϧεΫϦϓτͷόάϑΟοΫεؚ͕·ΕΔ 39 https://tam5917.hatenablog.com/entry/2019/04/30/213321 https://github.com/peisuke/babiniku

Slide 39

Slide 39 text

1Z"VEJPͱ1Z2Uͷ૊Έ߹Θͤ • ೾ܗϞχλϦϯάʢʮ͍͋͏͓͑ʯʣ • ࢀߟ ϦΞϧλΠϜʹมԽ͢ΔԻ੠ͷ೾ܗΛදࣔ͠ଓ͚ΔPythonεΫϦϓτʢฐϒϩάʣ 40 https://tam5917.hatenablog.com/entry/2019/04/28/130641

Slide 40

Slide 40 text

1Z"VEJPͱ1Z2Uͷ૊Έ߹Θͤ • FFTεϖΫτϧϞχλϦϯάʢʮ͍͋͏͓͑ʯʣ • ࢀߟ ϦΞϧλΠϜʹมԽ͢ΔԻ੠ͷFFTεϖΫτϧΛදࣔ͢ΔPythonεΫϦϓτʢฐϒϩάʣ 41 https://tam5917.hatenablog.com/entry/2019/04/28/125857

Slide 41

Slide 41 text

1Z"VEJPͱ1Z2Uͷ૊Έ߹Θͤ • εϖΫτϧแབྷϞχλϦϯά ʢʮ͍͋͏͓͑ʯʣ ࢀߟ ϦΞϧλΠϜʹมԽ͢ΔԻ੠ͷεϖΫτϧแབྷΛදࣔ͢ΔPythonεΫϦϓτʢฐϒϩάʣ 42 https://tam5917.hatenablog.com/entry/2019/04/28/130641

Slide 42

Slide 42 text

1Z415, • Ի੠৘ใॲཧπʔϧΩοτSPTKͷPythonϥούʔ • SPTKࣗମ͸LinuxίϚϯυ܈ • Իڹಛ௃ྔநग़ʹ࢖͏ͷ͕ศར • Ի੠෼ੳ߹੒΋Ͱ͖Δ͕ɺ඼࣭ࣗମ͸WORLDͷ΄͏্͕ • Πϯετʔϧ pip install pysptk • ެࣜϖʔδ https://pysptk.readthedocs.io/en/latest/ 43

Slide 43

Slide 43 text

OONOLXJJ • DNNԻ੠߹੒ʹ໾ཱͭϞδϡʔϧΛूΊͨύοέʔδ • ͲͪΒ͔ͱ͍͏ͱݚڀ༻్ • લॲཧ΍Իڹಛ௃ྔநग़ͷΫϥε͕Ұ௨Γଗ͍ͬͯΔ • ࿦จͷ࠶ݱ࣮૷Λ͢Δͱ͖ͳͲʹେ͍ʹ໾ཱͭ • Πϯετʔϧ pip install nnmnkwii • ެࣜϖʔδ https://r9y9.github.io/nnmnkwii/stable/index.html 44

Slide 44

Slide 44 text

1ZEVC • Pydub • ೾ܗฤूʹศརͳϞδϡʔϧΛूΊͨύοέʔδ • αϙʔτ͢ΔϑΝΠϧܗࣜ΋๛෋ʢwav, mp3, mp4, wma, aac, ...ʣ • ػೳ ੾Γग़͠ɺ෼ׂɺϛοΫεɺϑΣʔυΠϯΞ΢τɺແԻૠೖɺͳͲͳͲ • Ұ෦ͷػೳ͸ pysoxͷ΄͏͕ߴ଎ͱ͍͏ӟ?ʢະ֬ೝʣ • Πϯετʔϧ pip install pydub • ެࣜϖʔδ http://pydub.com/ 45

Slide 45

Slide 45 text

TQSPDLFU • ౷ܭత੠࣭ม׵ͷͨΊͷπʔϧΩοτ (not ύοέʔδ) • ͲͪΒ͔ͱ͍͏ͱݚڀ༻ ʢMITϥΠηϯεʣ • ݚڀͷʮϕʔεϥΠϯʯߏஙʹ࠷ద • ެࣜϖʔδ https://github.com/k2kobayashi/sprocket • ղઆ࿦จ ʰ౷ܭత੠࣭ม׵ιϑτ΢ΣΞೖ໳ʱ https://www.jstage.jst.go.jp/article/isciesci/62/2/62_69/_article/-char/ja/ • νϡʔτϦΞϧ (εϥΠυ & notebook) https://github.com/kan-bayashi/INTERSPEECH19_TUTORIAL 46

Slide 46

Slide 46 text

"VEBDJUZ ೖΕ͓ͯ͘ͱ҆৺ • ϑϦʔͷ೾ܗฤूιϑτɺϚϧνϓϥοτϑΥʔϜ • ๛෋ͳα΢ϯυΤϑΣΫτՃ޻ػೳ • ެࣜϖʔδ https://www.audacityteam.org/ 47

Slide 47

Slide 47 text

ͦͷଞ • Voidol • ϦΞϧλΠϜ੠࣭ม׵ʢ༗ྉʣ • ࣗ෼ͷ੠ΛλʔήοτͷΩϟϥʹม׵ • Realtime Yukarin • ϦΞϧλΠϜ੠࣭ม׵ • Gachikoe! Core • ϦΞϧλΠϜ੠࣭ม׵ 48 https://crimsontech.jp/apps/voidol/ https://blog.hiroshiba.jp/realtime-yukarin-introduction/ https://booth.pm/ja/items/1236505

Slide 48

Slide 48 text

͓·͚ • GoogleԻ੠߹੒ɺԻ੠ೝࣝɺ຋༁APIͷPythonϥούʔ ⇒ gTTS, SpeechRecognition, googletrans • ͦΕͧΕΛ͓ࢼ͠Ͱ࢖ͬͯΈΔʹ͸े෼ • ࢀߟ GoogleԻ੠ೝࣝͷ݁ՌΛGoogle຋༁͠ɺGoogle Text-to-SpeechͰ Ի੠ʹ໭͢PythonεΫϦϓτʢฐϒϩάʣ https://tam5917.hatenablog.com/entry/2019/04/28/191946 49

Slide 49

Slide 49 text

͓ΘΓʹ • Ի੠৘ใॲཧͷʢ௒ʣೖ໳ Ի੠ʹؚ·ΕΔ৘ใʢݴޠɾύϥݴޠɾඇݴޠʣΛॲཧ ⇒ Ի੠߹੒΍Ի੠ೝࣝɺ࿩ऀ/ײ৘ೝࣝɺԻ੠ม׵ ͳͲ • Ի੠ม׵ٕज़ͷ঺հ ϘΠενΣϯδϟʔͷ࢓૊ΈɿԻ੠ͷੜ੒աఔͷϞσϧԽ͕جૅ • ศརͳʢPythonʣύοέʔδΛ঺հ 50 Ի੠৘ใॲཧɾԻ੠ม׵ٕज़ʹ ڵຯΛ࣋ͬͯ΋Β͑ͨΒ޾͍Ͱ͢