...

歌唱合成音声のハスキーボイス化 A method for transforming normal

by user

on
Category: Documents
7

views

Report

Comments

Transcript

歌唱合成音声のハスキーボイス化 A method for transforming normal
Vol.2013-MUS-99 No.36
2013/5/12
᝟ሗฎ⌮Ꮫ఍◊✲ሗ࿌
IPSJ SIG Technical Report
ḷၐྜᡂ㡢ኌࡢࣁࢫ࣮࢟࣎࢖ࢫ໬
బࠎᮌ ᫍኴᮁ†1 ⓑᮌ ၿᑦ
†1
ḷၐྜᡂ㡢ኌ࡜ࡣ㸪㡢ኌྜᡂ࣭ࢹࢫࢡࢺࢵࣉ࣑࣮ࣗࢪࢵࢡ㸦DTM㸧ࢯࣇࢺ࢙࢘࢔➼ࢆ⏝࠸࡚㸪ࢥࣥࣆ࣮ࣗࢱୖ࡛⏕
ᡂࡉࢀࡓḷၐ㡢ኌ࡛࠶ࡿ㸬ᮏ◊✲࡛ࡣ㸪ḷၐྜᡂ㡢ኌࡢࣁࢫ࣮࢟࣎࢖ࢫ໬ࡢ᪉ἲࢆᥦ᱌ࡍࡿ㸬ࣁࢫ࣮࢟࣎࢖ࢫ໬ࡢᇶ
ᮏⓗ࡞ᡭἲࡣ㸪㡢ኌಙྕࡢኌᖏ᝟ሗ࡛࠶ࡿṧᕪಙྕ࡟ฎ⌮ࢆຍ࠼ࡿࠕṧᕪኚ᥮࡛ࠖ࠶ࡿ㸬ࡇࡢࠕṧᕪኚ᥮ࠖ࡟ࡘ࠸࡚
ᮏ◊✲࡛ࡣ஧✀㢮ࡢ᪉ἲࢆᥦ᱌ࡋ࡚࠸ࡿ㸬ࡉࡽ࡟ࠕṧᕪኚ᥮ࠖฎ⌮ᚋࡢ㡢㉁ྥୖࢆᅗࡿࡓࡵ࢖ࢥࣛ࢖ࢨ࣮ࡢᛂ⏝ࢆヨ
ࡳࡿ㸬ࡲࡓṧᕪಙྕࡢ᣺ᖜ᝟ሗ࠿ࡽࠕṧᕪኚ᥮ࠖฎ⌮ࡢࣦ࢙ࣞࣝࢆỴᐃࡍࡿࡇ࡜࡟ࡼࡗ࡚㸪ࣁࢫ࣮࢟࣎࢖ࢫࡢ⮬ື⏕
ᡂࢆ⾜࡞࠺㸬ᥦ᱌ࡋࡓ஧✀㢮ࡢࠕṧᕪኚ᥮ࠖἲ㸪࢖ࢥࣛ࢖ࢨ࣮ᛂ⏝ἲ㸪⮬ື⏕ᡂἲ㸪ࡢ㸲ࡘࡢᡭἲࢆ⏝࠸࡚ྜᡂࡋࡓ
㡢ኌࡢ⫈ྲྀᐇ㦂ࢆ⾜࠸㸪㨩ຊⓗ࡞ࣁࢫ࣮࢟࣎࢖ࢫ໬ࡢࡓࡵ࡟ࡣ࡝ࡢࡼ࠺࡞ᡭἲࡸᣦᶆ࡟᭷⏝ᛶࡀ࠶ࡿࡢ࠿࡟ࡘ࠸࡚㸪
᳨ドࡍࡿ㸬
A method for transforming normal singing voice into husky
synthesized singing voice
Seitarou Sasaki†1
and Yoshinao Shiraki†1
The singing synthesized voice , is a singing voice using speech synthesis and the desktop music (DTM) software,
etc., is generated on a computer. In this paper, we propose a method for transforming normal singing voice into
husky synthesized singing voice. The residual signal is a vocal information of the speech signal. "This conversion
to the residual signal" is the basic techniques of husky voice conversion. In this study, we propose two methods for
this "residual conversion". In order to improve sound quality of the "residual conversion" treatment, we try to
apply the equalizer. Further by determining the Revel "residual conversion" process from the amplitude
information of the residual signal, we perform the automatic generation of husky voice. For generating synthesized
speech using the four methods of the two types of proposed "residual conversion" method, the equalizer
application method, an automatic generation method, we do listening tests. We do verification for two things;
Which approach or increase the degree of attraction husky voice ? What indicators suitable to measure the degree
of glamor husky voice ?
1. ࡣࡌࡵ࡟
㨩ຊࢆᣢࡘኌࡢ୍ࡘ࡜ࡋ࡚ᣲࡆࡽࢀࡿࡢࡀࣁࢫ࣮࢟࣎࢖ࢫ
ḷၐྜᡂ㡢ኌ࡜ࡣ㸪㡢ኌྜᡂ࣭ࢹࢫࢡࢺࢵࣉ࣑࣮ࣗࢪࢵ
ḷኌ࡜ࡣ␗࡞ࡾ㸪ࠕࡋࡷࡀࢀࠖࡸࠕ࠿ࡍࢀࠖ࡜࠸ࡗࡓ୍ぢ㸪
ࢡ㸦DTM㸧ࢯࣇࢺ࢙࢘࢔➼ࢆ౑⏝ࡋ㸪ࢥࣥࣆ࣮ࣗࢱୖ࡛⏕
⨾ࡋࡉ࡜┦཯ࡍࡿࡼ࠺࡞ᛶ㉁ࢆᣢࡘ㸬ࡋ࠿ࡋ㸪ယឋ࡞࡝࠸
ᡂࡉࢀࡓḷၐ㡢ኌ࡛࠶ࡿ㸬௦⾲ⓗ࡞౛࡜ࡋ࡚㸪ࢡࣜࣉࢺ࣭ࣥ
ࡗࡓே㛫ⓗ㨩ຊࢆឤࡌࡉࡏࡿሙྜࡀ࠶ࡾ㸪ᐇ㝿࡟ࣁࢫ࣮࢟
ࣇ࣮ࣗࢳ࣮࣭࣓ࣕࢹ࢕࢔࠿ࡽⓎ኎ࡉࢀࡓࠕึ㡢࣑ࢡࠖࡸ㣩
࣎࢖ࢫࢆᣢࡘከࡃࡢ᭷ྡ࡞࢔࣮ࢸ࢕ࢫࢺࡀᏑᅾࡍࡿ㸬ᮏ◊
࡛࠶ࡿ㸬ࣁࢫ࣮࢟࣎࢖ࢫࡣ㸪࢜࣌ࣛ࡞࡝࡛⪥࡟ࡍࡿ⨾ࡋ࠸
ᒇ㸭Ⳮⵦ㸦࠶ࡵࡸ࣭࠶ࡸࡵ㸧ไస࡟ࡼࡿࣇ࣮࢙ࣜ࢘࢔ࠕUTAUࠖ
✲࡛ࡣ㸪≉࡟㸪ࡇࡢࣁࢫ࣮࢟࣎࢖ࢫ࡟↔Ⅼࢆᙜ࡚㸪ḷၐྜ
ࡀᣲࡆࡽࢀࡿ[1]㸬㏆ᖺ㸪ࢹࢫࢡࢺࢵࣉ࣑࣮ࣗࢪࢵࢡ㸦DTM㸧
ᡂ㡢ኌࢆࣁࢫ࣮࢟࣎࢖ࢫ໬ࡍࡿ᪉ἲࢆᥦ᱌ࡍࡿ㸬ࡲࡓ㸪ࣁ
ࡸ࢚ࣞࢡࢺࣟࢲࣥࢫ࣑࣮ࣗࢪࢵࢡ㸦EDM㸧ࡢᬑཬ࣭ὶ⾜࡟
ࢫ࣮࢟࣎࢖ࢫ໬ࢆ㏻ࡋ㸪ே㛫ⓗ㨩ຊࢆᣦᶆ໬࣭≉ᚩ໬ࡍࡿ
క࠸㸪ࡇࢀࡽࡢḷၐྜᡂ㡢ኌࡀ౑⏝ࡉࢀࡿᶵ఍ࡀቑ࠼㸪ㄆ
஦ࡶᮏ◊✲ࡢ┠ⓗ࡛࠶ࡿ㸬
▱ᗘࡶ㧗ࡲࡗ࡚࠸ࡿ㸬ࡑࡋ࡚㸪ࡇࢀࡽࡢ㡢ኌࢆ⫗ኌ࡟㏆࠸
ᮏ◊✲࡛ࡣ㸪ࣁࢫ࣮࢟࣎࢖ࢫࢆḟࡢࡼ࠺࡟ᤊ࠼ࡿ㸬ࣁࢫ
ḷኌ࡜ࡋ࡚౑⏝ࡋࡓ࠸࡜࠸࠺㟂せࡀ㧗ࡲࡿ஦ࡶண᝿ࡉࢀࡿ㸬
࣮࢟࣎࢖ࢫ࡜ࡣ㸪࠿ࡍࢀࡓኌ㸪ࢲ࣑ኌ࡛࠶ࡾ㸪୺࡟ኌᖏ࠿
ᮏ◊✲࡛ࡣ⫗ኌ࡟㏆࠸ḷၐྜᡂ㡢ኌࡢ⏕ᡂࢆ┠ᶆ࡟⨨ࡃ㸬
ࡽ⏕ࡌࡿ㞧㡢ࢆྵࡴኌ࡛࠶ࡿ㸬ኌᖏࡢჾ㉁ⓗ⑓ኚࡸ㐠ື㯞
┠ᶆ㐩ᡂࡢࡓࡵࡢ୍ࡘࡢഃ㠃࡜ࡋ࡚㸪⊂≉࡞ḷኌ㸪ḷ࠸᪉
⑷࠿ࡽ⏕ࡌࡿ⑕≧࡛࠶ࡿႤኌࡀ㸪࠿ࡍࢀࡓኌ㸪ࢲ࣑ኌ࡞࡝
࡞࡝࡜ゝࡗࡓᴦ㆕ୖ࡟⾲ࡏ࡞࠸せ⣲㸪࠸ࢃࡤࠕே㛫ⓗ㨩ຊࠖ
ࡢ㸪㡢Ⰽ࡟㛵ࡍࡿ⥲ྜⓗ࡞ኌࡢ␗ᖖࢆᣦࡍⅭ[3]㸪ࣁࢫ࣮࢟
ࢆḷၐྜᡂ㡢ኌ࡟௜ຍࡍࡿ஦ࡀᣲࡆࡽࢀࡿ[2]㸬ࡇࡢே㛫ⓗ
࣎࢖ࢫࡶႤኌࡢ୍ᙧែ࡜ぢ࡞ࡏࡿ㸬㢮ఝࡋࡓኌ࡟ࢹࢫ࣎࢖
ࢫ࡞࡝ࡀᣲࡆࡽࢀࡿࡀ㸪ࡇࢀࡣኌᖏ࡜ࡣูࡢჾᐁ࠿ࡽ⏕ࡌ
†1 ᮾ㑥኱Ꮫ Toho University
[a]ࣁࢫ࣮࢟࣎࢖ࢫࡢ௦⾲ⓗ࡞౛࡜ࡋ࡚㸪㺙㺙㺼㺻㺓㺎㺷㺛㺞㺎㺛㺼ࡢ᱓⏣ె♸Ặࡢḷኌ
ࡢ୍㒊ࢆ౑⏝ࡋ࡚࠸ࡿ㸬
[b]ࣁࢫ࣮࢟࣎࢖ࢫ࡛ࡣ࡞࠸㡢ኌ࡜ࡋ࡚➨୍ⴭ⪅⮬㌟ࡢḷኌࢆ౑⏝ࡋ࡚࠸ࡿ㸬
ⓒ 2013 Information Processing Society of Japan
ࡓ㡢ࡀྵࡲࢀࡿኌ࡛࠶ࡿⅭ㸪ࣁࢫ࣮࢟࣎࢖ࢫ࡜ࡣ༊ูࡉࢀ
ࡿࡇ࡜ࡀከ࠸[4],[5]㸬ࡲࡓᅗ 1 ࡢ౛࡟♧ࡍࡼ࠺࡟㸪ࣁࢫ࢟
[a]
࣮࣎࢖ࢫࡢ㡢ኌ ࡣࣁࢫ࣮࢟࣎࢖ࢫ࡛ࡣ࡞࠸㡢ኌ[b]࡜ẚ࡭㸪
࿘Ἴᩘᡂศࡀᗈᖏᇦ࡟ศᩓศᕸࡋ࡚࠸ࡿ஦ࡀ☜ㄆࡉࢀࡿ㸬
1
Vol.2013-MUS-99 No.36
2013/5/12
᝟ሗฎ⌮Ꮫ఍◊✲ሗ࿌
IPSJ SIG Technical Report
ᮏ◊✲࡛ࡣ㸪ࡇࡢ࿘Ἴᩘᡂศࡢᗈᖏᇦศᩓศᕸࢆࣁࢫ࣮࢟
࡚࠸ࡿ㸬ḟࡢᅗ 2 ࡟ࠕṧᕪኚ᥮ἲࠖࢆᇶᮏ࡜ࡍࡿࣁࢫ࣮࢟
࣎࢖ࢫࡢ୺せ࡞≉ᚩ࡜ぢ࡞ࡍ㸬
࣎࢖ࢫ໬ࡢฎ⌮ࣇ࣮ࣟࢆ♧ࡍ㸬
ࣁࢫ࣮࢟࣎࢖ࢫ
୍⯡ࡢேࡢኌ
ᅗ 2 ṧᕪኚ᥮ἲ࡟ᇶ࡙ࡃࣁࢫ࣮࢟࣎࢖ࢫ໬ࡢฎ⌮ࣇ࣮ࣟ
Fig. 2 Processing flow of the husky voice of based on residual
conversion method.
ᮏ◊✲࡛ࡣ㸪ࡇࡢṧᕪኚ᥮ἲ࡟㛵ࡋ࡚௨ୗ࡟㏙࡭ࡿ஧ࡘࡢ
ᡭἲࢆ᳨ウࡍࡿ㸬
ᅗ 1 ࿘Ἴᩘᡂศศᯒ࡟ࡼࡿࣁࢫ࣮࢟࣎࢖ࢫࡢẚ㍑ᅗ
[ c]
Fig. 1 Comparison with a normal voice and husky voice with a
frequency component analysis.
(2) ṧᕪኚ᥮ἲ㸯㸦ࣀ࢖ࢬᙉㄪἲ㸧
ኌᖏ᝟ሗ࡛࠶ࡿṧᕪಙྕࡣ㸪࿘ᮇᛶࡢ࠶ࡿ᣺ᖜࡢ኱ࡁ࠸
ࣃࣝࢫᡂศ࡜㸪ࡑࢀ௨እࡢ᣺ᖜࡢᑠࡉ࠸ࣀ࢖ࢬᡂศ࡟኱ู
ᮏ◊✲࡛ࡣ㸪ḷၐྜᡂ㡢ኌࢆ⥺ᙧண ศᯒ㸦LPC㸧ࡍࡿ
࡛ࡁࡿ㸦ᅗ 3㸧㸬ࣀ࢖ࢬᙉㄪἲࡣ㸪ࡇࡢࣀ࢖ࢬᡂศࡢ᣺ᖜࢆ
஦࡟ࡼࡗ࡚ᚓࡽࢀࡿṧᕪಙྕ࡟ὀ┠ࡋ㸪ṧᕪಙྕ࡟ኚ᥮㸪
ቑᖜࡉࡏࡿ஦࡟ࡼࡾ㸪෌ྜᡂᚋࡢ㡢ኌࡢࣀ࢖ࢬࢆᙉㄪࡉࡏ㸪
ฎ⌮ࢆຍ࠼ࡿ஦࡟ࡼࡗ࡚ࣁࢫ࣮࢟࣎࢖ࢫ໬ࢆ⾜࠺㸬ࡲࡓ㸪
ࣁࢫ࣮࢟࣎࢖ࢫ໬ࢆᅗࡿᡭἲ࡛࠶ࡿ.
㡢㉁㠃㸪ỗ⏝ᛶࢆ⪃៖ࡋࡓ᪉ἲࡢ⪃᱌ࡶ⾜࠺㸬㡢㉁ࡢಖᣢ
ࣃࣝࢫᡂศ࡜ࣀ࢖ࢬᡂศࡢุᐃ࡟ࡣ㸪ṧᕪಙྕෆࡢ᣺ᖜ
࡟ࡣ࢖ࢥࣛ࢖ࢨ࣮ࡢᢏ⾡ࢆᛂ⏝ࡋ㸪ỗ⏝ᛶࡢ㠃࡛ࡣ㸪ࣁࢫ
ࡢ⤯ᑐ್ࢆ౑⏝ࡍࡿ㸬᣺ᖜࡢ್ࡀ⤯ᑐ್ࡢᖹᆒ௨ୖࡢሙྜ㸪
࣮࢟࣎࢖ࢫࡢ⮬ື⏕ᡂࢆヨࡳࡿ㸬⮬ື⏕ᡂࡢᢏ⾡ࡣ㸪ᅵᖌ
ࣃࣝࢫᡂศ࡜ุᐃ㸪ᖹᆒ௨ୗࡢሙྜ㸪ࣀ࢖ࢬᡂศ࡜ุᐃࡍ
▱⾜,ᴙᩗⶶ,∦ᒸⱥᖾ,⸨ᮌᬸஓⴭࠕṧᕪἼᙧศᯒ࡟ࡼࡿႤ
ࡿ㸦ᅗ 4㸧㸬
ࣃࣝࢫᡂศ
ኌࡢᐈほⓗホ౯ࠖᩥ⊩[6]ࢆཧ⪃࡜ࡋ࡚࠸ࡿ㸬ᩥ⊩[6]࡛グ㏙
ࡉࢀࡓࠕႤኌࡢᗘྜ࠸ࢆ ࡿᣦᶆ࡜ࡋ࡚㸪ᝈ⪅ࡢṧᕪἼᙧ
ࡢࣄࢫࢺࢢ࣒ࣛࢆ౑⏝ࡍࡿࠖ࡜࠸࠺᪉ἲࢆࣄࣥࢺ࡜ࡋ࡚㸪
ࣁࢫ࣮࢟࣎࢖ࢫ໬࡟࠾ࡅࡿ㨩ຊᗘࢆィࡿᣦᶆ㸪⮬ື⏕ᡂࡢ
ᡭἲࢆ⪃᱌ࡍࡿ㸬᭱ᚋ࡟㸪ᥦ᱌ࡋࡓ஧✀㢮ࡢࠕṧᕪኚ᥮ࠖ
ἲ㸪࢖ࢥࣛ࢖ࢨ࣮ᛂ⏝ἲ㸪⮬ື⏕ᡂἲ㸪ࡢ㸲ࡘࡢᡭἲࢆ⏝
࠸࡚ྜᡂࡋࡓ㡢ኌ࡛⫈ྲྀᐇ㦂ࢆ⾜࠸㸪ࡑࡢ᭷⏝ᛶ࡟㛵ࡋ࡚
ᐇドࢆヨࡳࡿ㸬
ࣀ࢖ࢬᡂศ
2. ࣁࢫ࣮࢟࣎࢖ࢫ໬ࡢᇶᮏⓗᡭἲ
ᅗ 3 ࣃࣝࢫᡂศ࡜ࣀ࢖ࢬᡂศࡢ࢖࣓࣮ࢪᅗ
Fig. 3
Image of pulse and noise components.
(1) ᇶᮏⓗᡭἲ
⤯ᑐ್
ḷၐ㡢ኌࢆ⥺ᙧண ศᯒ㸦LPC㸧ࡍࡿ࡜㸪⥺ᙧண ಀᩘ
࡜ṧᕪಙྕࡀᚓࡽࢀࡿ[7]㸬ධຊಙྕࡀࡇࡢࡼ࠺࡞ḷၐ㡢ኌ
ࡢሙྜ㸪⥺ᙧண ಀᩘࡣḷモ㒊ศ࣭ኌ㐨᝟ሗ࡛࠶ࡾ㸪ṧᕪ
ಙྕࡣ࣓ࣟࢹ࢕㒊ศ࣭ኌᖏ᝟ሗ࡛࠶ࡿ࡜ぢ࡞ࡏࡿ㸬๓㡯࡛
グ㏙ࡋࡓࡼ࠺࡟㸪ࣁࢫ࣮࢟࣎࢖ࢫࡣኌᖏ࠿ࡽ⏕ࡌࡿ㞧㡢ࢆ
ྵࡴ㡢ኌ࡛࠶ࡿⅭ㸪ᮏ◊✲࡛ࡣኌᖏ᝟ሗ࡛࠶ࡿṧᕪಙྕ࡟
ฎ⌮࣭ኚ᥮ࢆຍ࠼ࡓᚋ㸪ඖࡢ㡢ኌࡢ⥺ᙧண ಀᩘ࡜෌ྜᡂ
ᅗ 4 ࣃࣝࢫᡂศ࡜ࣀ࢖ࢬᡂศࡢุᐃࡢ࢖࣓࣮ࢪᅗ
ࡍࡿࠕṧᕪኚ᥮ἲࠖࢆࣁࢫ࣮࢟࣎࢖ࢫ໬ࡢᇶᮏⓗᡭἲ࡜ࡋ
Fig. 4 Image of pulse and noise components at the time of
[c]࿘Ἴᩘゎᯒ࡟ࡣࣇ࣮ࣜࢯࣇࢺࠕSound Engineࠖࢆ౑⏝ࡋࡓ
judgment.
ⓒ 2013 Information Processing Society of Japan
2
Vol.2013-MUS-99 No.36
2013/5/12
᝟ሗฎ⌮Ꮫ఍◊✲ሗ࿌
IPSJ SIG Technical Report
(3) ṧᕪኚ᥮ἲ㸰㸦㛫ᘬࡁἲ㸧
ኌᖏ᝟ሗ࡛࠶ࡿṧᕪಙྕࡣ㸪ࢥࣥࣆ࣮ࣗࢱୖ࡟࠾࠸࡚
ࢧࣥࣉࣜࣥࢢ࿘Ἴᩘ
[ d]
࡛⾲⌧ࡉࢀࡓࢹࢪࢱࣝࢹ࣮ࢱ࡛࠶
⌧≧࡛ࡣಸ㡢௜㏆ࡢ࿘Ἴᩘᖏᇦ࡟ࣁࢫ࣮࢟࣎࢖ࢫ໬ฎ⌮ࢆ
⾜࠸㸪඼㡢࡜㧗࿘Ἴᩘᇦࡢ㞧㡢㒊ศࢆඖ㡢ࡢࡲࡲಖᣢࡍࡿ
᪉ἲࡀ᭱㐺࡜ุ᩿ࡋ࡚࠸ࡿ㸬
ࡿ㸬ࡇࡢࢧࣥࣉࣝࡢ୰࠿ࡽࣛࣥࢲ࣒࡟ࢧࣥࣉࣝࢆᢳฟࡋ㸪
ࢹ࣮ࢱ್ࢆ 0 ್࡟⨨ࡁ᥮࠼ࡿ㸦ᅗ 5㸧㸪ࡍ࡞ࢃࡕ㛫ᘬࡁࡍ
ࡿ஦࡟ࡼࡗ࡚ṧᕪಙྕࡢ࿘ᮇᛶࢆ୍㒊↓ࡃࡋ㸪␲ఝⓗࣀ
࢖ࢬࢆ⏕ࡌࡉࡏࡿᡭἲࡀࡇࡢ㛫ᘬࡁἲ࡛࠶ࡿ㸬㛫ᘬࡁฎ
⌮ࡢࣦ࢙ࣞࣝㄪᩚࡣ㸪඲యࡢࢹ࣮ࢱᩘࡢఱ㸣ࢆ 0 ್࡟⨨
ࡁ᥮࠼ࡿ㸦㛫ᘬࡁࡍࡿ㸧࠿ࢆᡭືධຊ࡟ࡼࡾᣦᐃࡍࡿ㸬
ᅗ 6
ศ㞳ࡍࡿ࿘Ἴᩘᖏᇦࢆ♧ࡋࡓᅗ
Fig. 6
[ e]
Diagram showing a frequency band to separate.
ᅗ 7
ᅗ 5
࢖ࢥࣛ࢖ࢨ࣮ᛂ⏝ἲࡢฎ⌮ࣇ࣮ࣟ
Fig. 7 Processing flow of the application equalizer method.
㛫ᘬࡁฎ⌮ࡢ࢖࣓࣮ࢪᅗ
Fig. 5 Image of thinning processing.
3.2 ࣁࢫ࣮࢟࣎࢖ࢫࡢ⮬ື⏕ᡂ
3. ṧᕪኚ᥮ࡢᛂ⏝ἲ
㛫ᘬࡁἲࡢᛂ⏝࡜ࡋ࡚㸪㛫ᘬࡁฎ⌮ࡢࣦ࢙ࣞࣝࢆ⮬ືⓗ
ື⏕ᡂἲࠖࢆ⪃᱌ࡋࡓ㸬ࡇࢀࢆḟࡢ㸦1㸧
㸦2㸧
㸦3㸧࡟グ㏙ࡍ
3.1 ࢖ࢥࣛ࢖ࢨ࣮ᢏ⾡ࡢᛂ⏝
ࡿ㸬
࡟Ỵᐃࡍࡿ஦࡟ࡼࡾ㸪ࣁࢫ࣮࢟࣎࢖ࢫࢆ⮬ື⏕ᡂࡍࡿࠕ⮬
ࣁࢫ࣮࢟࣎࢖ࢫ໬࡟࠾࠸࡚᭱ࡶ༴᝹ࡋ࡞ࡅࢀࡤ࡞ࡽ࡞࠸
せ⣲ࡣ㡢㉁ࡢຎ໬࡛࠶ࡿ㸬ࣁࢫ࣮࢟࣎࢖ࢫࡣඖ᮶㸪㞧㡢ࢆ
(1)
ྵࡴ㡢ኌ࡛࠶ࡿⅭ㸪ࣁࢫ࣮࢟࣎࢖ࢫ⏕ᡂࢆヨࡳࡿ࡜࡝࠺ࡋ
⮬ື⏕ᡂࡢ㝿ࡢࣦ࢙ࣞࣝỴᐃᣦᶆ࡟ࡣṧᕪಙྕࡢ᣺ᖜ
࡚ࡶ㸪వィ࡞㞧㡢ࡢᡂศࡀቑ࠼㸪㡢㉁ࡢຎ໬࡟⧅ࡀࡗ࡚ࡋ
ࡢࣄࢫࢺࢢ࣒ࣛࢆ฼⏝ࡍࡿ㸬ࡇࢀࡣ୍⯡ࡢேࡢኌ࡜ẚ࡭㸪
ࡲ࠺㸬ࡋࡓࡀࡗ࡚㸪ࣁࢫ࣮࢟࣎࢖ࢫ໬ࡢ㝿࡟ࡣ㡢㉁ಖᣢࡀ
ࣁࢫ࣮࢟࣎࢖ࢫࡢṧᕪಙྕࡢ᣺ᖜࡢศᕸࡀ኱ࡁࡃ␗࡞ࡿⅭ
᭱ඃඛ஦㡯࡜࡞ࡿ㸬
࡛࠶ࡿ㸬ᮏ◊✲࡛ࡣ᱓⏣ె♸Ặࡢኌࢆ⌮᝿ⓗ࡞ࣁࢫ࣮࢟࣎
㛫ᘬࡁࣦ࢙ࣞࣝỴᐃᣦᶆ
ࡇࡢ㡢㉁ಖᣢ࡟ࡘ࠸࡚ࡣ㸪ṧᕪኚ᥮᫬࡟࢖ࢥࣛ࢖ࢨ࣮ࡢ
࢖ࢫࡢࣔࢹࣝ࡜ࡋ㸪᣺ᖜࡢࣄࢫࢺࢢ࣒ࣛ࡟࠾࠸࡚ࣔࢹࣝ࡜
ᢏ⾡ࢆᛂ⏝ࡍࡿ஦࡟ࡼࡗ୍࡚ᐃࡢຠᯝࡀ☜ㄆ࡛ࡁࡓ㸬ࡇࢀ
ࡢㄗᕪࢆ⟬ฟ㸪ࡇࡢㄗᕪࢆᣦᶆ࡜ࡋ࡚᭱㐺࡞ฎ⌮ࡢࣦ࢙ࣞ
ࡣ୍ᐃࡢ࿘Ἴᩘࡢࡳ࡟ṧᕪኚ᥮ࢆ⾜࠸㸪ཎ㡢ࡢୗᆅࢆ୍㒊
ࣝࢆỴᐃࡋ࡚࠸ࡿ㸬ࡲࡓ㸪ኌ࡟࠾࠸࡚ࡢṧᕪಙྕࡢ᣺ᖜࡢ
ṧࡍࡇ࡜࡟ࡼࡗ࡚㸪㡢㉁ࢆಖᣢࡍࡿ࡜࠸࠺ᡭἲ࡛࠶ࡿ㸦ᅗ
ศᕸࡀ 5D\OHLJK ศᕸ>8@࡟ᚑ࠺࡜௬ᐃࡍࡿ஦࡟ࡼࡗ࡚㸪ࣃ
6㸧㸬௨ᚋ㸪ࡇࡢᡭἲࢆࠕ࢖ࢥࣛ࢖ࢨ࣮ᛂ⏝ἲࠖ࡜࿧ࡪ㸬
࣓࣮ࣛࢱ࣮1 ࡘ㸦௨㝆㸪ㄗᕪ Į㸧࡛ㄗᕪࢆ⾲⌧ࡍࡿ஦ࡀྍ
࢖ࢥࣛ࢖ࢨ࣮ᛂ⏝ἲ࡛ࡣ㸪ࣁࢫ࣮࢟࣎࢖ࢫ໬࡟⏝࠸ࡿཎ
⬟࡜࡞ࡗࡓ㸦ᅗ 8㸧㸬ᐇ㝿࡟ࣄࢫࢺࢢ࣒ࣛ࡜ㄗᕪࢆ⟬ฟࡋࡓ
㡢ࢆ FIR ࣇ࢕ࣝࢱ࡟ࡼࡗ࡚タィࡉࢀࡓࣂࣥࢻࣃࢫཬࡧࢫࢺ
㝿ࡢ⤖ᯝࡀᅗ 9 ࡛࠶ࡿ㸬
ࢵࣉࣇ࢕ࣝࢱࢆ⏝࠸࡚㸪ṧᕪኚ᥮ࢆ⾜࠺ᖏᇦ࡜㸪ࡑࢀ௨እ
ࡢᖏᇦ࡟ศ㞳ࡍࡿ㸬ศ㞳ࡋࡓ୍᪉ࡢ㡢ኌ࡟ṧᕪኚ᥮࣭෌ྜ
ᡂࢆ⾜ࡗࡓᚋ㸪ࡇࡢ஧ࡘࢆ㊊ࡋྜࢃࡏࡿ஦࡟ࡼࡾ㸪୍ᐃ࿘
Ἴᩘࡢࡳ࡟ṧᕪኚ᥮ࢆ⾜࠺ࡇ࡜ࡀྍ⬟࡜࡞ࡿ㸦ᅗ 7㸧㸬ᑦ㸪
[d]ࢧࣥࣉࣜࣥࢢ࿘Ἴᩘࡣ 44100Hz
ⓒ 2013 Information Processing Society of Japan
[e]࿘Ἴᩘゎᯒ࡟ࡣࣇ࣮ࣜࢯࣇࢺࠕWave toneࠖࢆ౑⏝ࡋࡓ
3
Vol.2013-MUS-99 No.36
2013/5/12
᝟ሗฎ⌮Ꮫ఍◊✲ሗ࿌
IPSJ SIG Technical Report
ᅗ 10
ᅗ 8
ṧᕪಙྕࡢࣄࢫࢺࢢ࣒ࣛ࡜ Rayleigh ศᕸ
㛫ᘬࡁᗘࢆ 1%ࡎࡘⴠ࡜ࡋ࡚࠸ࡗࡓ㝿ࡢㄗᕪࡢኚ໬
Fig. 10
Error change when went down by 1% the degree
of thinning.
(3)
ḷၐ༊㛫๭ࡾฟࡋ࡟ࡼࡿฎ⌮ไᗘࡢྥୖ
Fig. 8 Rayleigh distribution and the histogram of the residual
ḷၐࢆ཰㘓ࡋࡓ㡢ኌࢹ࣮ࢱࡢሙྜ㸪ᜥ⥅ࡂࡢ㝿ࡢᴟࡵ࡚
signals.
᣺ᖜࡢᑠࡉ࠸㒊ศࡸ㸪ኌࢆฟࡋ࡚࠸࡞࠸↓㡢ࡢ㒊ศࡀᚲࡎ
Ꮡᅾࡍࡿ㸬ࣁࢫ࣮࢟࣎࢖ࢫ໬ࢆ⾜࠺㝿㸪ࡇࡢ↓㡢㒊ศ࡟ࡲ
࡛ฎ⌮ࢆ᪋ࡍ࡜㸪ணᮇࡏࡠ㞧㡢ࡀ௜ຍࡉࢀ࡚ࡋࡲ࠺࡞࡝࡜
㸦㉥㸧᱓⏣Ặࡢḷၐ㡢ኌ㻌
࠸ࡗࡓ⢭ᗘຎ໬ࡢཎᅉ࡜࡞ࡿ㸬ࡲࡓ㸪➨୍᭱ᑠㄗᕪ๭ࡾฟ
㸦㟷㸧ḷၐྜᡂ㡢ኌ㻌
ࡋࡢไᗘࡶⴠࡕ࡚ࡋࡲ࠺㸬ࡑࡇ࡛㸪⮬ື⏕ᡂἲࡢ㝿࡟ࡣ㸪
᣺ᖜࡢᑠࡉ࠸༊㛫ࡸ↓㡢ࡢ༊㛫࡜㸪ḷၐࡢ༊㛫ࢆศ㞳ࡋ㸪
ḷၐࡢ༊㛫ࡢࡳ࡟ฎ⌮ࢆ⾜࠺஦࡟ࡼࡗ࡚⢭ᗘࡢྥୖࢆᅗࡿ㸬
ḷၐ༊㛫ࡣ㸪ධຊ㡢ኌࡢ᣺ᖜ࡟ࡼࡾ⟬ฟࡍࡿ㸬௵ពࡢ༊
Į 㛫㸦⌧≧࡛ࡣ୍⛊㛫ࢆ 100 ศ๭㸧ࡢ᣺ᖜࡢ⤯ᑐ್ࡢᖹᆒࢆ
ồࡵ㸪ࡇࡢᖹᆒ್ࡀᑠࡉ࠸㒊ศࢆ↓㡢༊㛫࡜ุ᩿ࡋ㸪ḷၐ
༊㛫ࡢ㛤ጞ࣭⤊஢Ⅼ࡜ࡍࡿ㸦ᅗ 11㸧.
ࡇ࠺ࡋ࡚⟬ฟࡋࡓྛḷၐ༊㛫ࡈ࡜࡟ฎ⌮ࢆ⾜࠸㸪᭱⤊ⓗ
ᅗ 9
Fig. 9
ṧᕪಙྕࡢࣄࢫࢺࢢ࣒ࣛ࡜ㄗᕪ⟬ฟ
࡟↓㡢༊㛫ࢆྵࡵ඲࡚ࡢ༊㛫ࢆ⤖ྜࡍࡿ஦࡟ࡼࡾ㸪ฎ⌮ไ
Histograms of the residual signal and error calculation.
ᗘࡢྥୖࢆᅗࡿ㸬
(2)
➨୍᭱ᑠㄗᕪ࡟ࡼࡿ᭱㐺ฎ⌮ࣦ࢙ࣞࣝࡢỴᐃ
1࡛グ㏙ࡋࡓ᣺ᖜࡢࣄࢫࢺࢢ࣒ࣛ࠿ࡽ⟬ฟࡍࡿ㸪ㄗᕪ Į
ࡣ㸪㛫ᘬࡁฎ⌮ࡢࣦ࢙ࣞࣝࢆୖࡆ࡚࠸ࡃ࡜㸪࠶ࡿ୍ᐃࡢᩘ
್ࡲ࡛ࡣୗࡀࡾ㸪ࡑࢀ௨㝆ࡣ୙つ๎࡞ࡩࡿࡲ࠸ࢆࡍࡿ≉ᛶ
ࡀ࠶ࡿ㸦ᅗ 10㸧㸬ࡇࡢㄗᕪࡀ᭱ᑠ࡟࡞ࡿึᅇࡢ࣏࢖ࣥࢺࢆ
ࠕ➨୍᭱ᑠㄗᕪࠖ࡜࿧ࡧ㸪ࡇࡢ➨୍᭱ᑠㄗᕪ࡛ฎ⌮ࣦ࢙ࣞ
ࣝࢆỴᐃࡍࡿ஦࡟ࡼࡾ㸪᣺ᖜࡢࣄࢫࢺࢢ࣒ࣛୖ࡛ࡣࣔࢹࣝ
࡜ྠࡌࣁࢫ࣮࢟࣎࢖ࢫࡢ⏕ᡂࡀྍ⬟࡜࡞ࡿ㸬ᑦ㸪➨୍᭱ᑠ
ㄗᕪࡢ⟬ฟࡣ㸪᳨ドࡢ⤖ᯝ㸪㛫ᘬࡁࣦ࢙ࣞࣝࢆ 1㸣ࡎࡘୖ
ࡆࡿࡓࡧ࡟ุᐃࢆ⾜࠺᪉ἲࡀ᭱㐺࡛࠶ࡿ࡜㸪ุ᫂ࡋࡓ㸬
ᅗ 11 ḷၐ༊㛫ࡢ๭ࡾฟࡋࡢᵝᏊ
Fig. 11
Extraction of singing section.
4. ⫈ྲྀᐇ㦂
3 ❶ࡲ࡛࡟㏙࡭ࡓ஧✀ࡢṧᕪኚ᥮ἲ㸪ཬࡧᛂ⏝ἲ࡟ࡼࡗ
࡚ࣁࢫ࣮࢟࣎࢖ࢫ໬ࡋࡓ㡢ኌ࡟ᑐࡍࡿ⫈ྲྀᐇ㦂ࢆ⾜ࡗࡓ㸬
ᐇ㦂ࡢ᪉ἲࡣ㸪஧ࡘࡢࣁࢫ࣮࢟໬ྜᡂḷၐ㡢ኌࢆ⪺ࡁẚ
ⓒ 2013 Information Processing Society of Japan
4
Vol.2013-MUS-99 No.36
2013/5/12
᝟ሗฎ⌮Ꮫ఍◊✲ሗ࿌
IPSJ SIG Technical Report
࡭㸪࡝ࡕࡽࡀዲࡳ࠿ࢆุ᩿ࡉࡏࡿ஧ⅬႴዲἲࢆ⏝࠸ࡓ㸬⿕
ᯝ࡛ࡣ㸪ࣀ࢖ࢬᙉㄪἲࡶከࡃ⚊ࢆ㞟ࡵࡓࡀ㸪ࡇࡢᡭἲࡣࣃ
㦂⪅ࡣ⏨ᛶ㸶ྡ㸪ዪᛶ㸲ྡࡢィ㸯㸰ྡ࡛࠶ࡿ㸬
࣓࣮ࣛࢱ࣮タᐃࢆᡭືධຊ࡛⾜࠺ᚲせᛶࡀ࠶ࡿⅭ㸪ỗ⏝ᛶ
ࡢ㠃ࢆ⪃៖ࡋ࡚ࡶ㸪࢖ࢥࣛ࢖ࢨ࣮ᛂ⏝ἲ㸩⮬ື⏕ᡂἲࡀ᭱
⫈ྲྀᐇ㦂࡟౑⏝ࡋࡓ㡢ኌࡣ
ࡶ᭷ຠ࡞ࣁࢫ࣮࢟࣎࢖ࢫ໬ࡢᡭẁ࡛࠶ࡿ࡜⪃࠼ࡽࢀࡿ㸬
࣭ࣀ࢖ࢬᙉㄪἲ
࣭㛫ᘬࡁἲ
࣭࢖ࢥࣛ࢖ࢨ࣮ᛂ⏝ἲ
㛫ᘬࡁἲ࡜ࣀ࢖ࢬᙉㄪἲࡢẚ㍑
࣭⮬ື⏕ᡂἲ
12
࣭࢖ࢥࣛ࢖ࢨ࣮ᛂ⏝ἲ㸩⮬ື⏕ᡂἲ
10
[ f]
ࡢ஬ࡘ࡛࠶ࡿ 㸬
ྜィⅬ
8
6
⫈ྲྀᐇ㦂ࡢ㞟ィ⤖ᯝࡣ௨ୗࡢᅗ 12 ࡢࡼ࠺࡟࡞ࡗࡓ㸬
඲య㞟ィ⤖ᯝ
4
ྜィⅬ
2
50
0
㛫ᘬࡁἲ
40
ࣀ࢖ࢬᙉㄪἲ
ᅗ 13 㛫ᘬࡁἲ࡜ࣀ࢖ࢬᙉㄪἲࡢẚ㍑
30
Fig. 13 Comparison of noise enhancement and thinning
20
method.
10
࢖ࢥࣛ࢖ࢨ࣮ᛂ⏝ࡢ᭷↓࡟ࡘ࠸࡚ࡢẚ㍑
0
12
-10
10
ྜィⅬ
8
6
4
2
ᅗ 12 ⫈ྲྀᐇ㦂ࡢ㞟ィ⤖ᯝ
Fig. 12
0
Aggregate results of the listening tests.
↓ࡋ
᭷ࡾ
ᅗ 14 ࢖ࢥࣛ࢖ࢨ࣮ᛂ⏝ࡢ᭷↓࡟ࡘ࠸࡚ࡢẚ㍑
᭱ࡶ⚊ࢆ㞟ࡵࡓࡢࡀ㸪࢖ࢥࣛ࢖ࢨ࣮ᛂ⏝ἲ࡜⮬ື⏕ᡂἲ
Fig. 14 Comparison of the presence or absence of
ࢆ᥃ࡅྜࢃࡏࡓࡶࡢ࡛࠶ࡗࡓ㸬⥆࠸࡚㸪ࣀ࢖ࢬᙉㄪἲࡶ኱
ࡁ࡞⚊ࢆ㞟ࡵࡓ㸬ࡉࡽ࡟ྛᡭἲࢆᑐ࡛ẚ࡭ࡓ⤖ᯝ࡟ὀ┠ࡋ
equalizer application.
࡚࠸ࡃ㸬ᇶᮏⓗ࡞ṧᕪኚ᥮ἲ࡟ࡘ࠸࡚ࡣ㸪ࣀ࢖ࢬᙉㄪἲ࡟
⮬ື⏕ᡂࡢ᭷↓࡟ࡘ࠸࡚ࡢẚ㍑
᭷⏝ᛶࡀ࠶ࡿ࡜☜ㄆࡉࢀࡓ㸦ᅗ 13㸧㸬ࡇࢀࡣ㸪ࣀ࢖ࢬᙉㄪ
ἲࡀཎ㡢࡟ྵࡲࢀࡿ㡢ኌࢆ฼⏝ࡋ࡚࠸ࡿࡓࡵ㸪␲ఝࣀ࢖ࢬ
10
ࢆⓎ⏕ࡉࡏࡿ㛫ᘬࡁἲࡼࡾࡶ㸪ኚ᥮ᚋࡢ㡢ኌࡢ⮬↛ᛶࡀಖ
8
ᣢࡉࢀࡓࡓࡵ࡜᥎ ࡉࢀࡿ㸬ࡑࡢࡓࡵ㸪඲య㞟ィ࡛ࡶࣀ࢖
ࢬᙉㄪἲࡣ኱ࡁ࡞⚊ࢆ㞟ࡵࡓ㸬࢖ࢥࣛ࢖ࢨ࣮ᛂ⏝ἲࡢ᭷↓
6
ࡢ⤖ᯝ㸦ᅗ 14㸧࡛ࡣ㸪᭷ࡾ࡟⚊ࡀ㞟ࡲࡾ㸪࢖ࢥࣛ࢖ࢨ࣮ᛂ
4
⏝ἲ࡟ࡘ࠸࡚᭷⏝ᛶࡀ☜ㄆࡉࢀࡓ㸬⮬ື⏕ᡂἲࡢ᭷↓ࡢ⤖
2
ᯝ㸦ᅗ 15㸧ࡣ㸪ከᑡ᳨ᐃࢆ⦆ࡵ㸪P൑0.15 ࡜ࡋࡓሙྜ࡟㝈
ྜィⅬ
0
ࡾ㸪᭷⏝ᛶࡀ☜ㄆࡉࢀࡓ㸬⤖ᯝⓗ࡟஧ࡘࡢᛂ⏝ἲ㸪ඹ࡟᭷
↓ࡋ
᭷ࡾ
⏝ᛶࡀ࠶ࡾ㸪ࡲࡓ඲య㞟ィ࡛㸪࢖ࢥࣛ࢖ࢨ࣮ᛂ⏝ἲ㸩⮬ື
ᅗ 15 ⮬ື⏕ᡂࡢ᭷↓࡟ࡘ࠸࡚ࡢẚ㍑
⏕ᡂἲࡀ᭱ࡶ⚊ࢆ㞟ࡵࡓ஦࠿ࡽ㸪୧⪅ࡢᛂ⏝ἲ࡟㛵ࡋ࡚㸪
㔜ࡡ᥃ࡅ┦஌ຠᯝࡶᮇᚅ࡛ࡁࡿ࡜ุ᫂ࡋࡓ㸬඲య㞟ィࡢ⤖
Fig. 15
Comparison for the presence of auto-generated.
[ f]ඖ㡢࡜࡞ࡿḷၐྜᡂ㡢ኌࡣ㸪ࣇ࣮ࣜࢯࣇࢺࠕ㹓㹒㸿㹓ࠖࢆ౑⏝ࡋ࡚ไస
ⓒ 2013 Information Processing Society of Japan
5
Vol.2013-MUS-99 No.36
2013/5/12
᝟ሗฎ⌮Ꮫ఍◊✲ሗ࿌
IPSJ SIG Technical Report
ࡇࡇ࡛㸪᭱ࡶ⾲ࢆྲྀࡗࡓ࢖ࢥࣛ࢖ࢨ࣮ᛂ⏝ἲ㸩⮬ື⏕ᡂἲ
ࡢ࿘Ἴᩘゎᯒࡢ⤖ᯝࢆ♧ࡋ࡚࠾ࡃ㸦ᅗ 16㸧㸬⤖ᯝ࠿ࡽࢃ࠿
ㅰ㎡ ◊✲࡟࠶ࡓࡾ㸪༠ຊࢆࡋ࡚࠸ࡓࡔ࠸ࡓ඲࡚ࡢⓙᵝ
ࡿࡼ࠺࡟࿘Ἴᩘᖏᇦࢆᩓᕸࡉࡏࡿ஦࡟ᡂຌࡋ࡚࠸ࡿ㸬
࡟ឤㅰࡢពࢆ⾲ࡍࡿ㸬
ඖ㡢
ฎ⌮ᚋࡢ㡢ኌ
ᅗ 16 ฎ⌮ᚋࡢ࿘Ἴᩘゎᯒ⤖ᯝ
Figure 16
Frequency analysis results after treatment
5. ࠾ࢃࡾ࡟
ཧ⪃ᩥ⊩
[1] 㣩ᒇ㸭Ⳮⵦ㸦࠶ࡵࡸ࣭࠶ࡸࡵ㸧: ḷኌྜᡂࢶ࣮ࣝ㹓㹒㸿㹓ࢧ
࣏࣮ࢺ࣮࣌ࢪ (http://utau2008.web.fc2.com/), Kenchan : UTAU ࡢ
࣮࣒࣮࣍࣌ࢪ http://kenchan22.web.fc2.com/homepageright.html
[2] ᚨ⏣ᜨ୍, ኱ᾆᆂ୍㑻: ⮬ືᏛ⩦࡟ࡼࡾே㛫ࡢࡼ࠺࡟ḷ࠺㡢
ኌྜᡂࢩࢫࢸ࣒ -Sinsy-, ᝟ሗฎ⌮Ꮫ఍◊✲ሗ࿌, (2012.2.3)
[3] 㧗ஂྐ㯢, ⊷⏣ா⏨, ໭ᮧ᝷୍㑻㸸⚟஭ḟ▮: ᐙᗞ་Ꮫ኱඲
໬, ἲ◊, (2010.10.10)
[4] ຍ⸨ᆂ㐀, ఀ⸨ᙲ๎: ࢢࣟ࢘ࣝཬࡧࢫࢡ࣮࣒ࣜḷၐࡢྜᡂ
࡟ྥࡅࡓ㡢㡪ⓗ≉ᚩࡢศᯒ, ᝟ሗฎ⌮Ꮫ఍◊✲ሗ࿌, (2012),
No.14
[5] ຍ⸨ᆂ㐀, ఀ⸨ᙲ๎: 」ᩘḷၐ⪅࡟ࡼࡿࢫࢡ࣮࣒ࣜḷၐ㡢ኌ
ࡢ㡢㡪ⓗ≉ᚩࡢศᯒ, ᪥ᮏ㡢㡪Ꮫ఍⛅Ꮨㅮ₇ㄽᩥ㞟,1-2-9,(2012)
[6] ᅵᖌ▱⾜, ᴙᩗⶶ, ∦ᒸⱥᖾ, ⸨ᮌᬸஓ: ṧᕪἼᙧศᯒ࡟ࡼ
ࡿႤኌࡢᐈほⓗホ౯, ⪥㰯⮫ᗋ 86: 7; 1013㹼1018, (1993.3.13)
[7] 㟷ᮌ┤ྐ㸸ࢹ࢕ࢩࢱ࣭ࣝࢧ࢘ࣥࢻฎ⌮ධ㛛, CQ ฟ∧♫
(2006.04)
[8] L ࣐ࢮࣝ: ☜⋡࣭⤫ィ࣭ࣛࣥࢲ࣒㐣⛬ (బ⸨ᖹඵヂ), 86㹼90
㡫, ᳃໭ฟ∧, ᮾி, (1980)
ᮏ◊✲࡛ࡣ㸪ḷၐྜᡂ㡢ኌࡢࣁࢫ࣮࢟࣎࢖ࢫ໬ࡢ᪉ἲ࡟ࡘ
࠸࡚஧ࡘࡢṧᕪኚ᥮ἲ㸪ཬࡧᛂ⏝ἲࢆᥦ᱌ࡋ㸪ࡑࢀࡽࡢᡭ
ἲ࡛⏕ᡂࡋࡓ㡢ኌ࡟ᑐࡍࡿ⫈ྲྀᐇ㦂ࢆ⾜ࡗࡓ㸬ὀ┠ࡋࡓ࠸
ࡢࡣ㸪ࡇࡢ⫈ྲྀᐇ㦂࡛⮬ື⏕ᡂἲ࡟᭷⏝ᛶࡀ☜ㄆࡉࢀࡓ஦
࡛࠶ࡿ㸬ࡇࡢ⮬ື⏕ᡂἲࡣ㸪᣺ᖜࡢࣄࢫࢺࢢ࣒ࣛ࠿ࡽ⟬ฟ
ࡉࢀࡿㄗᕪ Į ࢆඖ࡟ฎ⌮ࣦ࢙ࣞࣝࢆỴᐃࡋ࡚࠸ࡿ㸬ࡍ࡞ࢃ
ࡕ㸪ࡇࡢㄗᕪ Į ࡣࣁࢫ࣮࢟࣎࢖ࢫࡢ㨩ຊࢆᅗࡿ୍ࡘࡢᣦᶆ
࡛࠶ࡿ࡜⪃࠼ࡽࢀ㸪ࡇࢀࡣ┠ᶆ࡜ࡋ࡚࠸ࡓே㛫ⓗ㨩ຊࢆᅗ
ࡿᣦᶆ࡜ࡶ࡞ࡾᚓࡿྍ⬟ᛶࡀ࠶ࡿ㸬௒ᚋ㸪ࡉࡽ࡟◊✲ࢆ῝
ࡵ㸪ࡇࡢே㛫ⓗ㨩ຊࢆᅗࡿᣦᶆࡢ☜❧ࢆᅗࡾࡓ࠸㸬
ࡲࡓ㸪ᮏ◊✲࡛ࡣ⏨ᛶࡢྜᡂ㡢ኌࡢࡳࢆᑐ㇟࡟ࣁࢫ࣮࢟
࣎࢖ࢫ໬ࢆ⾜ࡗ࡚ࡁࡓⅭ㸪ዪᛶࡢྜᡂ㡢ኌ࡟㛵ࡋ࡚ࡶࣁࢫ
࣮࢟࣎࢖ࢫ໬ࢆྍ⬟࡜ࡍࡿࡢࡀ➨୍ࡢㄢ㢟࡛࠶ࡿ㸬ዪᛶࡢ
ྜᡂ㡢ኌࡢࣁࢫ࣮࢟࣎࢖ࢫ໬࡟㝿ࡋ࡚ࡣ㸪⏨ᛶࡢࡶࡢ࡜ẚ
࡭࿘Ἴᩘࡀ㧗࠸Ⅼࡸ㸪ࣇ࢛࣐ࣝࣥࢺࡢᙧ≧ࡀ␗࡞ࡿⅬࢆ⪃
៖ࡋࡓฎ⌮ࡀᚲせ࡛࠶ࡿ࡜᥎ ࡉࢀࡿ㸬
᭱ᚋ࡟㸪ྜᡂ㡢ኌࡢࣁࢫ࣮࢟࣎࢖ࢫ໬ࡢ฼⏝᪉ἲ࡟ࡘ࠸
࡚グ㏙ࡍࡿ㸬ࡇࢀࡣࡶࡕࢁࢇ㸪㡢ᴦไస࡬ࡢ㉳⏝ࡀ➨୍࡜
ࡋ࡚ᣲࡆࡽࢀࡿࡀ㸪ࡑࢀ௨እ࡟ࡶ㸪౛࠼ࡤ㸪㆙ഛࢩࢫࢸ࣒
ࡢ㆙࿌㡢ኌ࡞࡝㸪┦ᡭ࡟ᚰ⌮ⓗᙳ㡪ࢆ୚࠼ࡓ࠸ሙྜ࡟ࡶ㸪
ࡇࡢࡼ࠺࡞≉Ṧ࡞ྜᡂ㡢ኌࡀຠᯝࢆⓎ᥹ࡍࡿྍ⬟ᛶࡀ࠶ࡿ㸬
ࡍ࡛࡟஺㏻ᶵ㛵࡞࡝ࡢ࢔ࢼ࢘ࣥࢫ࡛㉳⏝ࡉࢀ࡚࠸ࡿࡼ࠺࡟㸪
௒ᚋ㸪♫఍ࡢᵝࠎ࡞ሙ㠃࡛ࡢྜᡂ㡢ኌࡢ౑⏝ࡀቑຍࡍࡿ஦
ࡀண᝿ࡉࢀࡿ㸬ࡑࡢࡓࡵ㸪ྜᡂ㡢ኌ࡟ே㛫ⓗ㨩ຊࡸ≉Ṧ࡞
ຠᯝࢆ௜ຍࡍࡿᢏ⾡ࡢᚲせᛶࡣᚲࡎ㸪㧗ࡲࡗ࡚࠸ࡃ࡛࠶ࢁ
࠺㸬ࡑ࠺࡞ࡗࡓ᫬࡟㸪ᮏ◊✲ࡢᡂᯝࡀ㈉⊩࡛ࡁࡿࡼ࠺㸪ࡇ
ࢀ࠿ࡽࡶ◊✲ࢆ⥆ࡅ㸪ࡉࡽ࡞ࡿⓎぢ㸪㐍ᒎ࡟⮳ࢀࢀࡤⰋ࠸
࡜⪃࠼࡚࠸ࡿ㸬
ⓒ 2013 Information Processing Society of Japan
6
Fly UP