Icyitegererezo
Icyitegererezo cyerekana imyigishirize yigisha imashini injyana yimvugo, injyana, amajwi, guhangayika, no kwihuta kugendera hejuru yamagambo.
Incamake
It is what separates a flat robotic voice from one that sounds genuinely human.
Kwibira cyane
Prosody numuziki wururimi: kuzamuka no kugwa kwijwi (intonasiyo), amajwi areshya (igihe), amajwi (ingufu), hamwe nibishimangira. Ibi bimenyetso bitwara ibisobanuro amagambo yonyine atabikora, yerekana ibibazo nibisobanuro, gusebanya, byihutirwa, cyangwa ni irihe jambo ari ngombwa. Sisitemu igezweho-y-imvugo yerekana icyitegererezo hamwe numuyoboro uhuza ibinyabuzima uteganya imiterere, igihe cya fonema, nimbaraga zituruka kumyandiko. Tacotron 2 yize byinshi muribi bitaziguye binyuze mubitekerezo, mugihe FastSpeech 2 yabisobanuye neza mu guhanura igihe, ikibuga, nimbaraga nkibintu bitandukanye byitozwa. Ibyiza byiza biterwa nurwego sisitemu idashobora kubona kuruhuka rwonyine, niyo mpamvu abanyamideli bagenda bakoresha interuro zikikije ndetse n'amajwi yerekana kugirango bashireho ijwi ryiza.
Ubushishozi
Ikibanza gikurikiranwa nkinshuro zifatizo (F0) yijwi, igipimo cyijwi ryinyeganyeza. Moderi nka FastSpeech 2 yongeramo adaptate itandukanya guhanura F0, ingufu, hamwe na foneme igihe cyayo nkinzira zitandukanye, hanyuma ugashyiraho decoder ya spekrogramu. Kuberako inyandiko idasobanura neza prosody (interuro imwe ifite ibyasomwe byinshi byemewe), iki nikibazo cyumuntu umwe kuri benshi, sisitemu rero ikoresha ibyihuta bitandukanye cyangwa kodegisi zerekana guhitamo ikintu runaka aho kugereranya muri monotone.
Ingaruka z'Ingamba
Kugera no kugera
Itezimbere kugerwaho binyuze mu kwandukura, kuvuga, no guhuza amajwi.
Igiciro na bije
Amatsinda yibitangazamakuru arashobora kohereza amajwi yihuse hamwe na bije nto.
Umuvuduko n'igipimo
Sisitemu ireba abakiriya irashobora gutunganya imikoranire ivugwa murwego runini.
Ejo hazaza h'icyitegererezo
Ubushinjacyaha bugenda bwerekeza kumiterere yibice byose hamwe nibiganiro, bityo uwatanze inkuru arashobora kubaka impagarara cyangwa ikiganiro gishobora guhuza numukoresha. Imvugo nini nicyitegererezo cyururimi rwiga prosody hamwe nubusobanuro, bigafasha ipfundo rishobora kugenzurwa kugirango ushimangire, amarangamutima, nuburyo bwo kuvuga ukoresheje amabwiriza asobanutse neza. Tegereza ibitabo byamajwi, dubbing, hamwe nabafasha batandukana muburyo busanzwe, wongeyeho kugenzura neza ibitagenda neza no guhumeka kugirango wambuke igice cya nyuma cyikibaya kidasanzwe.
Gushyira mu bikorwa Isi
Sisitemu yo kuvuga amajwi ya audiobook itandukanya ikibanza no kwihuta kuburyo ibice byumvikana neza kuruta monotone
Abafasha ba Virtual bazamura intonasiyo kurangiza yego / ntakibazo kuburyo byumvikana nkikibazo
Ibikoresho bya firime na videwo bihuza gushimangira nigitekerezo cyo gutanga umukinnyi wambere
Soma abasomyi kugirango bagerweho bashimangira amagambo yingenzi kuburyo abakoresha impumyi basobanukirwa interuro isobanura byihuse
Ingaruka & Kurinda
Gukoresha nabi amajwi no kwigira ibyago byiyongera mugihe uruhushya rubuze.
Ukuri kurashobora kugabanuka hejuru yimvugo, imvugo, cyangwa urusaku rwibidukikije.
Amajwi yubukorikori arashobora kwibeshya kumvugo yukuri nta kirango gisobanutse.
Igishushanyo mbonera
Shaka uruhushya rusobanutse rwo gufata amajwi, gukoroniza, no gukoresha.
Ikizamini cyiza mubiganiro bitandukanye hamwe nuburyo bwimbere.
Sobanura igihe umuntu agomba gusuzuma cyangwa kwemeza ibisubizo.
Andika amajwi yubukorikori kandi ugumane inyandiko zerekana kubazwa.
Komeza Ubushakashatsi
Free newsletter
Get the daily AI briefing
Three verified AI stories every weekday morning, written in plain English. Free forever, no ads.
One email each weekday. Unsubscribe in one click. We never sell or share your address.
Test yourself
Take the Prosody Modeling quiz
Instant feedback on every answer, and a shareable certificate with a verifiable ID once you pass a course.
Support free AI education. AI Understanding is a 501(c)(3) nonprofit — no ads, no paywall, ever. Make a donation
Ubuyobozi bukurikira
Icyitegererezo cya XLNet
Ibibazo bikunze kubazwa
What is Prosody Modeling?
Icyitegererezo cyerekana imyigishirize yigisha imashini injyana yimvugo, injyana, amajwi, guhangayika, no kwihuta kugendera hejuru yamagambo. Nicyo gitandukanya ijwi rya robo iringaniye nijwi ryumvikana nkabantu.
Nibihe bintu biranga bisobanura neza imvugo mu mvugo?
Prosody bivuga imico isumba iyindi mvugo, intonasiyo (ikibanza), injyana nigihe bimara, guhangayika, nimbaraga (ijwi rirenga), bigendera hejuru yamagambo.
Mu kwerekana imideli, inshuro zifatizo (F0) zerekana iki?
F0 ninshuro yibanze yijwi, igipimo cyinyeganyeza cyijwi ryijwi, twumva nkikibanza cyangwa injyana yijwi.
Nigute FastSpeech 2 yatezimbere igenzura ryimikorere ya moderi zabanje?
FastSpeech 2 yazanye adaptateur itandukanya igihe, igihe, n'imbaraga ku buryo bweruye, bitanga uburyo butaziguye kandi butajegajega ku miburanishirize kuruta kwitabwaho gusa.
Ni ukubera iki guhanura imanza ziva mu nyandiko byonyine bifatwa nk'ikibazo kuri benshi?
Amagambo amwe arashobora gutangwa muburyo butabarika bitewe nubushake n'amarangamutima, moderi rero igomba guhitamo mubisomwa byinshi byemewe aho kubara igisubizo kimwe.
Nibihe bimenyetso byerekana neza ko interuro yicyongereza ivugwa ari yego / ntakibazo?
Yego / ntakibazo mucyongereza mubisanzwe birangirana no kuzamuka kwintonasiyo, umurongo wa prosodic ubatandukanya namagambo niyo magambo amwe.