VITS Ƙarshe-zuwa-Ƙarshe Maganar Magana
VITS samfuri ne na rubutu-zuwa-magana wanda ke juya rubutu kai tsaye zuwa ga danyen sautin raƙuman sauti a cikin tsarin horarwa guda ɗaya, yana tsallake bututun matakai biyu na yau da kullun.
Dubawa
By combining variational inference with adversarial training, it produces remarkably natural, expressive speech.
Zurfafa nutsewa
VITS (Bambancin Bambanci tare da koyo na gaba don ƙarshen-zuwa-ƙarshen Rubutu-zuwa-Magana), wanda Kim, Kong, da Son suka gabatar a cikin 2021, ya haɗa ra'ayoyi guda uku waɗanda tsofaffin tsarin ke kiyaye su. Maɓallin bambance-bambancen autoencoder (VAE) yana koyan latent wakilci na magana, daidaita magudanar ruwa yana sa wannan rarrabawar latent ya zama mai sauƙi don ɗaukar cikakkun bayanai masu kyau, kuma mai nuna wariya irin na GAN yana tura tsarin motsin da aka samar zuwa ga gaskiya. Mahimmanci, VITS tana horar da ƙirar sauti da vocoder tare maimakon a matsayin matakai biyu, yana kawar da rashin daidaituwa wanda ke lalata inganci lokacin da aka horar da na'urori daban. Hakanan yana gabatar da tsinkaya na tsawon lokaci, don haka za'a iya yin magana iri ɗaya tare da mabambantan ƙarar sautin yanayi kowane lokaci.
Fahimtar Fasaha
VITS yana magance matsalar daidaitawa tare da Monoton Alignment Search (MAS), wanda ke nemo mafi kyawun taswira tsakanin alamun rubutu da firam ɗin sauti yayin horo ba tare da masu daidaitawa na waje ba. Ana lissafta na baya na VAE daga ainihin sautin, yayin da aka canza yanayin da aka riga aka tsara akan rubutu ta hanyar daidaita magudanar ruwa don dacewa da shi. A cikin ƙididdiga, kuna yin samfuri daga rubutun da ya gabata kuma ku yanke hukunci kai tsaye zuwa tsarin igiyar ruwa, don haka babu keɓantaccen mel-spectrogram kuma babu keɓaɓɓen vocoder da ake buƙata.
Dabarun Tasiri
Shiga ku isa
Yana inganta samun dama ta hanyar rubutu, ba da labari, da mu'amalar murya.
Kudin da kasafin kuɗi
Ƙungiyoyin kafofin watsa labaru na iya jigilar sauti mai gogewa cikin sauri tare da ƙaramin kasafin kuɗi.
Gudu da sikelin
Tsarin fuskantar abokin ciniki na iya aiwatar da hulɗar magana a mafi girman ma'auni.
Makomar VITS Ƙarshen-zuwa-Ƙarshen Maganar Magana
VITS ta haifar da dangin magada waɗanda suka mamaye tushen tushen TTS. VITS2 ya sauƙaƙa tsarin gine-gine da haɓaka dabi'a, yayin da YourTTS da Coqui XTTS da aka yi amfani da su da yawa sun tsawaita hanyar yin amfani da sautin murya mai harbi da kuma yaruka da yawa. Yi tsammanin ci gaba da aiki akan bambance-bambancen na'urori masu sauƙi, ainihin lokaci akan na'urori, mafi kyawun ɗaukar hoto don harsuna masu ƙarancin albarkatu, da tsauraran iko akan motsin rai da salon magana, tunda ƙirar ƙarshe zuwa ƙarshen abu ne mai ban sha'awa, ingantaccen tushe don ginawa.
Aiwatar da Gaskiyar Duniya
Coqui TTS yana jigilar samfuran tushen tushen VITS waɗanda masu haɓakawa da kyau-tunani don haɗa takamaiman muryar mai ba da labari don littattafan mai jiwuwa.
Mataimakan murya na buɗe tushen akan kayan aikin Rasberi Pi-class suna amfani da ƙaramin ƙirar VITS don cikakken fitowar magana ta layi.
Ka'idodin koyon harshe suna haifar da misalan lafuzza na yanayi ta amfani da bambance-bambancen VITS na harsuna da yawa kamar YourTTS.
Studios na wasan Indie suna haɗa layukan tattaunawa na NPC daban-daban, dogaro da ma'aunin tsinkaya na tsawon lokaci don rhythm mara amfani.
Hatsari & Tsare-tsare
Rashin amfani da murya da haɗarin kwaikwaya yana ƙaruwa lokacin da aka rasa izini.
Daidaituwa na iya faɗuwa cikin lafuzza, yaruka, ko mahalli masu hayaniya.
Ana iya kuskuren sauti na roba don ingantacciyar magana ba tare da bayyananniyar lakabi ba.
Taswirar Hanya
Sami tabbataccen izini don ɗaukar murya, cloning, da sake amfani.
Gwajin ingantattun masu magana daban-daban da yanayin baya.
Ƙayyade lokacin da dole ne ɗan adam ya duba ko ya amince da abubuwan da aka fitar.
Yi lakabin sauti na roba da kuma adana bayanan da aka tabbatar don yin lissafi.
Ci gaba da Bincike
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 VITS End-to-End Speech Synthesis 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
Jagora na gaba
Maganar Maganar Tafiya
Tambayoyin da ake yawan yi
What is VITS End-to-End Speech Synthesis?
VITS samfuri ne na rubutu-zuwa-magana wanda ke juya rubutu kai tsaye zuwa ga danyen sautin raƙuman sauti a cikin tsarin horarwa guda ɗaya, yana tsallake bututun matakai biyu na yau da kullun. Ta hanyar haɗa bambance-bambancen ra'ayi tare da horo na gaba, yana haifar da yanayi mai ban mamaki, magana mai bayyanawa.
Menene ma'anar gajarta VITS take nufi?
VITS tana tsaye don Bambancin Bambance-bambance tare da koyo na gaba don Rubutu-zuwa-Magana daga ƙarshen zuwa-ƙarshe, yana nuna ainihin abubuwan da ke cikin sa guda uku.
Menene babban bambancin gine-gine tsakanin VITS da bututun TTS na gargajiya?
VITS ƙarshen-zuwa-ƙarshe: yana koyon rubutu-zuwa-waveform a cikin tsari ɗaya, yana guje wa rashin daidaituwar ƙirar sauti da murya daban.
Ta yaya VITS ke koyon jeri tsakanin rubutu da firam ɗin sauti?
VITS tana amfani da Neman Daidaita Juyin Halitta don gano mafi kyawun daidaitawar rubutu-zuwa-firam a ciki, ba tare da aligner na waje da ake buƙata ba.
Me yasa VITS ta haɗa da tsinkayar tsawon lokaci?
Hasashen stochastic na tsawon lokaci yana ba da damar yin magana iri ɗaya tare da raye-raye na yanayi daban-daban, tare da guje wa lebur, ƙirar mutum-mutumi.
Wane bangare ne ke tura fitowar VITS zuwa ga sauti mai sauti na gaske?
VITS tana amfani da horon adawa (GAN), inda mai nuna wariya ke yanke hukunci ko sifofin raƙuman ruwa suna da gaske, suna haɓaka ingancin fahimta.