Imvugo Kamere hamwe na Diffusion Yatinze TTS
NaturalSpeech ni umurongo wa Microsoft Ubushakashatsi bwa TTS bugamije ireme ryimvugo yumuntu, hamwe na verisiyo yanyuma ikwirakwizwa ryihishwa kugirango itange amajwi akungahaye, karemano.
Incamake
It shows how diffusion models, famous for images, can produce expressive, controllable audio.
Kwibira cyane
Umwimerere wa NaturalSpeech (2022) niyo sisitemu ya mbere yavuzwe ko yageze ku rwego rw’umuntu ku gipimo cya LJSpeech, igenzurwa n’abayumva batabashaga kubivuga mu buryo bwizewe. Yakoresheje autoencoder itandukanye hamwe na priors ihuye neza kugirango yuzuze icyuho kiri hagati yimyitozo numwanzuro. NaturalSpeech 2 noneho yakoresheje uburyo bwo gukwirakwiza bwihishwa: imvugo yashizwemo na codec ya majwi ya necode mu bice bikomeza byihishwa, kandi moderi yo gukwirakwiza yiga kubyara ayo mabanga avuye mu nyandiko, bigatuma ijwi rikomeye rya zeru riva mu gihe gito. NaturalSpeech 3 yatangije gukwirakwiza ibintu, gutandukanya imvugo mubintu bitandukanijwe nkibirimo, prosody, timbre, na acoustic ibisobanuro, kuburyo buriwese ashobora kwiganwa no kugenzurwa yigenga kubwizerwa no guhinduka.
Ubushishozi
Gukwirakwiza bidatinze bikora wongeyeho urusaku rwihishwa rwerekana imvugo no gutoza umuyoboro kugirango uhindure urusaku intambwe ku yindi. Aho kwerekana imiterere mbisi cyangwa spekrogramu yuzuye, NaturalSpeech 2 yerekana codec yihishe, iri munsi-yoroheje kandi yoroshye kwerekana. Gutondekanya kumyandiko hamwe nijwi ryerekana ijwi riyobora ikwirakwizwa ryinyuma, bityo ibyanyuma byatoranijwe byinjira mumvugo ihuye nibisabwa hamwe nibiranga abavuga.
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'Imvugo Kamere no Gutinda Bitinze TTS
Diffusion ishingiye kandi yibintu bya TTS yerekeza kumajwi adasanzwe gusa ariko arashobora kugenzurwa neza, kureka abayikoresha bagahindura timbre, amarangamutima, hamwe nibyiza nkibiganiro byigenga. Tegereza icyitegererezo cyihuse ukoresheje disillation hamwe nintambwe nke zo gukwirakwizwa, gukomera kwa zeru kurasa kuva amasegonda y'amajwi, no guhuza cyane hamwe nururimi runini rwo gutanga imvugo. Iterambere kandi rishimangira gukenera ibimenyetso byamazi no kwemererwa kurinda, kubera ko kwizerwa kwinshi bitera ingaruka mbi zo gukoresha nabi.
Gushyira mu bikorwa Isi
Sitidiyo ya Dubbing yakusanyije ijwi ryumukinnyi kuva sample ngufi kugirango yerekane firime, ukoresheje NaturalSpeech yuburyo bwa 2-zero-shoti.
Amajwi ya Audiobook atanga ibisobanuro byurwego rwabantu abumva baharanira gutandukanya impano yijwi ryukuri.
Ibikoresho bigerwaho byongera ijwi ryumuntu uhereye kumajwi ya kera kubantu bataye ijambo.
Ibikoresho byo kurema ibintu reka reka abanditsi bigenga bahindure timbre na prosody, bakoresheje NaturalSpeech 3 yibiranga.
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 NaturalSpeech and Latent Diffusion TTS 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
Ijwi rihamye Gutandukana
Ibibazo bikunze kubazwa
What is NaturalSpeech and Latent Diffusion TTS?
NaturalSpeech ni umurongo wa Microsoft Ubushakashatsi bwa TTS bugamije ireme ryimvugo yumuntu, hamwe na verisiyo yanyuma ikwirakwizwa ryihishwa kugirango itange amajwi akungahaye, karemano. Irerekana uburyo ikwirakwizwa rya moderi, rizwi cyane kumashusho, rishobora kubyara amajwi yerekana, agenzurwa.
Ni ubuhe buryo bukomeye bwavuzwe na NaturalSpeech y'umwimerere (2022) bwagezweho?
NaturalSpeech yavuzwe nka sisitemu ya mbere igeze ku rwego rw’umuntu kuri LJSpeech, abayumva ntibashobora kuyitandukanya neza n’imvugo nyayo.
Niki NaturalSpeech 2 yerekana ikwirakwizwa ryihishe mubyukuri?
NaturalSpeech 2 ikwirakwira hejuru ya codec yihishe, iroroshye kandi yoroshye kuyigereranya kuruta imivumba mbisi, hanyuma ikayihindura amajwi.
Nigute moderi yo gukwirakwiza itanga amakuru kurwego rwo hejuru?
Diffusion yongeraho urusaku mugihe cyamahugurwa kandi ikiga kuyihindura, ikabyara ingero mukwerekana buhoro buhoro urusaku rudasanzwe.
Ni ubuhe bushobozi bw'ingenzi butangwa bwihishwa butanga NaturalSpeech 2?
Mugutondekanya kumajwi magufi, NaturalSpeech 2 irashobora guhuza ijwi ryumuvugizi itigeze itozwa neza.
Ni ikihe gitekerezo nyamukuru cya NaturalSpeech 3 'gukwirakwiza ibintu'?
Gukwirakwiza ibintu bitandukanya ibirimo, prosody, timbre, na acoustic birambuye kuburyo buri kiranga gishobora kugereranywa no kugenzurwa ukundi.