VibeVoice

Speaker 1

Emiselweyo IsiNgesi Neutral VibeVoice

{igama} yi neutral AI yelizwi elinamandla eVibeVoice umbhalo-kwi-speech model. Eli standard-level ilizwi lithetha IsiNgesi kwaye linika studio-quality speech synthesis. Nge Ixesha elifutshane unikezelo lwesantya kunye nomgangatho womgangatho we 5/5, Speaker 1 ulungele podcasts, dialogues, long-form narration, multi-speaker content. I-VibeVoice injini iphuhliswe ngu Microsoft under the MIT license, iyenza ikhuseleke kwimisebenzi yentengiso. Iinkqubo eziphambili ziquka: {iimpawu}.

Akukho manqaku

VibeVoiceUlwazi lwemodeli

Imodeli VibeVoice
Umbhekisi phambili Microsoft
Umgangatho
Isantya I-Fixed
Ilayisensi MIT
Ukuklona Ayifumaneki
I-Tier Eqhelekileyo (2x uphawu)
Iiparamitha 1.5B
Uyilo lwezindlu LLM + DAC
Uqeqesho lwe Data 100000 iiyure
Iminyaka 2025

Iinkqubo ezilungileyo zokusetyenziswa Speaker 1

Iinkqubo ezicetyiswayo ezisekelwe kwiimpawu zalo msindo

Iincwadi ezinesandi & Uxwebhu

Sebenzisa i {igama} ukuchaza imixholo yefom ende nge-prosody eqhelekileyo ne-expression.

Ividiyo

Yongeza ukuthetha okuzimeleyo kwiividiyo zeYouTube, iintengiso, kunye nemixholo yemidiya yoluntu.

Iinkqubo & Zokufikelela

Ukwenziwa ngokukhawuleza kwenza le lizwi lilungele iinkqubo zexesha elibonakalayo, abafundi bekhusi, kunye neezixhobo zokufikelela.

Ipodcasts & Ukusasazwa

Imveliso elungileyo yestudio elungele iipodcasts, umculo, kunye nokusasazwa okuzimeleyo.

I-More VibeVoice IiNkokheli

Ezinye iingoma zemodeli efanayo ye-TTS

Speaker 1 (Chinese)

IsiTshayina Neutral

Speaker 2

IsiNgesi Neutral

Speaker 2 (Chinese)

IsiTshayina Neutral

Speaker 3

IsiNgesi Neutral

Speaker 4

IsiNgesi Neutral

Imibuzo ebuzwa rhoqo

I VibeVoice ye Microsoft ivela ngeendlela ezimbini: imodeli ye 1. 5B yezinto eziqulethe iifomu ezide (zifike kwimizuzu engama- 90, abavakalisi aba- 4) kunye nexesha elipheleleyo lemodeli ye 0. 5B yosasazo nge ~200ms yokuqala yokungasebenzi kakuhle kwesandi. I 1. 5B ilungele iipodcasts kunye neencwadi zesandi ezinesandi esizinzileyo ngaphezulu kweendawo ezide. Qaphela: I Microsoft isuse ikhowudi ye TTS kwindawo yokugcina kwaye ivelise isandi equlethe i- AI ekwazi ukuvakala.

IVibeVoice yaphuhliswa yiMicrosoft kwaye ikhutshwa phantsi kwe MIT (ukukhangela-ku-intent) ilayisenisi, evumela ukusetyenziswa korhwebo lwesandi esiveliswe.

I-VibeVoice inkxaso ye 1 ulwimi: IsiNgesi.

I-VibeVoice ikwinqanaba eliphezulu — ii-credits ezi-4 ngabasebenzi aba-1,000. Ungajonga ngaphambili nayiphi na i-VibeVoice yesandi ngaphandle kokuvelisa isandi esipheleleyo.

I VibeVoice inesantya sokwakha esiphakathi. Ukwakha kuthatha imizuzwana emibini exhomekeke kubude bombhalo.

I-VibeVoice ifunyenwe 5/5 kwi-audio quality kwi-TTS.ai. Inikezela nge-studio-grade, ukuthetha okunjalo nomuntu.

Hayi, i-VibeVoice isebenzisa iseti emiselweyo yelizwi elingaphakathi. Ukwenza ilizwi lifana, zama iimodeli ezifana ne-CosyVoice 2, GPT-SoVITS, okanye i-Chatterbox.

Ewe, iVibeVoice icetyiswa ngokukodwa kwiipodcasts, iincwadi zesandi, imixholo emide yomthumeli-omninzi. Umthumeli-omninzi wayo, ukuya kwimizuzu engama-90, ukwenziwa kwepodcast kwenza ukuba ibe lukhetho olulungileyo kule meko yokusetyenziswa.

Ewe, iVibeVoice ilayisensiwe phantsi kweMIT (ukukhangela-ku-intent), evumela ukusetyenziswa korhwebo. Isandi esiveliswe ngeVibeVoice sesandi singasetyenziswa kwividiyo, iipodcasts, iinkqubo, imidlalo, nakweyiphi na enye iprojekthi yorhwebo.

Ewe, zonke iingoma kwi-TTS.ai zisebenzisa iimodyuli ze-open-source ezilayisensiweyo ngentengiso (MIT, Apache 2.0). Isandi esiveliswe yiyo yakho ukuyisebenzisa kwividiyo, iipodcasts, iiapps, imidlalo, nakweyiphi na enye inkqubo yentengiso.

Thumela isicelo se POST ku /api/v1/tts/ ngegama lemodeli ne ID yesandi. Bona iphepha lethu le-API Documentation ngemizekelo yekhowudi kwi-Python, JavaScript, Go, kunye ne-cURL.

Ewe, nqakraza iqhosha lokudlala kweli phepha ukuva isampuli. Ungabhala umbhalo oqhelekileyo kwiphepha lombhalo ukuya kukuthetha kwaye wenze ukujonga kuqala simahla ngelizwi elithile.

Zama Speaker 1 Ngoku

Bhala nawuphi na umbhalo uze uyiva ithetha ngu Speaker 1. Ifumaneka simahla.