Umbhalo ukuya kuSpeech Abathunywa be-AI I-Music Umbhalo we-audio I-Voice Cloning Uyilo lwesandi I-Marketplace I-More
Ingxelo ye Bug / Isicelo se Feature

Umenzi Wengoma we-AI

Yenza iingoma ezibhaliweyo ukusuka kwiinkcazelo zombhalo. Chaza uhlobo, ubume, okanye uhlobo kwaye uvumele i-AI ihlanganise kuwe.

Ubhaliso simahla
Asikho nasiphi na isandi se-TTS kwisiNgesi sakho. Nceda uncedo lwethu ukongeza isandi sakho! Intengiso yelizwi lakho
Bhalisa to generate music

Ukudibanisa umculo wakho...

Ukudala umculo kunokuthatha ixesha elide. Isandi sakho siza kubonakala kwi- Imbali yokudala xa ilungile.
Ingoma Izalisekisiwe
0:00 0:00
Layishela phantsi egronjiweyo

Imodeli

ACE- Inyathelo v1 yi 3. 5B- parameter diffusion model evelisa i 48 kHz stereo music ukusuka kumbhalo kwiilwimi ezili-17. Apache 2. 0 licensed — completely royalty- free.

Ixesha lokusebenza

5s 30s 30s
Ixesha elide lisebenzisa uphawu oluninzi kwaye lithatha ixesha elide ukudala.

Iingoma (ekhethiweyo)

Yishiye engenanto iingoma zesiculo. ACE- Step ixhasa ulwimi 17.

Iqondo lobushushu

Ubugcisa / Ukungaqhelekanga 1.0
0. 5 (Iqhotyoshelweyo) 1. 5 (Ezimeleyo)

Iingcebiso Ezicelayo

  • Cacisa uhlobo:
  • Iinkqubo zokuzichaza:
  • Ichaza ubume: \
  • Cwangcisa ixesha:
  • Uhlobo lobhekiso:

Iinketho ze projekti

Cofa ukuze uzame:

Indlela i-AI Music Generation isebenza ngayo

Yenza umculo wesiqalo kwinyathelo lesithathu esilula. Akukho lwazi lomculo lufunekayo.

Inyathelo 1

Inkcazelo

Write a text prompt describing the music you want. Mention genre, mood, instruments, tempo, and style. Use the quick- select tags to build your prompt faster.

Inyathelo lesi-2

I-AI Composes

Imodeli ye-AI ichaza umyalezo wakho kwaye ivelise umculo wokuqala. Uqhubekeko olukhawulezayo lwe-GPU luqinisekisa iziphumo ezikhawulezayo, ngokuqhelekileyo 10-30 imizuzwana ngokuxhomekeke kwixesha lokusebenza.

Inyathelo lesi-3

Layisha ezantsi

Imboniselo yangaphambili yengoma yakho eyenziweyo ngomdlali wesandi ofakwe ngaphakathi. Layisha ezantsi kwifomati ye-WAV yomgangatho ophezulu. Yenza kwakhona ngemimiselo eyahlukeneyo de ufumane ingoma egqibeleleyo.

Iimodeli ze-AI Music

Uthelekiso lweemodeli ze-AI ezifumanekayo zokwakha iingoma. Imodeli nganye inezinto ezinamandla, iimpawu, kunye neendlela zemveliso.

ACE-Step v1

Efumanekayo
Umbhekisi phambili:
StepFun & ACE Studio
Ilayisensi:
Apache 2.0

3. 5B- parameter diffusion transformer for full songs. Apache 2. 0 weights with no gated dependencies. Optional lyrics in 17 languages. Generates a 4- minute song in roughly 20 seconds on an A100.

Iingoma Ezipheleleyo Iingoma Iilwimi 48kHz Stereo

YuE

Iyafika
Umbhekisi phambili:
Tencent
Ilayisensi:
Apache 2.0

Imodeli yokudala umculo opheleleyo okwaziyo ukwenza iingoma ezipheleleyo ezinemiculo, iingoma, kunye noncedo lwezixhobo ezisuka kwimiyalezo ekhawulezayo.

Iingoma Ezipheleleyo Iingoma + Iingoma zengoma IiNtaka Ezininzi 44.1kHz Audio

DiffRhythm

Iyafika
Umbhekisi phambili:
ASLP@NPU
Ilayisensi:
Apache 2.0

Imodeli yokudala iingoma ezipheleleyo ezisekelwe kwi-diffusion. Ivelisa iingoma ezipheleleyo ezithembekileyo kakhulu usebenzisa ubume besakhiwo esingaqhelekanga.

I-Diffusion-Based Iingoma Zobude Obupheleleyo High Fidelity 48kHz Audio

Iinkqubo zokuvelisa iingoma

Qala ngokukhululekileyo, uphucule xa ufuna okuninzi

I-akhawunti Ekhululekileyo
  • Iiclip ezili-30 zesekondi
  • ACE- Inyathelo v1 (Apache 2. 0)
  • Iingoma ezikhethiweyo kwiilwimi ezili-17
  • 15,000 iimpawu kwi-signup
  • Stereo WAV
Yenza phezulu
Ethandwa Kakhulu
Isiqalisi / Esincinci
  • Iiclip ezili-30 zesekondi
  • Uphawu oluphezulu lwenyanga
  • Igama leqela lenjongo ethile
  • Ukusetyenziswa korhwebo ngaphandle kwelayisensi
Ubhaliso
I-Pro
  • Iiclip ezili-100 zemizuzu
  • Uhlobo lweqela
  • Uhlobo lweqela
  • Ufikelelo lwe-REST API
Yenza phezulu

Imibuzo ebuzwa rhoqo

Uhlobo lwemiculo ye-AI isebenzisa imodeli yokufunda ngokunzulu ukwenza iingoma zokuqala ukusuka kwiinkcazo zombhalo. Chaza uhlobo, ubume, izixhobo, kunye nexesha ofuna, kwaye i-AI iqulethe inxalenye ekhethekileyo yengoma. Akukho lwazi lwemiculo lufunekayo.

TTS.ai inikezela nge Bark (Suno, ilayisensi ye MIT) yengoma, iziphumo zesandi, kunye nokwenziwa kwengoma. Iimodeli ezizayo ziquka i YuE (Apache 2. 0) yokwenziwa kwengoma epheleleyo ngeengoma kunye ne DiffRhythm yengoma epheleleyo ukuya kumaxesha angama- 285.

Ewe. ACE-Step yi-Apache 2.0 — zombini ikhowudi yemodeli kunye nesisindo — ngoko ke yonke imiculo eyenziwe nge-TTS.ai iyinikwe wena ukuyisebenzisa ngokurhweba. Ungasebenzisa imiculo eyenziwe kwi-YouTube videos, iipodcasts, imidlalo, izikhumbuzo, nakweyiphi na iprojekthi yorhwebo ngaphandle kweerhafu okanye ukunikezelwa.

I-Bark ivelisa iziqhoboshi ukuya kuthi ga kwimizuzwana engama-30. Iimodeli ezizayo ezinjenge-YuE ne-DiffRhythm ziya kuxhasa iingoma ezipheleleyo ukuya kuthi ga kwimizuzwana engama-285 (imalunga nemizuzu emi-5).

Ewe. Chaza uhlobo olufunayo (i-rock, electronic, jazz, classical, lo- fi, ambient), ubume (obunethemba, obunesiyezi, obunengqondo, obuzolileyo), izixhobo (i-piano, i-guitar, synth, i-drums), nexesha kumyalezo okhawulezayo. Imodeli ichaza inkcazelo yakho ukuvelisa ingoma ehambelanayo.

Ewe. ACE- Step v1 ixhasa indawo yengoma ekhethiweyo. Nceda unikezele ngengoma yakho (imalunga ne 4, 000 iimpawu) kunye nesimo sokuchaza kwaye imodeli izakwenza ingoma epheleleyo enengoma. Iingoma zixhaswa kwiilwimi ezili-17 kubandakanya isiNgesi, isiSpanish, isiFrentshi, isiJamani, isiTshayina, isiJaphani, kunye nezinye.

I-music eyenziweyo iphuma kwi-48 kHz WAV formati yomgangatho ophezulu. I-API ixhasa i-OGG output. Ungaguqula kwi-MP3, FLAC, okanye M4A usebenzisa isixhobo sethu esikhululekileyo sokutshintsha isandi.

Iimodeli ze-AI music zisebenzisa i-transformer architectures eziqeqeshiweyo kwi-data ye-music. Imodeli ifunda iipateni ze-melody, harmoni, rhythm, ne-timbre, emva koko ivelisa isandi esitsha ngokujika-jika ngokuzenzekelayo ukusuka kwi-description yombhalo. Ifana nendlela iimodyuli ze-language ezivelisa ngayo umbhalo, kodwa kwi-audio tokens.

Ewe. Umculo owenziwe yi-AI ukusuka kwi-TTS.ai yimixholo yokuqala eyenziwe ngesicelo. Akuyi kubangela izilumkiso ze-ID yeMixholo kuba ayiyikopi yomculo okhoyo. Uyakwazi ukwenza imali ngeevidiyo usebenzisa lo mculo ngaphandle kweengxaki ze-copyright.

I-Bark ifumaneka kwinqanaba eliqhelekileyo (i-2 credits ngexesha ngalinye). Ii-akhawunti ezikhululekileyo zifumana i-15 credits xa zibhalisa. Iinkqubo ezihlawulwayo ziqala kwi- $9/inyanga nge-500 credits.

Ewe. I-REST API yethu ibonisa /api/v1/music/ kubasebenzi abaqinisekisiweyo. Thumela isiqu se-JSON esicela, ixesha lokuphila, kunye nengoma ekhethiweyo; ufumane umsebenzi we-UUID ukuvota kwi-URL yesandi eyenziweyo. Ifumaneka kuzo zonke iinkqubo ezihlawulwayo.
5.0/5 (1)

Yintoni esinokuyilungisa? Ulwazi lwakho olufunyenweyo lunceda silungise iingxaki.

Qala ukudala umculo nge-AI

Yenza umculo wokuqala ukusuka kwiinkcazelo zombhalo. Bhalisa ngokukhululekileyo kwaye ufumane amakhadi angama-50 ukuqalisa ukudweba.

Ubhaliso simahla Bonisa ixabiso