TTS wā tūturu

Ka puta te kupu-ki-te-kōrero me te whakateretanga o te reo tuatahi. I hangaia mō ngā māngai reo me ngā taupānga ora.

Kua mahia e tātau Whakapā atu i tō tou reo

Huinga

Ka puta te rerenga
0/5,000 Huānga ~0.3s oro tuatahi

Ka taea te whakarerekē i te reo

Kāhua ā-waiho anake.

Whakarewa ā-rau

Ka tirohia te pūwhitinga kia ine ai i te whakaterenga-pūrongo tuatahi

Pāpāho

Ka whakapāpāhohia ēnei wāhanga oro i te wā e rere ana rātau.

0:00
Ko te wāhanga tuatahi:
Ko te katoa o nga hautanga: 0
Wā katoa:

He pēhea te mahi o te Streaming TTS

1. Whakahaua te kupu

POST te kupu ki te /v1/tts/stream/ hei tono Hōu-Tuhiu-whakataua.

2. Ka hangaia te tauira

Ko te Kokoro e whakaroa ana i te kupu me te whakaputa oro-i-te-tauira i runga i te GPU.

3. Te Raupapa o te Reo

Ka tae mai ngā wāhanga WAV whakawaeheretia-pūtake64 i runga i te SSE, ā, ka tīmata te whakatere i te wā kotahi.

4. Heoi mai

Ka whakarongo te kaiwhakaari i te tīmatanga o te rerenga i raro i te rua, ahakoa i runga i ngā tāurunga roa.

Ka whakamahia ngā take

I reira ka kati te wāhanga-rua i ngā wheako hōu.

Kaikōrero reo

Ko ngā tāngata kōrero e whakautu ana i te tere pēnei i te tangata.

Whakapāpāho ora

Ka whakamāoritia me te whakarerekē i tētahi rerenga i te wā tūturu me te kore whakawhāititanga.

Kīhei

Ko te tauwhitinga NPC e urupare ana ki ngā kōwhiringa o te kaitapere, kāore he VO i whakaritea i mua.

Kitenga

Ko ngā kaimātai mata me ngā utauta āwhina e tīmata ana ki te kōrero i te wā e pāpāho ana te kaiwhakaari.

Whakahau tōtika TTS

Ka tīmata wātea, whakapiki ina hiahiatia he nui ake

Waihoki
  • Kokoro te rerenga (tauira wātea)
  • 500 ngā pūāhua i ia whakatupuranga
  • 10 ngā rerenga wātea/rohe mō ia kaimahi ā-ringa
  • Ko te whakateretanga o te audio tuatahi o te tuarua
  • SSE te rerenga puta noa i HTTPS
Ko te tino rongonui
Kāreti
  • 15,000 ngā tohu i te whakaingoatanga
  • 5,000 ngā pūāhua i ia rerenga
  • Paepae API mō te uru ki te papatono
  • Ko te hītori whakawhanaketanga
  • Kāore he tepe rerenga i ia rā
Ka tāuru i te wātea
Ka taea
  • MOSS-TTS-Realtime (ina ora)
  • 100,000 ngā pūāhua i ia rerenga
  • GPU whakahau tuatahi
  • Kaikōrero reo + Twilio taupānga
  • Ko ngā tepe o te mokatere tiketike ake
Whakahauhau

E pā ana ngā pātai

Ko te rerenga kupu-ki-rongo-rerenga-

Ka whakaputaina e te TTS noa iho te faila oro katoa i mua i te hokinga o tētahi mea — ka tūpono koe, kātahi ka whakarongo i te rerenga katoa i te wā kotahi. Ka whakamahia e te TTS te wā tūturu ngā Hītori-Tērā-Tērā (SSE) hei whakawhitiwhiti i ngā wāhanga oro poto i te wā e whakaputaina ana e te tauira. Ka whakarongo te kaiwhakamaori i te tīmatanga o te rerenga tata tonu, ahakoa i ngā tāuru roa.

Ko te Kokoro te pūtake pūmau — e whakaoho ana i te oro 100x tere ake i te wā tūturu i runga i tētahi GPU o nāianei. E whakauru ana tātau i te MOSS-TTS-Realtime hei whirinoa whai hua nui ake; Ka taea e ngā kaiwhakaari te kōwhiria tono ina puta mai.

Ko te āhua o te āhua o te oro-whakahaere tuatahi i runga i te Kokoro ko te 300-800ms i runga i tētahi hononga tūmatanui. Ka whakahaeretia te haerenga paetata i muri i taua wā. Ka puta te pātū i te wā-ki-te-rongo-whakahaere-tūturu i te UI kia kitea ai e koe te roa o ia tono.

Ko ngā māngai reo e urupare ana i ngā kōrero, te whakarerekētanga ora mō ngā pāpāho whakawhitiwhiti, ngā NPC kēmu whakawhitiwhiti, ngā kaimātai āhei ki te kōrero i te wā e pāpāho ana te kaiwhakaari, me ētahi atu taupānga e rua, e toru rānei ngā takiwā mō te oro, ka pōturi te āhua.

He. POST ki te https://api.tts.ai/v1/tts/stream/ me te tinana ōrite ki te /v1/tts/ te wāhi mutunga. He rerenga SSE te urupare o ngā wāhanga WAV whakawaeheretia-pūtake64. E tautoko ana te taumata wātea i ngā whakatupuranga 10 i ia rā mō ia kaimahi whakawātea; ka whiwhi ngā kaimahi whakamātautau i te āheitanga pūāhua-whakahaere katoa.

Ko te Kokoro e whakamahi ana i ngā oro i whakaakona i mua, ā, kāore e tārua. MOSS-TTS-Realtime (ina whakaurua) e tautoko ana i te tārua reo kore-ki te tohutoro 3-sekondi. Mō te tārua reo katoa i tēnei rā, ka whakamahia te /tuhi-ki-te-kōrero/ pae noa me te Chatterbox, GPT-SoVITS rānei — kāore ēnei i te āhei ki te whakawhitiwhiti, engari ka whakaputaina ngā oro whakawhanake.

Ko te utu āhuarite ōrite ki te wāhi mutunga TTS noa. Ko te Kokoro he taumata-whāiti (1x utu). Ka haere a MOSS-TTS-Realtime ki te taumata paerewa (2x utu) ina whakahohetia. Kāore te tikanga rerenga i te tāpiri i tētahi utu tāpiri.

He — e hono ana te pito mutunga o te rerenga ki tētahi Twilio webukapu reo hei whakawātea i te oro ora ki roto i tētahi whakapāho. Kua mahi kē tēnei a tātau papatono māngai reo mō IVR me te whakapāhotanga. Ko te āhua o te whakawhāititanga o te mutunga ki te mutunga i runga i tētahi whakapāho, ko te 1-2 ngā takirua, tae atu ki te urupare STT me LLM.

Mēnā ka ngaro tētahi wāhanga i roto i te whakawhitinga o tōtou whatunga, ka ākiritia te kaiwhakaari rerenga i mua, kaua e tū. Mō ngā taupānga kāore e taea te whakamātau i ngā wāhanga, ka hoki ki te wāhi mutunga kore-rerenga noa iho, ki te pūrere 500ms rānei o te oro i mua i te tīmatanga o te whakaari.
5.0/5 (1)

He aha ka taea e tātau te whakapai ake? Ka āwhina tātau ki te whakaoti i ngā raruraru.

Ka whakawhitiwhiti kōrero i te wā tūturu

Whakawhanake mo te 10 o ngā whakatupuranga tuatahia rā. Ka whakataki hei whakakōkī i te āheitanga pūāhua katoa me te āheitanga API.