Dellu ci xibaar yi
ProduitAI Understanding

Google genne na model TTS 3.8 Flash

Google DeepMind genne na Gemini 3.8 Flash TTS ak Flash-Lite TTS, xeetu bind-ci-kaddu yu bees yu jàppandi ci Google AI Studio ak jaaraleko ci Gemini API ak sos SIDyn baat buñ personaalise bu baax.

4 min readRead the primary source
Source-provided image accompanying Google launches Gemini 3.8 Flash TTS models
Këyitu xët bu njëkkSource biñ enregistre
Siiwalkat
deepmind.google
Lëkkalekaayu cosaan
deepmind.googlehttps://deepmind.google/blog/say-hello-to-gemini-38-text-to-speech/
Xeetu balluwaay
Këyitu njëkk - ab yëgle ofisel, këyit, dosiye, wala xëtu pàrti bu njëkk bi ñuy jàng ci saasi.
KontekstXam lii ci 60 seconde

Tambalil fii

Term yu am solo

API (interfaasu prograam aplikaasioŋ)
Benn anam buñ tëral ngir benn sistem losisel mëna yónnee ay laaj ak jot tontu ci beneen sistem.
Figrana
Dugalal siñaal buñu mëna gis ci biir mbind wala media bu IA defar suko defee ñu mëna ko xàmmee ni masin moo ko defar.
Référence
Test buñ yamale wala ensemble done yuñ jëfandikoo ngir natt ak méngale liggéeyu model bi.
Nattal sa boppModèlu IA leeral quiz

Lu xew

Google DeepMind xamlena genne ñaari xeetu bind-ci-kaddu yu bees, ZXQAI1QXZ 3.8 Flash TTS ak Gemini 3.8 Flash-Lite TTS. Modèle yooyu am nañu ci saasi ci Google AI Studio ak jaaraleko ci API Gemini, boole ci platform yu melni Agora, LiveKit, ak Vercel. Li ñuy génne dafay yaatal famiy Gemini Audio, yokk ci kàttan ngir defar kàdduy personage yuñ personaaliseer ak di doxal waxtaanu seen ak doxal bu jaar yoon ci joxe gi.

Google DeepMind 3.8 Flash TTS ak Gemini 3.8 Flash-Lite TTS, muy leen wax ni ñooy model yiy gëna fësal audio ci famiy Gemini. Koom-koom bi dafa wax ni xeetu model yooyu dañuy soppi defarum baat joge ci preset yu taxaw dem ci studio kreatif buy yëngu, loolu mooy tax ñu mëna defar baat yu personage yi ak yoon wi ñuy jaar ngir waxtaan ci seen.

Modèle yi ñu ngi jàppandi tay ci Google AI Studio, fu ñuy liggéey ni barabu liggéeyu jëmmal baat. Jëfandikukat yi mën nañu laaj dàntite vocal yu bees ci noonu wala ñu niroole seen baat, ba noppi jëfandikoo editëru senaario bu am ñaari kàddu ngir mëna joxe liiñ par liiñ. Accès bi itam jaaraleko ci Gemini API, loolu mooy tax platform defarkat yu melni Agora, LiveKit, Pipecat, ak Vercel mëna tabax ak jëfandikoo jaar-jaari defar kàddu.

Google dafa wax ni Gemini 3.8 Flash TTS moo jël #1 barab bi ci Hume AI's Voice Design ak poñ bu 71.4 te jiitu ci modelaasioŋ ak poñ bu 60.8. Ñari xeetu model yi deñoo am #1 ak #2 ci Index bu Kalite bu Hume AI. Ci jàngat yiñ def ci tànneefi nit ñi ci Voice Arena, neena ñu xeetu nit ñi ñoo ëppu doole ci lakk yi gëna am solo ci àdduna bi lu ci melni Japon, Portugais bu Bresil, Vietnamien, Araab bu bees, Españool bu Mexico, ak Hindi, ak ndimmbalu lu ëppu 100 lakk.

Li ñuy génne dafa amaale ay pexe yu jëm ci kaaraange ngir niroole baat, di sàkku ci jëfandikukat yi ñu joxe ndigalu enregistrement bu bawoo ci boroom baat bi méngoo ak kàddukat bi ñuy jëfandikoo balaa ñuy mëna sos baat. Rax ci dolli, bépp clip audio bu model yooyu defar dañu ko def watermark ak SynthID, watermark buñu dul gis, ñu defaree ko ngir mu mëna gis kàddu yi IA defar ngir moytu joxe xibaar bu dul dëgg.

Ay leeral ci cosaan: deepmind.google

Lu tax mu am solo

Ubbite bii dafay màndargaal coppite bu am solo ci defar baat IA joge ci preset yu fiix dem ci defar audio buy yëngu, buñ mëna mengale. Suñu may developpeur yi ñu sos dàntite vocal yu bees wala ñu niroole ay baat yuñ xool ndax dañu nangu, model yi dañuy ubbi pexe yu bees ci localisation media yi, agent conversationnel yi, ak defar ëmbiit yu bees. Dugalug SynthID dafay saafara jafe-jafe yiy gëna yokk ci joxe xibaar bu dul dëgg bu IA defar, di joxe kaaraange xarala ngir ëmbiit li leer.

Dugalug xeetu model yooyu dafay may developpeur yi ak entreprise yi jumtukaay yu ñuy mëna defaree ay audio yu gëna riis, gëna fësal seen xalaat te duñu yéem ci preset yu baat yu fiks. Mën-mën boobu dafa am solo lool ci liggéey yi soxla aksan yu bari ci diwaan yi ngir lokalisasioŋ media yi wala kàddu màrk yu méngoo ngir ndawu waxtaan yi.

Partenariat bi ñu am ak sosiete yu melni Figma, HeyGen, Linguana, Wondercraft, 99.co, ak Ollang dafay wane ni mën nañu ko jëfandikoo ci saasi ngir gaawlu dubbing àdduna bi ak dooleel kàdduy waxtaan ci anam wu yaatu. Loolu dafay tekki ni dañu wara dugg ci boole kàttanu TTS yu xarañ yi ci liggéeyu sos lu bees ak liggéeyu liggéeyukaay.

Bëgg-bëgg bi ñu am ci xool ndax nangu nañu baat bi ñuy baamtu ak jëfandikoo SynthID dafay wax ci jafe-jafe yi gëna am solo ci wàllu ethique ak kaaraange ci defar audio IA. Kaaraange yooyu liñuy fexe mooy aar dàntite talentu baat bi ak fexe ba ëmbiit li leer, te loolu dafay gëna am solo ginaaw bi mejaa yi IA defar di gëna bari.

Interactive Mechanism

Mekanism buy weccoo xalaat: naka lay doxee

Saytu xarala yu bees yi ci ginaaw yokkute bii ci anam wu weccoo xalaat.

Agent Lifecycle Stage:
1
User Intent & Planning: "Audit customer refund request #4092 and settle payment."
2
Tool Calling: Emits structured JSON call crm_get_transaction(id='4092').
3
Guardrail & Verification:🛡️ Paused: High-value action requires human operator sign-off.
4
Final Settlement: Refund recorded, email receipt dispatched, and audit log stored.
Core takeaway: An AI agent is not just a language model—it is a closed loop of planning, tool invocation, and environment feedback. Production systems require self-healing retries and strict human approval guardrails.
Saytu konsept buy weccoo xalaat+10 Points
AI Models Explained Quiz

In AI, what are a model's "parameters"?

Li nga wara seetaan ci topp

Xool bu baax ni sosiete yiy jàppalante ak ñoom lu melni Figma ak HeyGen di jëfandikoo xeetu mbir yooyu ngir dubbing àdduna bi yépp ak lokalisasioŋ ci mejaa yi. Seetal jàngat yu moom seen bopp ci wàllu kaaraange baat ak njariñu SynthID ci gis kàddu yi IA defar. Rax ci dolli, xoolal ni defarkat yi di jëfandikoo editëru senaario bu am ñaari kàddu yi ci Google AI Studio ngir nettali audio yu jafee xam.

Xoolal ni liggéeyukaay yi ñuy jàppale di boolee xeetu produit yooyu ci seeni produit, rawatina ci wàllu dubbing global ak localisation media, ngir jàngat performance yi ci àdduna bi ak ni jëfandikukat yi di nangu.

Saytu jàngat yi moom seen bopp ci ñeneen ñi ci anam yi ñuy nangu baat bi ak ni ñu mëna gisee filigran SynthID, ndax loolu dafay am solo ngir mëna wóolu kaaraange ak njubte gi ci xarala yi.

Toppal yokkuteg editëru senaario bu am ñaari kàddukat ci Google AI Studio, ndax man-man bii dafay wane interface bu bees ngir teg waxtaan bu IA defar bu mëna indi coppite ci ni defarkat yi di jegee nettali audio.

Gid ak quiz yu ci méngoo

Model IA leeral nañu koJikko yu AIAgent IANatt li nga xam — natt quiz IA bu amul faydaSeetal benn baat IA ci sunu glossaire
Gis nga lii am njariñ?