AI Briefing#text-to-speech
The latest AI and developer news about #text-to-speech, with the original source and a short summary.
The latest AI and developer news about #text-to-speech, with the original source and a short summary.
The platform offers 70,000 licensed Brazilian Portuguese titles and free access to Machado de Assis's Dom Casmurro narrated by Fábio Porchat.
The update reduces voice roughness and break-up while maintaining ~300 ms first-audio latency on laptop CPUs.
Eleven v4 introduces contextual Audio Tags like [pause] and [long pause] to replace fixed SSML break tags for natural timing.
Microsoft launched MAI-Transcribe-2-Streaming, which ranks No. 1 on Artificial Analysis, along with two new voice models, MAI-Voice-2.1 and MAI-Voice-2.1-Flash, designed for low-latency conversational AI.
Eleven v4 supports phoneme rules and inline IPA but does not support SSML phoneme tags.
Audio Tags allow users to direct emotional performance in Eleven v4, which ranks first in the Artificial Analysis Speech Arena with 1319 Elo.
The new leaderboard uses objective metrics like WER and RTFx to evaluate open-source TTS models, reducing assessment time from weeks to hours.
Creator+ subscribers receive 3x credits for Eleven v4 until October 12.
Eleven v4 Turbo delivers ~100 ms median inference latency and supports over 90 languages within the ElevenAgents platform.
Eleven v4 Turbo achieves a median inference latency of ~100ms, enabling low-latency conversational agents.
The new model uses Glow-TTS architecture and targets edge devices and CPU inference.
The plan offers university students in the US, Canada, EU27, Australia, and UK free access to ElevenCreative, ElevenAgents, ElevenAPI, and ElevenReader Ultra.
Google has unveiled the Gemini 3.8 Flash TTS series, which enables voice generation and cloning via natural language prompts.
Two models supporting over 100 languages and speaker control are now available via the unified API.
StepAudio 3 Gen achieves state-of-the-art performance in TTS and multimodal audio generation using an RVQ token-based discrete generation architecture.
Havells has launched an Agentic AI Voice Mode in the Havells One app, powered by ElevenLabs technology, enabling users to control home appliances in 8 languages.
Hunar AI integrated ElevenLabs' TTS to increase onboarding speed for Indian field workers by 90% and improve retention rates by 35%.
Prosody, comprising rhythm, intonation, and stress patterns, is a core element of TTS models that determines the human-likeness of AI voices.
Halo NeuroAI has released Sopro V2 Turbo, a 120M-parameter open-source voice cloning model that natively supports European Portuguese and enables real-time streaming on laptop CPUs.
FireRedTeam released FireRedAudio, a 9B parameter unified audio language model that decouples understanding and generation.
Neural TTS is driving market growth by replacing traditional methods with deep learning-based natural speech synthesis.
Fish Audio's speech synthesis and recognition models are available for free on Vercel AI Gateway until September 18.
audio.cpp version 0.6 has been released, supporting five new model families including MiniMax-H3 and a WebUI.
Voice cloning APIs generate multilingual synthetic speech using the voice of a specific individual.