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 release includes three model sizes ranging from 7.2M to 82M parameters, all supporting 24 kHz audio output and CPU-only execution via int8 ONNX weights.
Paid APIs provide turnkey enterprise-grade speech synthesis with SLAs, while open source models offer data sovereignty and full architectural control at the cost of compute and engineering overhead.
The release includes Mini, Micro, and Atom variants, with the 34.31M parameter Mini model supporting reference voice tuning and 48 kHz enhancement on CPU and CUDA.
The model supports 43 languages and runs on phones, laptops, and browsers using Apache 2.0 licensed weights.
The guide defines seven API architectures and details ElevenAPI's use of single-use tokens expiring after 15 minutes for secure client-side connections.
The project uses a 21.7 MB iSTFTNet vocoder for mobile devices and a 27.9 MB HiFi-GAN for desktops to achieve 0.75x real-time factor synthesis.
Gradium's Phonon TTS, available in 31m and 40m sizes, tops the Hugging Face Open TTS Leaderboard's CPU streaming benchmark for time-to-first audio.
The guide outlines the three-step process of voice synthesis, distinguishes between instant and professional cloning, and details detection methods for AI-generated audio.
A practical guide details how to build an AI voice assistant using Eleven v4 Turbo and Scribe v2 Realtime to achieve low-latency, high-accuracy interactions.
The partnership covers six audio categories including speech-to-text, streaming TTS, realtime voice agents, music, sound effects, and multi-voice dialogue, accessible via a single API key.
This article explains the mechanics of text-to-speech technology, detailing the neural synthesis pipeline and introducing ElevenLabs' new Eleven v4 model.
Destined AI releases Hurston-1, a 3B parameter Apache-2.0 model fine-tuned from Maya1 to generate accented English speech with native telephone-band output.
The system uses BookLLM to profile characters with 97% accuracy in English and 94% in Spanish, enabling distinct and consistent voices across full-length books.
The model runs at 72x real-time on CPU and 1120x on GPU, though it may slur difficult words or long numbers.
The guide distinguishes voice over from dubbing and acting, and outlines a five-step process for generating audio using ElevenCreative Studio.
The 1.7B parameter model supports zero-shot voice cloning in 20+ languages and expressive emphasis control without requiring a GPU.
The 0.5B parameter model supports voice design, 7 languages, and emotion/style control.
The guide covers Python and TypeScript setup, model selection for 90+ languages, and troubleshooting for common API errors.
The new English models support voice cloning and are designed to run on CPU, allowing GPUs to remain dedicated to LLM inference.
The inference stack targets 10+ backends including CPUs, GPUs, and NPUs, with most components written in Rust.
The release includes a LiveKit-based Voice Agent SDK and speech models priced from $0.0035 per minute for transcription and $13 per 1M characters for TTS.
The model achieves ~1.8% average WER on Open TTS Leaderboard English sets and supports 51 style tags plus zero-shot cloning.
ElevenLabs supports SSML tags like <break> and <phoneme> on older models, while Eleven v3 and v4 replace them with natural language Audio Tags.
The model is currently in preview and offers an 80% discount for 30 days through November 12.