@multimodalart: they extracted only the audio bit of LTX-2.3, fine-tuned for TTS task and achieved SOTA TTS emotional control??? try it…
Summary
A fine-tuned version of the LTX-2.3 model's audio component achieves state-of-the-art emotional control in text-to-speech, now available as a Hugging Face Space called DramaBox by ResembleAI.
View Cached Full Text
Cached at: 05/15/26, 11:09 PM
they extracted only the audio bit of LTX-2.3, fine-tuned for TTS task and achieved SOTA TTS emotional control???
try it for yourself. So far I’m very impressed!
https://t.co/A8qD9dc78b
DramaBox - a Hugging Face Space by ResembleAI
Source: https://huggingface.co/spaces/ResembleAI/Dramabox Fetching metadata from the HF Docker repository...
Similar Articles
DramaBox - Most Expressive Voice model ever based on LTX 2.3
DramaBox is a highly expressive voice model based on LTX 2.3, released by Resemble AI with open-source code and models on GitHub and Hugging Face.
ResembleAI/Dramabox
Dramabox is an expressive text-to-speech model by Resemble AI that uses prompt-driven control for speaker identity, emotion, and delivery, with optional voice cloning via a 10-second reference. Built on the LTX-2.3 audio diffusion transformer, it is open-sourced on Hugging Face.
DramaBox: An Open-Weight TTS Model Built Around Stage Directions
DramaBox is an open-weight TTS model fine-tuned from LTX-2.3 that uses stage directions as prompts to generate expressive speech, with optional voice cloning from a 10-second sample.
Kitten-TTS 2: New Open-source SOTA 1.58 bit TTS (CPU-only with voice cloning, multilingual, emotion control)
KittenML 开源发布了 Kitten-TTS 2，这是全球首个 1.58-bit（三值化权重）TTS 模型，1.7B 参数，可在纯 CPU 上实时推理，支持少于 5 秒参考音频的零样本语音克隆、20+ 语言多语种支持以及情感与强调控制。
huggingface/speech-to-speech
Hugging Face's speech-to-speech is an open-source, modular voice-agent pipeline (VAD->STT->LLM->TTS) with an OpenAI Realtime-compatible WebSocket API, supporting swappable components and local or hosted models.