Input an audio and a photo into the AI model, you can make the characters in the picture sing and speak, let Audrey Hepburn sing "Shangchun Mountain", the pottery lady speaks English RAP, and Einstein speaks Chinese jokes. Not long ago, this model, called EMO, became popular at home and abroad because of a paper by Ali Tongyi Lab. Today, EMO officially launched the Tongyi App, which is free for all users.
Open the Tongyi App, enter the "National Stage" channel, and you can come to the EMO product page "National Singing". Choose any template from songs, memes, and memes, upload a portrait photo, and EMO can be combined into a video.
Tongyi App has launched more than 80 EMO templates in the first batch, including the hit songs "Shangchun Mountain" and "Wolf Disco", etc., as well as the Internet hot memes "Bowl Chicken" and "Back to Hand". At present, Tongyi App does not allow user-defined audio, and users can only select the preset audio of the App to generate videos.
EMO is an AI model developed by Tongyi Laboratory, and the Talking Head technology behind it is currently a hot AIGC field. Before EMO, the Talking Head technology required 3D modeling of human faces, heads, or body parts, and Tongyi Lab was the first in the industry to propose a weak control design, which can drive portraits to speak without modeling, which not only reduces the cost of video generation, but also greatly improves the quality of video generation.
EMO learns and encodes the ability of humans to express emotions, can match audio content with characters' expressions and mouth shapes, and can also reflect the tone characteristics and emotional colors of audio to the characters' micro-expressions.
At the end of February, Tongyi Lab published related papers, and EMO became one of the most concerned AI models after SORA. In the future, EMO technology is expected to be applied to scenarios such as digital humans, digital education, film and television production, virtual companionship, and e-commerce live streaming.
The phenomenon app "National Dance King" has set off a boom of "Terracotta Warriors Dancing Subject 3" and "Golden Retriever Dancing" on social platforms, and the model behind the National Dance King, Animate Anyone, is also from Tongyi Lab. Recently, Tongyi APP has also successively launched free and practical functions such as ultra-long document analysis, AI coding assistant, and AI meeting assistant. The Tongyi model is becoming a super AI assistant for more and more users.