Synthesia has made significant progress in developing its platform by introducing the feature of Expressive Avatars.
The emerging artificial intelligence company supported by Nvidia revealed digitally generated avatars capable of conveying human emotions using user text inputs.
These AI-powered avatars move a step forward from traditional digital avatars by adjusting their tone, facial expressions, and body language based on the context of the content they present.
The Expressive Avatars technology in Synthesia is highly realistic and helps clients create AI-powered video clips to attract the target audience.
The company stated that its expressive avatars could blur the boundaries between the virtual world and real personalities.
The goal of expressive avatars is to eliminate the need for cameras, microphones, actors, lengthy editing processes, and other costs in professional video production operations.
Synthesia’s team has a studio in London where actors read texts in front of a green screen to train the system.
The company demonstrated three lines of text on its platform representing joy, sadness, and frustration, and the AI-generated actor read the text in a manner that suited each emotion.
Compared to a real presenter in the video, digital avatars cannot change their tone, expressions, or gestures to align with the text.
According to Synthesia, new AI-powered avatars are now available that can understand the context and expressed emotions in a part of the text, and adjust their tone and expressions to deliver a speech.
The company claims the ability to display a variety of emotions using precise modifications in facial expressions, eyelash movements, and even eye movements to match the speech.
The company uses the EXPRESS-1 deep learning model trained on numerous texts and videos to achieve this goal, demonstrating how this text is pronounced in real life after several hours of training.
EXPRESS-1 company expects every movement and facial expression in real-time, focusing on timing and spoken language.
Over 55,000 companies use Synthesia’s technology to create digital avatars for presentations and training videos, including half of the companies listed in the Fortune 100.
Synthesia was founded in 2017 and managed to raise $90 million from investors in the past year, with a valuation of around $1 billion, making it one of the newest artificial intelligence companies in Britain.
