D-ID, the company behind Deep Nostalgia, lets you create AI-generated videos from a single image

Israeli AI company D-ID, which offered the expertise for initiatives like Deep Nostalgia, is launching a new platform the place customers can add a single image and textual content to create a video. With this new web site, known as Creative Reality Platform, the company is focusing on sectors resembling company coaching and training, inside and exterior company communications, product advertising and gross sales.

The platform is sort of straightforward to make use of: customers can add a speaker image or choose considered one of the pre-made audio system to start out the video creation course of. Paid customers can entry premium presenters, that are extra “expressive” as a result of they’ve higher facial expressions and hand actions than the normal ones. Users can then kind from a script or just ship an audio recording of somebody talking. Users can then select the language (the platform helps 119 languages), voice and types resembling humorous, unhappy, animated and pleasant.

The company’s AI-based algorithms will generate videos primarily based on these parameters. Users can share videos wherever. The company claims that the algorithm makes use of solely half the size of a video to create a clip, however in our exams it took a number of minutes to create a one-minute video. This could fluctuate relying on the kind of speaker and language chosen.

“The COVID-19 pandemic has accelerated the want for digital content material round the world. Creating instructional content material is a huge problem for organizations. Reading paperwork and watching shows might be dry and boring. In addition, they must spend 1000’s of {dollars} on hiring actors and creating coaching videos. So we’re utilizing our AI to create audio system and tutors to duplicate people and make content material extra partaking and efficient,” Gill Perry, CEO of D-ID, advised TechCrunch.

Perry pointed to many use circumstances for this expertise, from a multilingual message from the CEO to staff to personalised needs for customers in the group.

D-ID launched the studio for testing in mid-August to iron out bugs earlier than the public launch. And whereas its major focus is serving corporations of all sizes, the company sees many pursuits from the creators on the platform.

Creating offensive deep pretend videos is a danger. That’s why the Israeli company has put in place obstacles resembling filtering out profanity and racist feedback, in addition to image recognition to keep away from utilizing the faces of well-known individuals. It makes use of the Microsoft Azure Text Moderation API to take away sexual feedback and offensive language from video scripts. D-ID stated the platform’s phrases of use prohibit customers from making political videos. If any of those guidelines are violated, the company could droop the violator’s account and take away the video from the library.

D-ID raised $25 million in Series B funding led by Macquarie Capital in March, and has raised $47 million to this point. Until now, the company has relied on others utilizing its API to create content material – ​​Deep Nostalgia is a prime instance of this – with shoppers resembling Modelez, Warner Bros. The company is now increasing its worthwhile merchandise by releasing a PowerPoint plugin alongside this self-service platform. The plugin provides an interactive presenter to the deck, so customers do not simply must learn the slides. They can select from totally different avatars, voices and languages ​​– identical to the self-service platform. But at the second there is no such thing as a provision for a particular speaker.

Once launched, customers will probably be in a position to join a free 14-day trial account and create as much as 5 minutes of 720p AI-generated video. After that, they will pay $49 monthly for entry to fifteen minutes of Full HD AI-generated video, a PowerPoint plugin, and e mail help.

D-ID is co-founded by Eliran Kuta, Gil Perry and Sella Blondheim

Users can even add their very own audio recordings for voice cloning. In addition, the company is engaged on a device that may permit customers to add their very own pictures to show the AI ​​to be extra expressive in order that it will probably higher imitate a individual in a video. All these options will probably be restricted to the enterprise degree of the company.

Although the company faces competitors from corporations like Rephrase.ai and Soul Machines in the AI-generated video house, it says there are nearly no corporations that may declare to create high-quality video from a single image.

Perry stated D-ID would not intend to restrict itself to company coaching, communications and advertising videos. It additionally goals to make it simpler to translate video calls in real-time and to clone audio system – making an avatar seem in the video as an alternative of you when you dictate the audio.

The company can be contemplating turning into a key participant in web3/metaverse improvement. “Because we’ve got the expertise of making a video from a single image. We’re excited about the right way to create digital avatars for the metaverse,” Perry stated.

Leave a Comment

Your email address will not be published.