AI video company Lightricks is bolstering games for professional content creators with major updates to its LTX video model, offering users the option to create videographic content up to 60 seconds at a time.
According to the company, LTXV is the first long AI video generation model to be published, with video lengths well above industry standard, which is just 8 seconds long. Rather than making short, sweet video clips, creators can now create more intense blows, one minute stories that are merely natural language prompts.
Focused on lasers on consumer use cases with visual content mobile applications such as VideoLeap and FaceTune, Lightricks highlights its pivot for professional content creators with the ability to turn another game into new AI model updates.
In addition to the long format of video, the company has made LTXV the first streaming-enabled model of its kind, and even better, it has provided users with the ability to oversee the video in real time by adding new prompts to various points along the video's timeline. This change will unprecedented control over the content people create.
https://www.youtube.com/watch?v=x-2_cs7ki00
Zeev Farbman, co-founder and CEO of Lightricks, believes these two new features will highlight the model, allowing users to operate longer video sequences for the first time. The new version of LTXV “translate coherent storytelling, demos or random clips with visual and semantic consistency into true media with creative intent,” he said in a statement.
Also Read: In the age of AI, trust is a real infrastructure
Enhanced real-time creation controls are made possible through Lightricks' new and improved auto-reflex video engine. After the first second of the user's long AI video is generated, additional prompts can be entered immediately to adapt the output on the fly, allowing the scene to evolve and fulfill your creative vision. Once each batch of video frames is generated, it “conditions” the next batch batch of these frames to ensure continuity and consistency as long form content is generated and streamed in real time.
Such capabilities were thought to open the door to generative storytelling and ongoing narrative ideas, according to the company's announcement.
Professional video creation
There are many use cases for the new LTXV Long Form Video feature. Social media influencers and ad studios can use it to generate vertical ads on demand. There is also new possibilities for video games as the model allows developers to quickly create live-rendered cutscenes in the game based on user actions used during gameplay.
Other use cases include augmented reality applications. This includes educational content that brings out characters generated in AI along with human performers, synchronized with real-time actions and evolves in real-time based on learner input.
This direction is in line with Lightricks' growing belief that as technology evolves and becomes more and more powerful, companies are increasingly interested in using and paying for AI applications. The company first began messing around with AI within its fun smartphone-based apps.
Later, Lightricks launched LTX Studio in 2023, will all-in-the-art AI technology, targeting very different audiences of advertising production teams, product marketers and aspiring filmmakers, with professional-grade creation tools to generate storyboards, pitch decks and even production-enabled footage.
LTX Studio is a subscription-based platform that dramatically enhances the basic functionality of the LTXV model. It is a suite of editing tools that allow creators to apply extensive project-level and shot-level controls to AI-generated content and convert them into smooth, professional videos without investing in film studios or human actors. This includes simple prompts, full script uploads, reference images, and even motion capture, feature initiation and medium-sized support.
AI as a joint partner
More recently, Lightricks has begun to emphasize the idea that AI could become a kind of co-partner of the human creators who care about all of their video productions.
In an interview with ITECHPOST earlier this year, Farbman said that true magic happens when treating AI as a kind of co-creator and using it to build ideas. “It starts with a spark of ideas from the start,” he said. “You share a vision, a rough storyline, maybe some characters. AI brings it back to life as a rich visual concept.”
The new features of LTXV seem to enhance that idea. “We have not only been encouraged by AI video, but we have reached a point where it was truly overseen,” added Lightricks co-founder and CTO Yaron Inger in a statement. The new model essentially stated that “turns AI video into a lengthy storytelling platform, not just visual tricks.”
Nevertheless, it is also clear that Lightricks has not completely forgotten its consumer roots, as the latest version of LTXV continues to be open source. This means that anyone can use whatever they can experiment. Like previous editions of the model, it can be found on platforms such as Face and Github's Clothes. And you can experiment with it at your own disposal, unless you are using it to generate millions of dollars in revenue.
The low resource requirements of the model should appeal to casual creators as well. Unlike other video models such as Openai's SORA, LTXV runs efficiently, allowing you to generate high-quality, high-resolution output on a single H100 GPU, and can also run on consumer-grade GPUs. In contrast, most video generation models require up to 8 GPUs to generate high-quality video clips.
Also Read: Aithority Interview with Lokesh Jindal, Axtria's Product Head
[To share your insights with us as part of editorial or sponsored content, please write to psen@itechseries.com]
