Midjourney Debuts AI Video Generator: Transforming Images into Short Animated Clips
Midjourney, the popular AI image generation platform, has officially launched its first foray into video creation, opening up a new realm of possibilities for its users. For now, the initial release focuses on generating short, dynamic clips from existing images, either those crafted within Midjourney itself or uploaded by users.
Think of it as taking a still photograph and breathing a little life into it. Once you've created an image in Midjourney, you'll notice a new "animate" button. Clicking this allows you to generate a brief, five-second video based on your vision, expressed via a text prompt. If you prefer, you can even use an image you've uploaded as the "starting frame" for your video masterpiece.
While the tool can automatically generate a simple animation that, as Midjourney puts it, "just makes things move," users have granular control. A "manual" button lets you specify the type of motion you desire, allowing you to fine-tune the animation to match your creative intent.
Taking Control of Your AI-Generated Video
Furthermore, you're not limited to a mere five seconds. You can extend an animation by four seconds, and repeat this process up to four times, resulting in a final video of up to 21 seconds. There are also "high" and "low" motion settings. Want both the subject and camera to move? Choose "high." Prefer to keep the camera steady while the subject comes alive? Select "low." The choice is yours.
Currently, Midjourney's AI video generator is accessible through its website and Discord server. To access it, a subscription is required, starting at $10 per month for 3.3 hours of "fast" GPU time. In terms of cost, Midjourney estimates that generating a video will be approximately eight times more expensive than creating an image, working out to about "one image worth of cost" per second of video.
However, this launch arrives amidst a legal backdrop. Disney and Universal have an ongoing lawsuit against Midjourney. They are concerned about the potential for copyright infringement, arguing that Midjourney's video generator could facilitate the creation of "endless unauthorized copies" of their copyrighted works.
Midjourney founder David Holz sees this initial video generator as merely a "stepping stone." The startup's ultimate goal is to develop "models capable of real-time open-world simulations." This vision aligns with the broader trend in the AI space, where tech giants like Google, OpenAI, and Meta are also developing AI video generators that respond to text prompts. Whether these advancements will revolutionize creative industries or raise further legal challenges remains to be seen.
Source: The Verge