I Tried The New Ray3 AI Video Maker by Luma Labs and I Found it…

Up to now yr, searches for “AI video maker” have skyrocketed by over 200%. This isn’t a shocker, as creators all over the place are searching for sooner, smarter methods to show concepts into visuals. In actual fact, the very thought of such AI instruments was to supply inventive freedom to everybody, whether or not with a videomaking background or not. Into this wave comes Luma AI Ray3, the latest video-generation mannequin from Luma Labs.

Positioned as a “visible thought companion,” Ray3 is constructed to assist creators ideate, storyboard, and refine their movies inside a single interface. Not less than that’s what the corporate goals at, as per its launch announcement. Powered by Ray3 for video and Photon for pictures, the system goals to push video creation from experimentation to execution, all in minutes.

I attempted my hand on the new Ray3. Right here, I share all about it, my expertise, and how one can attempt it out for your self. Learn on!

What’s Luma AI Ray3?

At its core, Luma AI Ray3 is the most recent improve to Luma Labs’ video-generation mannequin, designed to push previous the boundaries of Ray2. In contrast to earlier iterations that targeted totally on rendering quick clips, Ray3 introduces a unified inventive layer: you’ll be able to ideate, experiment, and direct visible outputs with out leaving the interface.

The system combines two engines underneath one roof – Ray3 for video and Photon for pictures. This pairing means you’ll be able to transfer fluidly between nonetheless pictures and movement, treating them as components of the identical inventive course of.

But, probably the most attention-grabbing function Ray3 brings is its reference-driven consistency. The brand new AI video-gen mannequin allows you to lock kinds or characters throughout a number of photographs, one thing many different AI video makers nonetheless battle with.

Key Options of Luma AI Ray3

Apart from the reference-driven video-making, Ray3 has a number of recent options. Here’s a deep dive into all of those.

1. Boards for Structured Creativity

Ray3 introduces a number of boards – Artboard, Storyboard, and Moodboard. These let creators organise uncooked concepts into structured visuals. As an alternative of leaping straight into video, you’ll be able to form the appear and feel earlier than committing to technology.

2. Modify with Pure Language

Want fast modifications? Merely kind directions like “make it anime” and the mannequin edits your picture or video accordingly. This bridges the hole between skilled enhancing instruments and on a regular basis inventive language.

3. Modify with Visible Cues

One of the vital novel options I’ve personally seen in an AI video maker, Ray3, now allows you to modify a video by including visible cues to it. The AI mannequin is ready to interpret these cues and make modifications to your video accordingly.

4. Reference for Fashion and Character Consistency

One of the vital highly effective additions: reference-based technology. By importing a visible reference, Ray3 can keep constant kinds, environments, and even recurring characters throughout a number of outputs. Most AI video makers nonetheless lack this capability.

5. Keyframe Management

Administrators typically assume when it comes to keyframes, and Ray3 embraces that. By defining a begin and finish body, you’ll be able to let the AI fill within the in-between movement, including a layer of management and predictability to generative video.

6. Brainstorm and Artistic Question

Past video technology, Ray3 doubles as an thought companion. Its question system can develop tough prompts into detailed inventive instructions, stopping inventive blocks.

7. Share & Remix

Each challenge comes with a built-in “behind the scenes” layer. Different creators can see your course of, remix your outputs, and construct collaboratively, making Ray3 as a lot a neighborhood device as a manufacturing engine.

8. HDR Era

Ray3 claims to be the world’s first mannequin that is ready to ship studio-grade HDR movies by means of native excessive dynamic vary shade technology. The AI video device is ready to generate gorgeous HDR movies from textual content prompts, SDR pictures, in addition to SDR movies. What’s extra, it will probably even export these as 16-bit EXR for seamless integration into professional workflows.

How Does It Work?

Most AI video makers (see the highest 10 right here) generate clips in isolation, however Luma AI Ray3 combines a number of fashions and controls into one pipeline. On the core are two engines working collectively:

• Ray3 (video mannequin): Generates movement sequences, interprets prompts, and handles transitions between frames.
• Photon (picture mannequin): Produces high-quality stills, references, and beginning factors that may stream immediately into video creation.
• Right here’s how the method unfolds whenever you use Ray3:
• Prompting or Referencing: You start with both a textual content immediate (“make it anime”) or a visible reference (a picture, a method, or a personality).
• Scene Era: Photon generates the stills, whereas Ray3 animates them into clean movement, making use of lighting, textures, and digital camera results.
• Keyframe Course: For those who present a begin and finish body, Ray3 fills within the in-betweens, simulating how a director would plan photographs.
• Edits and Modifications: You possibly can iterate immediately utilizing pure language, from fashion swaps to environmental modifications, with out touching conventional enhancing software program.
• Collaboration Layer: Your outputs and processes will be shared, remixed, or prolonged by different creators on the platform.

The best way to Entry Luma AI Ray3

For the time being, Dream Machine is out there on each the net and iOS. Ray3, nonetheless, appears to be in a trial mode at current on each.

Ray3 AI Video Maker: Palms-on

We tried the Ray3 for a novel video round a mystical theme utilizing the next immediate:

Generate a sweeping cinematic sequence of a lone wanderer – a tall, cloaked determine with glowing amber eyes – crossing an limitless ocean of clouds on a floating obsidian bridge. Every step sends faint ripples of sunshine by means of the bridge, as if it’s alive. The digital camera glides outward to disclose colossal celestial beings carved of residing starlight, suspended within the sky like historic guardians.

Their silhouettes stretch throughout the heavens, topped with halos of orbiting moons. Because the traveler strikes ahead, the beings slowly flip their heads, illuminating the trail with cascades of golden cosmic fireplace.

The traveler stops on the fringe of the bridge, cloak billowing, because the scene erupts in a symphony of auroras, capturing stars, and resonant cosmic hums.The temper is epic, surreal, and awe-inspiring – a visible surprise that blends fantasy, sci-fi, and dreamlike delusion.

Right here is the output that the AI device generated:

As you’ll be able to see, the video is kind of distinctive when it comes to high quality. It managed to ascertain a complete new state of affairs as talked about within the immediate, and got here up with fairly a nuanced and correct video accordingly. It even tried to include the finer particulars inside the immediate, just like the lighting on the ft of the determine as he walks on the bridge, in addition to his amber crimson eyes.

Having stated that, I can see some inconsistencies within the video too. A number of the objects on the far finish are randomly altering shapes all through the video, whereas the celestial beings proven are performing completely random acts. Even the hooded determine is strolling a bit unusually, together with his arm completely disappearing from imaginative and prescient after some extent.

Ray3 AI Video Maker: My Expertise

I attempted a number of different movies too, utilizing easy textual content technology prompts in addition to some with reference pictures. Whereas Ray3 is ready to catch the overall objective of the video fairly effectively, it largely tends to fail within the nuances of issues and the finer particulars. A video of a fighter jet flying over Noida, as an example, turned out a very bizarre output. The complete shot was shaky, and the jet modified form halfway by means of the video. Identical with a video of an Indian father-son duo enjoying billiards (pool). The balls on the pool desk appeared to look out of mid-air after which disappear simply as abruptly.

I did some analysis on how Luma Labs greatest suggests using Ray3. As per the corporate, the mannequin works greatest with “quick, pure language, easy prompts. Whereas Ray3 tries to be as correct as attainable to the immediate given, it’s most snug with easy and shorter prompts which can be extremely particular. This additionally means the mannequin is healthier suited to movies round objects and environments which were envisioned earlier than. Give it one thing completely distinctive and tends to fumble in the dead of night.

The place it shines is the only of animations round a reference picture. Merely add the picture you need to animate, and it will provide you with a high-quality video output, across the identical, even with out the necessity for a immediate. Working example, in the event you add a picture of a canine, it would animate it with the canine shaking its head or wagging its tail. No immediate wanted. In case there are modifications, you’ll be able to then point out the identical by means of follow-up prompts, and Ray3 will get these executed too. Fairly simple, and really useful.

Conclusion

Whereas Ray3 is spectacular, I really feel it nonetheless has a protracted option to go as a great AI video maker. The checklist of recent options it now presents is, little question, extraordinarily useful. Animating a picture, including a keyframe, and even modifying a video utilizing pure language or visible cues are gamechangers in some ways. As is its capability to give you HDR movies within the paid model.

However to name it an ideal AI video technology device, there may be going to be some time in that.