Ray 3.2 is video-to-video first
It starts from an existing source clip, then transforms the look while preserving the duration, motion, and structure you choose to keep.
What Is Ray 3.2?
Ray 3.2 is a controlled AI video model for transforming source clips. It is built for the moment when you already have footage and need to change the world around it.
Definition
Ray 3.2 turns source footage into directed variations
The simplest way to understand Ray 3.2: it keeps the useful parts of a source take and changes the parts your creative direction needs to change.
Keyframes make direction specific
Instead of hoping one prompt controls the whole clip, you can lock important moments and guide the transformation frame by frame.
Adherence controls the creative distance
Motion and Structure adherence decide whether the result stays close to the source or moves into a bigger stylistic change.
Ray 3.2 is about changing a shot without throwing the shot away
The easiest way to understand Ray 3.2 is to separate invention from transformation. Invention starts with a blank prompt or a loose reference and asks the model to imagine a shot. Transformation starts with a real source clip and asks the model to keep the useful production facts while changing the creative surface. That distinction matters. If a shot already has the right camera move, actor timing, hand placement, product reveal, or edit duration, a blank generation workflow can waste the work you already have. Ray 3.2 is useful because it treats the source clip as the foundation, not as a disposable suggestion.
The model is useful when continuity is part of the brief
Continuity can mean different things depending on the job. For a filmmaker, it may mean the transformed clip still lands on the same cut point. For an agency, it may mean the product stays in the same hand position while the market, background, or label changes. For a VFX team, it may mean a performer keeps the same gesture and expression while the character design changes. Ray 3.2 gives that continuity a practical control surface: source duration, keyframes, Motion adherence, Structure adherence, and character locks.
Ray 3.2 is not only a prettier render setting
The HDR and EXR story is important, but Ray 3.2 is not just about output quality. The bigger shift is directability. A video model becomes more useful when a creative team can point to a specific frame, preserve a specific motion, and change a specific visual layer without restarting the entire shot. Higher fidelity matters at export time; directability matters during every review cycle before export.
Capabilities
What Ray 3.2 can change
These capabilities all start from a source take. The point is not random generation; it is controlled transformation.
Motion transfer
Camera motion transfer
Character transformation
Visual effects
Environment change
Relighting
Product swap
Global campaign variations
Video examples
Four ways to understand Ray 3.2 through source footage
Each video shows a different reason Ray 3.2 exists: preserving motion, holding detail, choosing output quality, or transforming a cinematic environment.
Motion and source structure
This amusement-ride source has strong camera movement and a clear track path. It represents the kind of footage where Motion adherence is the main value: the transformed result should respect speed, perspective, and direction.
1080p detail and texture
The macro plant shot emphasizes fine detail, water droplets, and close-focus texture. It is useful for explaining why clean 1080p output matters when the generation needs to hold up beyond a small preview.
HDR output decision
The SDR/HDR comparison makes the output decision visible. Ray 3.2 is not only about creating a new look; it also gives teams a path toward brighter highlights, richer color separation, and finishing-friendly delivery.
Cinematic environment transformation
The wide cinematic landscape shows how environment, atmosphere, and composition can carry a shot. For Ray 3.2, this kind of clip is a reminder to preserve scale, horizon, and camera language when changing the world around a subject.
Use cases
Ray 3.2 is best when the source take still matters
Use it when the shot is not blank-page ideation anymore. Use it when continuity, timing, and production reuse matter.
01
A source clip is almost right, but the style or setting needs to change.
02
A product shot needs new markets, lighting, or backgrounds without reshooting.
03
A character performance should be preserved while the identity or world changes.
04
An edit needs the transformed result to keep the same duration and timing.
FAQ
What is Ray 3.2 FAQ
The short version before you choose a workflow.
What is Ray 3.2?+
Ray 3.2 is Luma's video model update for controlled video-to-video transformation. In plain terms, it helps you take a clip you already have and reshape the look, lighting, character, product, or environment while keeping the useful motion and timing.
What makes Ray 3.2 different?+
The important difference is control. Ray 3.2 is built around source footage, keyframes, Motion adherence, Structure adherence, and production-friendly output, so it is less about rolling a new idea and more about directing an existing shot.
Who should use Ray 3.2?+
It fits filmmakers, agencies, product teams, VFX artists, and creators who already have footage worth keeping. If the timing, camera move, performance, or product staging is valuable, Ray 3.2 is the right kind of workflow to consider.
Is Ray 3.2 only for professional studios?+
No. Professional teams benefit from the extra control, but the idea is simple enough for smaller creators too: start with a source clip, describe the result you want, then use keyframes and adherence settings when the shot needs more precision.
Does Ray 3.2 replace text-to-video?+
Not really. Text-to-video is useful when you are inventing a shot from scratch. Ray 3.2 is more useful when a shot already exists and you want to transform it without losing the structure that made it usable.
What kind of source video works best?+
Use a source clip with clear framing, readable subject motion, and a shot duration you actually want to keep. Clips with heavy compression, fast cuts, or distracting overlays are harder to transform cleanly.
Why do keyframes matter in Ray 3.2?+
Keyframes give the model exact moments to respect. They are especially helpful for product reveals, character poses, facial expressions, and final frames that need to land the same way every time.
When should I use HDR or EXR output?+
Use HDR or EXR when the clip needs to keep moving through a finishing workflow. For quick creative exploration, a draft is usually enough; for color, review, and delivery, higher-quality output matters.