Creating Motion That Wasn’t There
Cinematic motion is an illusion created by displaying a series of still images in rapid succession. The number of images, or frames, shown per second (fps) determines the smoothness of this motion. AI frame interpolation, also known as motion interpolation or frame generation, is a process that creates entirely new frames to insert between the originally filmed frames. For example, it can take 24 fps footage and intelligently generate the in-between frames needed to convert it to 48 fps or 60 fps, or it can be used to create extreme slow-motion effects from footage shot at a standard speed.
This technique represents a significant leap over older methods. Traditional frame interpolation often relied on ‘optical flow,’ which analyzed the vector movement of pixels between frames. This worked for simple movements but often produced bizarre, watery, and distracting visual artifacts when faced with complex motion, occlusions (where one object passes in front of another), or fast camera pans. AI-based methods, powered by deep learning, are far more sophisticated. Trained on vast datasets of video, these models learn to recognize objects and context, allowing them to generate whole, coherent frames that are much cleaner and more believable.
Applications in Modern Filmmaking
AI frame interpolation is a versatile tool used in various stages of post-production to solve both technical and creative problems.
-
Slow Motion (Slo-Mo): Creating compelling slow-motion footage normally requires shooting at a very high frame rate (e.g., 120 fps or higher) with a specialized camera. AI interpolation allows filmmakers to create buttery-smooth slow-motion sequences from footage shot at standard speeds (like 24 or 30 fps). This can be a creative choice or a practical one, saving a shot that was not originally intended to be in slow motion.
-
Frame Rate Conversion: Content often needs to be converted between different frame rates for various broadcast standards around the world. AI can perform these conversions more cleanly than traditional methods, reducing judder and motion artifacts.
-
Film Restoration: Archival footage is often damaged, with frames that are missing or unusable. AI interpolation can be used to generate new frames to fill these gaps, smoothing out jerky or incomplete sequences and making historical footage more watchable for modern audiences.
The High Frame Rate Controversy
The aesthetic implications of frame interpolation are deeply connected to the cinematic debate around High Frame Rate (HFR) filmmaking. Directors like Peter Jackson (The Hobbit trilogy at 48 fps) and Ang Lee (Gemini Man at 120 fps) have experimented with shooting and projecting films at higher-than-standard frame rates. The result is hyper-realistic, fluid motion that eliminates the traditional ‘stutter’ or motion blur associated with 24 fps film.
However, this look has been highly controversial. Many critics and viewers find HFR to be jarring and ‘unchinematic,’ arguing that it resembles a cheap soap opera, a live sports broadcast, or a video game—a phenomenon often called the ‘soap opera effect.’ The very motion blur and strobing of 24 fps that HFR eliminates is, for many, an integral part of the dream-like, larger-than-life quality of cinema. AI frame interpolation gives filmmakers the ability to create this HFR look in post-production, but in doing so, it forces them to confront the same artistic question: what is the fundamental texture of cinematic motion, and how does altering it change the audience’s emotional connection to the story?