Spatial video is a stereoscopic video format that records a scene through two horizontally offset cameras and stores both views in one file, so that a headset can show the left view to the left eye and the right view to the right eye. Among XR applications, it is the most widespread way for ordinary users to capture personal memories with real binocular depth and relive them later in extended reality.
Unlike 360-degree or 180-degree immersive formats, spatial video keeps a conventional rectangular frame. The viewer sees the footage as a floating window or a large virtual screen, and depth appears inside that window. The format became mainstream when smartphones and headsets gained native capture and playback, which turned stereo video from a specialist production technique into a consumer recording mode.
How spatial video captures and encodes depth
Spatial video works by recording two synchronized views whose small horizontal offset reproduces the way human eyes see the world from slightly different positions. The distance between the two lenses is the stereo baseline. When the brain fuses the two images, the horizontal disparity between matching points is interpreted as distance, which is the same cue that gives everyday vision its sense of depth.
On phones, the two views usually come from the main and ultra-wide cameras placed side by side when the device is held in landscape orientation. The wider image is cropped and matched to the main camera, and the system applies rectification, which aligns both views so that matching points lie on the same horizontal line. Headsets with forward cameras capture from lenses placed closer to the human interpupillary distance, which averages about 63 mm in adults.
The common storage format is MV-HEVC, the multiview extension of the HEVC video codec. It encodes the left view as a base layer and predicts the right view from it, so the second eye costs far less than a second independent stream. A player without multiview support simply decodes the base layer and shows ordinary 2D video, which keeps the files backward compatible. Metadata in the container describes the baseline, the horizontal field of view and a disparity adjustment that tells the player where to place the image plane.
The following table lists the main technical parameters and the values typical of current consumer spatial video.
| Parameter | Typical value | What it affects |
|---|---|---|
| Resolution per eye | 1080p, higher on dedicated cameras | Sharpness of the floating window |
| Frame rate | 30 fps, sometimes 60 fps | Smoothness of motion and comfort |
| Stereo baseline | From about 2 cm on phones to near 6 cm on headsets | Strength of the depth effect |
| Codec | MV-HEVC in a MOV or MP4 container | File size and player compatibility |
| Bitrate | Roughly 20 to 50 Mbit/s per file | Compression artifacts and storage use |
| Field of view | Narrower than the headset display | Size of the window in the scene |
Spatial video compared with other immersive formats
Spatial video differs from other immersive formats mainly in how much of the field of view it fills and whether the viewer can move around inside it. It gives stereoscopic depth within a frame, but the viewpoint is fixed by the camera, so leaning sideways does not reveal anything new.
The comparison below places spatial video next to the formats that are most often confused with it.
| Format | Field of view | Depth | Viewer movement |
|---|---|---|---|
| Spatial video | Framed window, often 60 to 70 degrees | Stereoscopic | Head rotation only |
| Monoscopic 360-degree video | Full sphere | None | Head rotation only |
| VR180 video | Front hemisphere | Stereoscopic | Head rotation only |
| Volumetric video | Free viewing angle | True 3D geometry | Six degrees of freedom |
| Conventional 3D film | Cinema or TV screen | Stereoscopic | None |
The practical consequence is that spatial video is cheap to capture and easy to share, while volumetric capture needs camera rigs or depth sensors and large data volumes. VR180 surrounds the viewer more completely, but it requires dedicated fisheye cameras and careful framing. Spatial video sits in the middle: a personal, low-effort depth format that fits well into a photo library.
Where spatial video is used
Spatial video is used wherever a sense of presence adds value but full immersion would be excessive or too costly to produce. The best-known use is personal memories: family events, travel and children growing up, viewed later at near life size. Other XR applications include product presentation, training clips and short documentary content.
- Personal archives recorded on phones and viewed in a headset photo app.
- Product and real estate presentation, where depth helps judge size and materials.
- Training and instruction, such as hand procedures where depth clarifies the movement.
- Remote presence, where a short clip shares a place with relatives or colleagues.
- Editorial and creative work, including short films designed for headset viewing.
Because the base layer stays playable in 2D, the same file can be posted to ordinary platforms while still offering full depth on a capable headset. This dual nature is one reason the format spread faster than earlier stereo video formats.
Recording spatial video that is comfortable to watch
Comfortable spatial video depends on controlled disparity, steady framing and adequate light. Because the brain must fuse two images, errors that are minor in 2D video become tiring in stereo. Experienced stereo producers follow the 1/30 rule of thumb, which keeps the baseline at about one thirtieth of the distance to the nearest subject; phone baselines are small, so subjects closer than roughly one meter quickly become uncomfortable.
The decision table below summarizes common recording situations and the choices that keep the result easy on the eyes.
| Situation | Recommended choice | Reason |
|---|---|---|
| Subject very close to the lens | Step back to about 1 to 2.5 m | Prevents excessive disparity and eye strain |
| Walking while filming | Stand still or move slowly | Camera motion in stereo can cause discomfort |
| Dim interior | Add light or choose a brighter spot | Noise differs between the two lenses |
| Wide landscape far away | Include a foreground element | Distant scenes show almost no depth |
| Objects cut by the frame edge | Reframe to keep them inside | Avoids stereo window violations |
A stereo window violation occurs when an object that appears in front of the screen plane is cut off by the frame edge. The brain receives contradictory cues, because the edge seems closer than the object it occludes. Players reduce the problem with floating window masks and disparity adjustment, but careful framing is the more reliable fix.
Limitations and common misconceptions
Spatial video has clear limits that follow from its fixed two-view geometry. It cannot offer parallax when the viewer moves the head sideways, and its depth effect weakens rapidly with distance, because disparity shrinks as objects move away from a small baseline.
| Symptom | Likely cause | Remedy |
|---|---|---|
| Flat, cardboard-like depth | Distant subject with a short baseline | Film closer subjects or layered scenes |
| Eye strain after a few seconds | Large disparity or vertical misalignment | Increase distance, keep the device level |
| Different colors in each eye | Lenses with unequal exposure or white balance | Shoot in even light |
| Blurry fine detail | Upscaled ultra-wide view and compression | Use good light and higher bitrate |
A common misconception is that spatial video is the same as volumetric or 6DoF video. It is not: it stores two flat images, not a 3D model of the scene. Another is that any stereo pair looks good in a headset; in practice, vergence-accommodation conflict and poor alignment can make badly recorded clips tiring, so short, steady takes give the best results.
Put simply
Spatial video is a recording made with two cameras that sit side by side, much like a pair of eyes. When the clip plays in a headset, each eye receives the picture taken by its own camera, and the brain combines them into a scene with real depth.
The result does not surround the viewer like a 360-degree film. It looks more like a window into a remembered moment, where people and objects stand at different distances. On an ordinary phone or computer, the same file simply plays as a normal flat video.
In practice
A parent records a child blowing out birthday candles with a phone held in landscape orientation, standing about 1.5 meters from the table and keeping the phone still for twenty seconds. The kitchen lights are on, and the cake sits in the foreground while relatives stand behind it.
Months later, the family watches the clip in a headset. The cake, the child and the relatives appear at clearly separate depths, and the moment feels close to being in the room again. The same file shared with grandparents who own no headset plays as an ordinary 2D video, so one recording serves everyone. The outcome is a personal memory that keeps its sense of space without any special production equipment.