The first examples of spatial video from the iPhone 15 Pro prompted predictable complaints: 1080p, 30 frames per second, and a small distance between the cameras. I understand the desire for more detail, but I am less convinced that a small baseline is always the wrong choice.

My immediate thought was close subjects. The distance between the two capture viewpoints affects how strongly a nearby object separates between the left and right images. A small baseline can make some close compositions more manageable, even while producing weaker disparity for distant objects.

The simplified stereo relationship is useful here:

disparity = focalLengthInPixels * baseline / distance

Keep everything else fixed and halve the baseline: disparity halves. Move the object twice as far away: disparity also halves. This does not describe every part of a spatial-video pipeline, but it captures the geometric relationship behind the observation.

Large disparity is not automatically better. It can make depth more pronounced while also making an image harder to view comfortably. The final experience depends on how the images are framed and displayed, the viewer's position, and the near and far content included in the shot.

A close object against a very distant background can be especially demanding. The two images must describe both depth ranges at once. A smaller capture baseline can reduce the disparity spread, potentially giving the composition more room before the difference becomes distracting.

That does not create missing image detail. Resolution and stereo geometry are separate questions. A close object can have comfortable depth while still revealing compression or limited spatial resolution. Similarly, a higher-resolution image can retain uncomfortable disparities. I would evaluate the two qualities separately.

There is also a difference between camera baseline and human eye separation. They are related to the appearance of stereo depth, but they are not interchangeable measurements. A captured stereo pair is presented through a display system, with its own scale and viewing geometry.

My expectation is that some intimate subjects could suit the phone's compact stereo arrangement: objects on a table, hands doing a small task, or a face at an appropriate distance. That remains a capture hypothesis. It needs examples judged in a headset, not just side-by-side frames on a monitor.

For such a comparison, I would keep the subject and lighting consistent and vary distance. Look for comfortable separation, boundary artifacts and how much useful detail survives. The goal would be to find the conditions the camera handles well, rather than decide that one baseline is universally ideal.

I am interested in spatial video as a format with a particular visual character. Understanding that character may be more productive than expecting a small phone camera to behave exactly like a dedicated stereo rig.