First, it appears that the AirPods do not take images for you to save. Instead, they send still images in the background for Visual Intelligence to process, but nothing gets saved to your Photos app or anything like that.
The cameras seemingly operate in two modes: active and passive. Active mode is when a request is made (such as asking Siri), and each earbud captures a 640×640 image and returns a 1024×1024 processed image to your iPhone. In Passive mode, the cameras take really low-resolution 320×320 or 512×512 images at a continuous rate. While the framework focuses on standard color images, there may be other capabilities not listed, such as infrared spectrum captures.
Passive mode seems to be for background environment detection…to see if there’s a person in front of you, if you moved to a busy street, and so on. The actual camera sensors may be higher resolution, but for a variety of reasons (bandwidth, battery life, processing speed, etc.) Apple seems to be using lower-res images.
There’s mention of some processing capabilities such as lens distortion correction, dual camera calibration, and compensation for head motion. While the earbuds can take multiple images in a series, there is no video recording. The framework includes control for a hardware indicator and its brightness, which likely means that both earbuds will have some sort of recording light.
This is all subject to change, of course, especially since these AirPods aren’t expected to arrive until next fall. Test features revealed in beta frameworks are sometimes incomplete or don’t take into account later optimizations or hardware changes.
