CameraX: Google’s New Weapon in the Photography Wars
Ever since it introduced Android, Google has struggled to provide a camera framework for developers that allows them to both build camera-enabled apps quickly and take advantage of the advanced capabilities being offered by phone makers. Its first Camera API was limited, and the second version is complicated. Neither has a vendor-independent way to activate some of the advanced modes that have been added to phones for improving image quality and adding stylistic effects. Now, Google is launching CameraX, a library that provides not just a simplified way for developers to access Android’s Camera2 functionality, but provides extensions for additional capabilities.
CameraX Is a Simpler Way to Harness the Power of Camera2
To work with CameraX a developer specifies a desired use case with configuration options. Listeners are then added to handle the data output by the CameraX library — which can be either in the form of a data stream or written directly to a file. Finally, the use cases are bound to Android Architecture Lifecycles, so that CameraX can handle some of the housekeeping associated with setup and teardown of resources for the application. One nice feature of CameraX is that multiple use cases can be run simultaneously, so a preview can remain live while images or video are analyzed and perhaps captured, for example.
CameraX supports several use cases: Preview, to get an image on the display; Image analysis, to get direct access to the image buffer; and Image capture, to save a fully-processed image or video. Google provides several sample applications for various use cases. As is the recent trend, the ones I looked at are written in Kotlin with alternate Java versions.
After a few lines of setup, CameraX commands are fairly straightforward. For example, to simply capture an image from the camera, there is a takePicture() method. Prior to calling it, another simple call lets an application select which camera to use, or to set other parameters. As you’d expect, CameraX requires the app to have CAMERA permissions, and WRITE_EXTERNAL_STORAGE permission on versions of Android prior to Q if you want to capture directly to the file system.
Extending RAW Image Capture to Multi-frame Scenarios
CameraX Offers Extensions for HDR, Night, Portrait, and Beauty
For an app to have access to an extension, the phone maker needs to add a hook to the CameraX library to the vendor’s own API. If a vendor doesn’t provide an implementation of a capability, CameraX just reports it as being unavailable. Currently, developers using the alpha version of CameraX can make use of extensions on the following phones:
- Samsung (HDR, Night, Beauty, Auto): Galaxy Note 10 series (pictured, top)(demonstrated at Samsung’s SDC19)
- Huawei (HDR, Portrait): Mate 20 series, P30 series, Honor Magic 2, Honor View 20
CameraX Rollout
CameraX is currently in alpha, but it’s expected to get to beta status — meaning final APIs according to Google — in December. Samsung showcased its support for CameraX this week at its Software Developer Conference, and it co-hosted a session with Google on how developers can take advantage of both CameraX and its extensions on the latest Samsung phones.
Continue reading
Intel’s 10nm Sapphire Rapids CPU Delidded, Photographed
Intel's Sapphire Rapids has been sighted in the wild. It looks as if Intel will hook the chip together a bit differently than AMD does, at least at the moment.
Mars Orbiter Photographs Canyon Five Times Deeper Than Grand Canyon
The ESA didn't need a fancy new mission to snap these images either. You can thank the venerable Mars Express orbiter, which has been in service for almost 20 years.
Scientists Use Powerful New Climate Model to Recreate Iconic ‘Blue Marble’ Photograph
The simulation on the right is a dead ringer for the authentic photo on the left.
Google Hired Photographers To Help Train the AI-Powered ‘Clips’ Camera
Google unveiled this device in October 2017, but it's still not available. It's expected to go on sale soon, and now Google has detailed how it tuned the camera's neural network to understand what is a significant moment.