Configuration options

You configure each CameraX use case to control different aspects of the use case's operations.

For example, with the image capture use case, you can set a target aspect ratio and a flash mode. The following code shows one example:

val imageCapture = ImageCapture.Builder()
    .setFlashMode(...)
    .setTargetAspectRatio(...)
    .build()

ImageCapture imageCapture =
    new ImageCapture.Builder()
        .setFlashMode(...)
        .setTargetAspectRatio(...)
        .build();

In addition to configuration options, some use cases expose APIs to dynamically alter settings after the use case has been created. For information on configuration that is specific to the individual use cases, see Implement a preview, Analyze images, and Image capture.

CameraXConfig

For simplicity, CameraX has default configurations such as internal executors and handlers that are suitable for most usage scenarios. However, if your application has special requirements or prefers to customize those configurations, CameraXConfig is the interface for that purpose.

With CameraXConfig, an application can do the following:

• Optimize startup latency with setAvailableCameraLimiter().
• Provide the application's executor to CameraX with setCameraExecutor().
• Replace the default scheduler handler with setSchedulerHandler().
• Change the logging level with setMinimumLoggingLevel().

Usage Model

The following procedure describes how to use CameraXConfig:

1. Create a CameraXConfig object with your customized configurations.
2. Implement the CameraXConfig.Provider interface in your Application, and return your CameraXConfig object in getCameraXConfig().
3. Add your Application class to your AndroidManifest.xml file, as described here.

For example, the following code sample restricts CameraX logging to error messages only:

class CameraApplication : Application(), CameraXConfig.Provider {
   override fun getCameraXConfig(): CameraXConfig {
       return CameraXConfig.Builder.fromConfig(Camera2Config.defaultConfig())
           .setMinimumLoggingLevel(Log.ERROR).build()
   }
}

Keep a local copy of the CameraXConfig object if your application needs to know the CameraX configuration after setting it.

Camera Limiter

During the first invocation of ProcessCameraProvider.getInstance(), CameraX enumerates and queries characteristics of the cameras available on the device. Because CameraX needs to communicate with hardware components, this process can take a non-trivial amount of time for each camera, particularly on low-end devices. If your application only uses specific cameras on the device, such as the default front camera, you can set CameraX to ignore other cameras, which can reduce startup latency for the cameras your application uses.

If the CameraSelector passed to CameraXConfig.Builder.setAvailableCamerasLimiter() filters out a camera, CameraX behaves as if that camera does not exist. For example, the following code limits the application to only use the device's default back camera:

class MainApplication : Application(), CameraXConfig.Provider {
   override fun getCameraXConfig(): CameraXConfig {
       return CameraXConfig.Builder.fromConfig(Camera2Config.defaultConfig())
              .setAvailableCamerasLimiter(CameraSelector.DEFAULT_BACK_CAMERA)
              .build()
   }
}

Threads

Many of the platform APIs on which CameraX is built require blocking interprocess communication (IPC) with hardware that can sometimes take hundreds of milliseconds to respond. For this reason, CameraX only calls these APIs from background threads, which ensures the main thread is not blocked and that the UI remains fluid. CameraX internally manages these background threads so that this behavior appears transparent. However, some applications require strict control of threads. CameraXConfig allows an application to set the background threads that will be used through CameraXConfig.Builder.setCameraExecutor() and CameraXConfig.Builder.setSchedulerHandler().

Camera Executor

The camera executor is used for all internal Camera platform API calls, as well as for callbacks from these APIs. CameraX allocates and manages an internal Executor to perform these tasks. However, if your application requires stricter control of threads, use CameraXConfig.Builder.setCameraExecutor().

Scheduler Handler

The scheduler handler is used to schedule internal tasks at fixed intervals, such as retrying opening the camera when it isn't available. This handler does not execute jobs, and only dispatches them to the camera executor. It is also sometimes used on the legacy API platforms that require a Handler for callbacks. In these cases, the callbacks are still only dispatched directly to the camera executor. CameraX allocates and manages an internal HandlerThread to perform these tasks, but you can overridde it with CameraXConfig.Builder.setSchedulerHandler().

Logging

CameraX logging allows applications to filter logcat messages, as it may be good practice to avoid verbose messages in your production code. CameraX supports four logging levels, from the most verbose to the most severe:

• Log.DEBUG (default)
• Log.INFO
• Log.WARN
• Log.ERROR

Refer to the Android Log documentation for detailed descriptions of these log levels. Use CameraXConfig.Builder.setMinimumLoggingLevel(int) to set the appropriate logging level for your application.

Automatic selection

CameraX automatically provides functionality that is specific to the device that your app is running on. For example, CameraX will automatically determine the best resolution to use if you don't specify a resolution, or if the resolution you specify is unsupported. All of this is handled by the library, eliminating the need for you to write device-specific code.

CameraX's goal is to successfully initialize a camera session. This means CameraX compromises on resolution and aspect ratios based on device capability. The compromise may happen because:

• The device doesn't support the requested resolution.
• The device has compatibility issues, such as legacy devices that require certain resolutions to operate correctly.
• On some devices, certain formats are only available at certain aspect ratios.
• The device has a preference for a "nearest mod16" for JPEG or video encoding. See SCALER_STREAM_CONFIGURATION_MAP for more information.

Although CameraX creates and manages the session, you should always check the returned image sizes on the use case output in your code and adjust accordingly.

Rotation

By default, the camera rotation is set to match the default display's rotation during the creation of the use case. In this default case, CameraX produces outputs to allow the app to easily match what you would expect to see in the preview. You can change the rotation to a custom value to support multi-display devices by passing in the current display orientation when configuring use case objects or dynamically after they have been created.

Your app can set the target rotation using configuration settings. It can then update rotation settings by using the methods from the use case APIs (such as ImageAnalysis.setTargetRotation()), even while the lifecycle is in a running state. You might use this when the app is locked to portrait mode—and so no reconfiguration occurs on rotation—but the photo or analysis use case needs to be aware of the current rotation of the device. For example, rotation awareness may be needed so faces are oriented correctly for face detection, or photos are set to landscape or portrait.

Data for captured images might be stored without rotation information. Exif data contains rotation information so that gallery applications can show the image in the correct orientation after saving.

To display preview data with the correct orientation, you can use the metadata output from Preview.PreviewOutput() to create transforms.

The following code sample shows how to set the rotation on an orientation event:

override fun onCreate() {
    val imageCapture = ImageCapture.Builder().build()

    val orientationEventListener = object : OrientationEventListener(this as Context) {
        override fun onOrientationChanged(orientation : Int) {
            // Monitors orientation values to determine the target rotation value
            val rotation : Int = when (orientation) {
                in 45..134 -> Surface.ROTATION_270
                in 135..224 -> Surface.ROTATION_180
                in 225..314 -> Surface.ROTATION_90
                else -> Surface.ROTATION_0
            }

            imageCapture.targetRotation = rotation
        }
    }
    orientationEventListener.enable()
}

@Override
public void onCreate() {
    ImageCapture imageCapture = new ImageCapture.Builder().build();

    OrientationEventListener orientationEventListener = new OrientationEventListener((Context)this) {
       @Override
       public void onOrientationChanged(int orientation) {
           int rotation;

           // Monitors orientation values to determine the target rotation value
           if (orientation >= 45 && orientation < 135) {
               rotation = Surface.ROTATION_270;
           } else if (orientation >= 135 && orientation < 225) {
               rotation = Surface.ROTATION_180;
           } else if (orientation >= 225 && orientation < 315) {
               rotation = Surface.ROTATION_90;
           } else {
               rotation = Surface.ROTATION_0;
           }

           imageCapture.setTargetRotation(rotation);
       }
    };

    orientationEventListener.enable();
}

Based on the set rotation, each use case will either rotate the image data directly or provide rotation metadata to the consumers of the non-rotated image data.

• Preview: Metadata output is provided so that the rotation of the target resolution is known using Preview.getTargetRotation().
• ImageAnalysis: Metadata output is provided so that image buffer coordinates are known relative to display coordinates.
• ImageCapture: The image Exif metadata, buffer, or both the buffer and metadata will be altered to note the rotation setting. The value altered depends upon the HAL implementation.

Crop rect

By default, the crop rect is the full buffer rect. You can customize it with ViewPort and UseCaseGroup. By grouping use cases and setting the viewport, CameraX guarantees that the crop rects of all the use cases in the group point to the same area in the camera sensor.

The following code snippet shows how to use these two classes:

val viewPort =  ViewPort.Builder(Rational(width, height), display.rotation).build()
val useCaseGroup = UseCaseGroup.Builder()
    .addUseCase(preview)
    .addUseCase(imageAnalysis)
    .addUseCase(imageCapture)
    .setViewPort(viewPort)
    .build()
cameraProvider.bindToLifecycle(lifecycleOwner, cameraSelector, useCaseGroup)

ViewPort viewPort = new ViewPort.Builder(
         new Rational(width, height),
         getDisplay().getRotation()).build();
UseCaseGroup useCaseGroup = new UseCaseGroup.Builder()
    .addUseCase(preview)
    .addUseCase(imageAnalysis)
    .addUseCase(imageCapture)
    .setViewPort(viewPort)
    .build();
cameraProvider.bindToLifecycle(lifecycleOwner, cameraSelector, useCaseGroup);

ViewPort defines the buffer rect visible to end users. Then CameraX calculates the largest possible crop rect based on the properties of the viewport and the attached use cases. Usually, to achieve a WYSIWYG effect, you should configure the viewport based on the preview use case. A simple way to get the viewport is to use PreviewView.

The following code snippets shows how to get the ViewPort object:

val viewport = findViewById<PreviewView>(R.id.preview_view).viewPort

ViewPort viewPort = ((PreviewView)findViewById(R.id.preview_view)).getViewPort();

In the preceding example, what the app gets from ImageAnalysis and ImageCapture matches what the end user sees in PreviewView, assuming the PreviewView's scale type is set to the default, FILL_CENTER. After applying the crop rect and rotation to the output buffer, the image from all use cases will be the same, though possibly with different resolutions. For more information on how to apply the transformation info, see transform output.

Camera selection

CameraX automatically selects the best camera device for your application’s requirements and use cases. If you wish to use a different device than the one selected for you, there are a few options:

• Request the default front facing camera with CameraSelector.DEFAULT_FRONT_CAMERA.
• Request the default rear facing camera with CameraSelector.DEFAULT_BACK_CAMERA.
• Filter the list of available devices by their CameraCharacteristics with CameraSelector.Builder.addCameraFilter().

The following code sample illustrates how to create a CameraSelector to influence device selection:

fun selectExternalOrBestCamera(provider: ProcessCameraProvider):CameraSelector? {
   val cam2Infos = provider.availableCameraInfos.map {
       Camera2CameraInfo.from(it)
   }.sortedByDescending {
       // HARDWARE_LEVEL is Int type, with the order of:
       // LEGACY < LIMITED < FULL < LEVEL_3 < EXTERNAL
       it.getCameraCharacteristic(CameraCharacteristics.INFO_SUPPORTED_HARDWARE_LEVEL)
   }

   return when {
       cam2Infos.isNotEmpty() -> {
           CameraSelector.Builder()
               .addCameraFilter {
                   it.filter { camInfo ->
                       // cam2Infos[0] is either EXTERNAL or best built-in camera
                       val thisCamId = Camera2CameraInfo.from(camInfo).cameraId
                       thisCamId == cam2Infos[0].cameraId
                   }
               }.build()
       }
       else -> null
    }
}

// create a CameraSelector for the USB camera (or highest level internal camera)
val selector = selectExternalOrBestCamera(processCameraProvider)
processCameraProvider.bindToLifecycle(this, selector, preview, analysis)

Camera resolution

You can choose to allow CameraX to set image resolution based on a combination of the device capabilities, device’s supported hardware level, use case, and provided aspect ratio. Alternatively, you can set a specific target resolution or a specific aspect ratio in use cases that support that configuration.

Automatic resolution

CameraX can automatically determine the best resolution settings based on the use cases specified in cameraProcessProvider.bindToLifecycle(). Whenever possible, specify all the use cases needed to run concurrently in a single session in a single bindToLifecycle() call. CameraX will determine resolutions based on the set of use cases bound by considering the device’s supported hardware level and by accounting for device-specific variance (where a device may exceed or not meet the stream configurations available). The intent is to allow the application to run on a wide variety of devices while minimizing device-specific code paths.

The default aspect ratio for image capture and image analysis use cases is 4:3.

Use cases have a configurable aspect ratio to allow the application to specify the desired aspect ratio based on UI design. CameraX output will be produced to match the aspect ratios requested as closely as the device supports. If there is no exact-match resolution supported, the one that fulfills the most conditions is selected. Thus the application dictates how the camera should appear in the app, and CameraX determines the best camera resolution settings to satisfy that on different devices.

For example an app can do any of the following:

• Specify a target resolution of 4:3 or 16:9 for a use case
• Specify a custom resolution, which CameraX will attempt to find the closest match to
• Specify a cropping aspect ratio for ImageCapture

CameraX will choose the internal Camera2 surface resolutions automatically. The following table shows the resolutions:

Specify a resolution

You can set specific resolutions when building use cases using the setTargetResolution(Size resolution) method, as shown in the following code sample:

val imageAnalysis = ImageAnalysis.Builder()
    .setTargetResolution(Size(1280, 720))
    .build()

ImageAnalysis imageAnalysis =
  new ImageAnalysis.Builder()
    .setTargetResolution(new Size(1280, 720))
    .build();

You can't set both target aspect ratio and target resolution on the same use case. Doing so will throw an IllegalArgumentException when building the config object.

Express the resolution Size in the coordinate frame after rotating the supported sizes by the target rotation. For example, a device with portrait natural orientation in natural target rotation requesting a portrait image may specify 480x640, and the same device, rotated 90 degrees and targeting landscape orientation may specify 640x480.

The target resolution attempts to establish a minimum bound for the image resolution. The actual image resolution will be the closest available resolution in size that is not smaller than the target resolution, as determined by the Camera implementation. However, if no resolution exists that is equal to or larger than the target resolution, the nearest available resolution smaller than the target resolution will be chosen. Resolutions with the same aspect ratio of the provided Size are given higher priority than resolutions of different aspect ratios.

CameraX will apply the best suitable resolution based on the requests. If the primary need is to satisfy aspect ratio, specify only setTargetAspectRatio, and CameraX will determine a specific resolution suitable based on the device. If the primary need of the app is to specify a resolution in order to make image processing more efficient (for example a small or mid-sized image based on device processing capability), use setTargetResolution(Size resolution).

If your app requires an exact resolution, see the table within createCaptureSession() to determine what maximum resolutions are supported by each hardware level. To check for the specific resolutions supported by the current device, see StreamConfigurationMap.getOutputSizes(int).

If your app is running on Android 10 or higher, you can use isSessionConfigurationSupported() to verify a specific SessionConfiguration.

Control camera output

In addition to letting you configure the camera output as-needed for each individual use case, CameraX also implements the following interfaces to support camera operations common to all bound use cases:

• CameraControl lets you configure common camera features.
• CameraInfo lets you query the states of those common camera features.

These are the supported camera features with CameraControl:

• Zoom
• Torch
• Focus and Metering (tap-to-focus)
• Exposure Compensation

Get instances of CameraControl and CameraInfo

Retrieve instances of CameraControl and CameraInfo using the Camera object returned by ProcessCameraProvider.bindToLifecyle(). The following code shows an example:

val camera = processCameraProvider.bindToLifecycle(lifecycleOwner, cameraSelector, preview)

// For performing operations that affect all outputs.
val cameraControl = camera.cameraControl
// For querying information and states.
val cameraInfo = camera.cameraInfo

Camera camera = processCameraProvider.bindToLifecycle(lifecycleOwner, cameraSelector, preview)

// For performing operations that affect all outputs.
CameraControl cameraControl = camera.getCameraControl()
// For querying information and states.
CameraInfo cameraInfo = camera.getCameraInfo()

For example, you can submit zoom and other CameraControl operations after calling bindToLifecycle(). After you stop or destroy the activity used to bind the camera instance, CameraControl can no longer execute operations and returns a failed ListenableFuture.

Zoom

CameraControl offers two methods for changing the zoom level:

• setZoomRatio() sets the zoom by the zoom ratio.

  The ratio must be within the range of CameraInfo.getZoomState().getValue().getMinZoomRatio() and CameraInfo.getZoomState().getValue().getMaxZoomRatio(). Otherwise the function returns a failed ListenableFuture.

• setLinearZoom() sets the current zoom with a linear zoom value ranging from 0 to 1.0.

  The advantage of linear zoom is that it ensures the field of view (FOV) scales with changes in zoom. This makes it ideal for use with a Slider view.

CameraInfo.getZoomState() returns a LiveData of the current zoom state. The value changes when the camera is initialized or if the zoom level is set using setZoomRatio() or setLinearZoom(). Calling either method sets the values backing ZoomState.getZoomRatio() and ZoomState.getLinearZoom(). This is helpful if you want to display zoom ratio text alongside a slider. Simply observe the ZoomState LiveData to update both without needing to do a conversion.

The ListenableFuture returned by both APIs offers the option for applications to be notified when a repeating request with the specified zoom value is completed. In addition, if you set a new zoom value while the previous operation is still executing, the previous zoom operation's ListenableFuture fails immediately.

Torch

CameraControl.enableTorch(boolean) enables or disables the torch (also known as the flashlight).

CameraInfo.getTorchState() can be used to query the current torch state. You can check the value returned by CameraInfo.hasFlashUnit() to determine whether a torch is available. If not, calling CameraControl.enableTorch(boolean) causes the returned ListenableFuture to complete immediately with a failed result and sets the torch state to TorchState.OFF.

When the torch is enabled, it remains on during photo and video capture regardless of the flashMode setting. The flashMode in ImageCapture works only when the torch is disabled.

Focus and Metering

CameraControl.startFocusAndMetering() triggers autofocus and exposure metering by setting AF/AE/AWB metering regions based on the given FocusMeteringAction. This is often used to implement the “tap to focus” feature in many camera applications.

MeteringPoint

To begin, create a MeteringPoint using MeteringPointFactory.createPoint(float x, float y, float size). A MeteringPoint represents a single point on the camera Surface. It’s stored in a normalized form so that it can be easily converted to sensor coordinates for specifying AF/AE/AWB regions.

The size of the MeteringPoint ranges from 0 to 1, with a default size of 0.15f. When calling MeteringPointFactory.createPoint(float x, float y, float size), CameraX creates a rectangle region centered at (x, y) for the provided size.

The following code demonstrates how to create a MeteringPoint:

// Use PreviewView.getMeteringPointFactory if PreviewView is used for preview.
previewView.setOnTouchListener((view, motionEvent) ->  {
val meteringPoint = previewView.meteringPointFactory
    .createPoint(motionEvent.x, motionEvent.y)
…
}

// Use DisplayOrientedMeteringPointFactory if SurfaceView / TextureView is used for
// preview. Please note that if the preview is scaled or cropped in the View,
// it’s the application's responsibility to transform the coordinates properly
// so that the width and height of this factory represents the full Preview FOV.
// And the (x,y) passed to create MeteringPoint may need to be adjusted with
// the offsets.
val meteringPointFactory = DisplayOrientedMeteringPointFactory(
     surfaceView.display,
     camera.cameraInfo,
     surfaceView.width,
     surfaceView.height
)

// Use SurfaceOrientedMeteringPointFactory if the point is specified in
// ImageAnalysis ImageProxy.
val meteringPointFactory = SurfaceOrientedMeteringPointFactory(
     imageWidth,
     imageHeight,
     imageAnalysis)

startFocusAndMetering and FocusMeteringAction

To invoke startFocusAndMetering(), applications must build a FocusMeteringAction, which consists of one or more MeteringPoints with optional metering mode combinations from FLAG_AF, FLAG_AE, FLAG_AWB. The follow code demonstrates this usage:

val meteringPoint1 = meteringPointFactory.createPoint(x1, x1)
val meteringPoint2 = meteringPointFactory.createPoint(x2, y2)
val action = FocusMeteringAction.Builder(meteringPoint1) // default AF|AE|AWB
      // Optionally add meteringPoint2 for AF/AE.
      .addPoint(meteringPoint2, FLAG_AF | FLAG_AE)
      // The action will be canceled in 3 seconds (if not set, default is 5s).
      .setAutoCancelDuration(3, TimeUnit.SECONDS)
      .build()

val result = cameraControl.startFocusAndMetering(action)
// Adds listener to the ListenableFuture if you need to know the focusMetering result.
result.addListener({
   // result.get().isFocusSuccessful returns if the auto focus is successful or not.
}, ContextCompat.getMainExecutor(this)

As shown in the code above, startFocusAndMetering() takes a FocusMeteringAction consisting of one MeteringPoint for AF/AE/AWB metering regions and another MeteringPoint for AF and AE only.

Internally, CameraX converts it into Camera2 MeteringRectangles and sets the corresponding CONTROL_AF_REGIONS / CONTROL_AE_REGIONS / CONTROL_AWB_REGIONS parameters to the capture request.

Since not every device supports AF/AE/AWB and multiple regions, CameraX executes the FocusMeteringAction with best effort. CameraX will use the maximum number of MeteringPoints supported, in the order that points were added. All MeteringPoints added after the maximum count are ignored. For example, if a FocusMeteringAction is supplied with 3 MeteringPoints on a platform supporting just 2, only the first 2 MeteringPoints are used. The final MeteringPoint is ignored by CameraX.

Exposure Compensation

Exposure compensation is useful when applications need to fine-tune exposure values (EV) beyond the auto exposure (AE) output result. Exposure compensation values are combined in the following way to determine the necessary exposure for current image conditions:

Exposure = ExposureCompensationIndex * ExposureCompensationStep

CameraX provides the Camera.CameraControl.setExposureCompensationIndex() function for setting the exposure compensation as an index value.

Positive index values make the image brighter, while negative values dim the image. Applications can query the supported range by CameraInfo.ExposureState.exposureCompensationRange() described in the next section. If the value is supported, the returned ListenableFuture completes when the value is successfully enabled in the capture request; if the specified index is out of the supported range, setExposureCompensationIndex() causes the returned ListenableFuture to complete immediately with a failed result.

CameraX keeps only the latest outstanding setExposureCompensationIndex() request, and calling the function multiple times before the previous request gets executed results in its cancellation.

The following snippet sets an exposure compensation index and registers a callback for when the exposure change request has been executed:

camera.cameraControl.setExposureCompensationIndex(exposureCompensationIndex)
   .addListener({
      // Get the current exposure compensation index, it may be
      // different from the asked value in case this request was
      // canceled by a newer setting request.
      val currentExposureIndex = camera.cameraInfo.exposureState.exposureCompensationIndex
      …
   }, mainExecutor)

• Camera.CameraInfo.getExposureState() retrieves the current ExposureState including:

  • The supportability of exposure compensation control.
  • The current exposure compensation index.
  • The exposure compensation index range.
  • The exposure compensation step used in exposure compensation value calculation.

For example, the following code initializes the settings for an exposure SeekBar with current ExposureState values:

val exposureState = camera.cameraInfo.exposureState
binding.seekBar.apply {
   isEnabled = exposureState.isExposureCompensationSupported
   max = exposureState.exposureCompensationRange.upper
   min = exposureState.exposureCompensationRange.lower
   progress = exposureState.exposureCompensationIndex
}

Additional resources

To learn more about CameraX, consult the following additional resources.

Codelab

Code sample

Developer community

Android CameraX Discussion Group