Camera2 | Android L Camera2 Demo | Camera library

I am writing an app using Camera2 API, which should show preview from camera and take a picture. Currently my code works as follows:

1. When Camera Fragment is instantiated, wait for TextureView.SurfaceTextureListener.onSurfaceTextureAvailable to be called
2. In ViewModel get available and suitable picture and preview sizes from CameraCharacteristics, and emit found preview size to Fragment with LiveData
3. Fragment observes preview size LiveData and calls setDefaultBufferSize with new size for its TextureView's SurfaceTexture
4. When new size is set, capture session is created, and repeating preview request is set, so TextureView starts to show image from camera
5. To avoid disrupting other camera apps' work, all camera-related things are cleared after Fragment's onPause and steps 1-4 are followed again after onResume
6. Surface instance is shared between Fragment and camera logic classes: the shared variable is initialized with it in TextureView.SurfaceTextureListener.onSurfaceTextureAvailable and is set to null when TextureView.SurfaceTextureListener.onSurfaceTextureDestroyed is called

This works fine for some devices of popular brands with modern Android versions, but the app should work on the particular generic Chinese tablet with Android 6 ("CameraManager: Using legacy camera HAL"), and there I face a problem.

• When the camera is instantiated and preview is started, I see that preview size is 640x480 (so the image is stretched), however, the size passed to setDefaultBufferSize is 1280x720
• Logcat also is full of continuous Surface::setBuffersUserDimensions(this=0x7f55fb5200,w=640,h=480) messages
• I've found on SO, that on some Samsung devices with Android 5 some resolutions may not really be available for Camera2, but here when I close the app and open it again, the preview resolution is 1280x720 as needed
• So my guess is that I may call setDefaultBufferSize too early on first Camera Fragment setup, and only when the view is recreated when after the app was minimized, the needed resolution is "picked up"
• I also tried to call setDefaultBufferSize in lambda passed to TextureView.post, and it solved the problem except for the case when I should ask for user's permissions on Camera Fragment (ie when user opens the camera for the first time), so the Fragment is paused a few times to show permissions pop-ups. However, without TextureView.post setDefaultBufferSize is also called in main thread, so I guess that delay caused by TextureView.post was the game changer here
• Also in setDefaultBufferSize docs I see: The new default buffer size will take effect the next time the image producer requests a buffer to fill. For Canvas this will be the next time Surface.lockCanvas is called. For OpenGL ES, the EGLSurface should be destroyed (via eglDestroySurface), made not-current (via eglMakeCurrent), and then recreated (via eglCreateWindowSurface) to ensure that the new default size has taken effect. It seems to me that it may be about the case

I'm receving the below error in API 31 devices during Firebase Auth UI library(Only Phone number credential),

My app process starts a foreground service when it receives BOOT_COMPLETED action

That service is for video recording (dashboard camera app) which uses camera and microphone.

To start a camera from a foreground service on Android 11+ one of the conditions must be met (Exemptions from the restrictions to starting a camera/micro from a foreground service) https://developer.android.com/guide/components/foreground-services#bg-access-restriction-exemptions

The only exception I can use is:

The service is started by an app that has the START_ACTIVITIES_FROM_BACKGROUND privileged permission.

This one doesn't require any interactions from a user, a user just wants a video recording to be started automatically when he starts driving his car, he doesn't want manually make app visible or start recording from a notification, he doesn't care about all of this, which is ok, it's 2021 and such things can be done automatically, but thanks Google for such restrictions in 2021, yes, we should care about safety of private personal data of a user, but they just could add BACKGROUND_CAMERA, BACKGROUND_RECORD_AUDIO permissions with manual review on Google Play Console like they did with ACCESS_BACKGROUND_LOCATION

If I understood correctly to get START_ACTIVITIES_FROM_BACKGROUND permission I must have one of the conditions be met (Exemptions from the restrictions to starting activities from the background) https://developer.android.com/guide/components/activities/background-starts#exceptions

So I chose SYSTEM_ALERT_WINDOW permission, because if it was granted by a user before then no interactions from the user is required and app can start an activity from background. This one can be granted with Intent(Settings.ACTION_MANAGE_OVERLAY_PERMISSION)

So do I have now START_ACTIVITIES_FROM_BACKGROUND permission granted or not? I don't understand...

If yes then I can use a camera in foreground service which was started from some background action

But still

I've been testing a migration to Android's new Camera X library and checking the EXIF data with exif $FILE. Some of the attributes are displayed, but the amount of values seems truncated & it displays an error:

I can't figure out why by adding a second object the first object stops animating and doesn't run anymore? How can I go about figuring out why my objects are canceling one another out and how to create two different or more multiple threejs items that animate at the same time? I would like to run more than on three.js script in some file. Someone can help me? Thank you

I added the code from

https://developer.android.com/training/camerax

link . The preview, permissions and everything else is working just fine, however I am not able to capture photo from the button. I declared the button to have a listener and connected the function takePhoto(), but it doesn't seem to work, in the past it works fine and I don't know if it's because of the new version of the CameraX feature is using, I don't know but it doesn't seem to be responding correctly. I have the Logcat open and when I press the button, it has the message that is pressed:

I am trying to use ARCORE to return a depth image and also use the CameraX to return an RGB image. I can do both individually but when I combine both together the cameraX doesn't work.

I see that I must allow the shared camera but as far I searched, it can only be possible using the camera2 API.

Does anyone know any way of using the CameraX instead?

How to achieve portrait mode like blurring the background of image in video of people programmatically, this feature exists in IOS natively now since the last update, I think we can achieve this using something like openGL shader or something or some setting in Camera2 API, the thing is I does not look straight forward at all. any help would be appreciated, https://www.xda-developers.com/how-to-use-portrait-mode/ this link has an image as an example of what I am trying to achieve, the article is not technical and not mentioning anything related to programming.

is there a way to record a video with android Camera2 API or AndroidX Camera and obtain the file size while taking the video?

I want to have something like this: The user records a video and sees while recoding the file size of the current capture.

Is there any callback that gives the file size of the temp video file? Or is there a chance to pass some kind of "TempFileObserver"? Or an existing lib?

Kind regards for reading and sharing your experience. PS: I don't need a total implementation, reference is enough