I just bought a Google Cardboard Viewer and decided to take a shot at making an app. I have played around with OpenGL and have no problem with the basics of that.
I created a new project following the source from the only sample I could find "Treasure Hunt" linked to from the "Getting Started" for Cardboard.
I can run my own test application and the Treasure Hunt sample application from Google on my Nexus 7 tablet, but I get an error on my phone:
JNI DETECTED ERROR IN APPLICATION: can't call void com.google.vrtoolkit.cardboard.CardboardView$StereoRenderer.onNewFrame(com.google.vrtoolkit.cardboard.HeadTransform) on instance of my.app.cardboardtest.CardboardRenderer
Since I get the same error from the sample from Google, I don't think I need to share any of my code specifically. This is a link to their code:
https://github.com/googlesamples/cardboard-java
I have a T-Mobile Galaxy Note 3 running Android 5.0
Is this the final tipping point for me to upgrade my phone? lol
This seems to be a problem with your build in Android Studio.
Try moving the cardboard.jar file (that contains the vrtoolkit.so binary) to a folder called /lib and point to it.
Upshot is, android cannot find vrtoolkit.so in the delployed apk. The binary is a new addition to the jar, and the docs are older than that.
Or, just stick with Eclipse until Studio is a little more rattled in.
Just a headsup the learning curves around VR development are turbulent like the event horizon of a black hole ...
Related
I am on a tutorial about deploying custom Models on AutoML onto edge, in this case an android device.
link is here: https://cloud.google.com/vision/automl/docs/tflite-android-tutorial
I followed the step to "customize" it with my own model and run, but i end up with a lot of buffoverflowexceptions.
What I did:
I exported my autoML models and i downloaded (out of the 3: metadata,.txt,.tflite) the .txt and the .tflite.
I followed up renaming them "graph.lite" and "labels.txt" like in the example and replaced the tutorial's example with mine on android studios. I also followed the tutorial's changes for custom models, like changing to byte etc.
However, running the app crashes the app straight away. with the following error:
https://i.stack.imgur.com/RLeSy.png
I was hoping for an app that could detect chairs and tables!
Help is greatly appreciated!
We've recently implemented Vuforia using Vumarkers and are about to purchase a license. We've tested with the 'tuxedo' test Vumarker from Vuforia. Both in the Editor and Android worked just fine.
As a final step in our development process, we've made our own Vumarker using the Illustrator template. Everything works fine, the marker was validated, uploaded to their website and we've generated some test markers. Everything works fine in the Unity Editor.
But when we build our project on any device no markers are detected at all and no errors are thrown. Does anyone else experience this problem? We've tested in all light environments and marker sizes to make sure it's not a camera problem.
We've used the following versions
Unity 2017.3.0f3 Windows 64bit
Vuforia 7.0.36
Android 7.x
If you have indeed successfully created vumark instances and downloaded the dataset, and you are sure nothing is detected and no errors are thrown, and if you are using the exact code from the sample and only replaced the dataset...
Then the first option that comes to mind is - did you load and activate the new dataset associated with the VuMarks?
I have tried to use the Project Tango unity examples found here; https://github.com/googlesamples/tango-examples-unity with no success.
I have never used Unity before, and have successfully completed the Codelab: Motion Tracking example fromt he developer resources: https://developers.google.com/project-tango/apis/unity/unity-codelab-motion-tracking.
I downloaded the example files mentioned at the top, open one of the projects (tried several, all have the same problem). Once the project is open, I open the scene that come with it, I go to Build Settings, change platform to Android, change the Bundle Identifier and set the Minimum API Level to Jelly Bean 17 and finally click "Build and Run".
Now it starts to build the project, everything goes fine and the Unity logo shows up on the Tango. And that's it, the Unity logo shows up, no error messages on the computer, but nothing more happens. The program stalls at the Unity logo.
Am I missing some crucial step to get the example projects to actually run?
Thanks in advance!
The example code is Unity 5 based project, so please use the 5.x version of Unity to open it. If you use Unity 4.x version, the scene won't be opened up properly.
However, the code lab is based on Unity 4.6 and it is independent from our sample code. I understand this is a little bit confusing now, we are working on make sure the version are unified.
In short, I would suggest you open our example code with Unity 5.x version. You will be able to run it without a problem.
It was blank empty scene build. thats why after the unity logo it didn't load any scene or do anything.
You need to go to Unity Build Setting and drag and drop all the scenes from the examples folder in it.
Once you have all the scenes. Make sure that AreaDescriptionManagement is on the top of the list.
That is the only step missing from you.
Now, build and run. Everything should be fine.
i got a webapp build via flambe/haxe which is wrapped by phonegap build to get an .ipa and .apk. For now the only remaining error is that the app looses its context if i switch between apps or pressing home-button resulting in the following error on android:
"W/Adreno200-EGL(31381): <qeglDrvAPI_eglSetBlobCacheFuncsQCOM:5575>: EGL_BAD_PARAMETER"
"E/libEGL(31381): eglSetBlobCacheFuncsANDROID resulted in an error: 0x300c"
"E/SurfaceTexture(31381): [unnamed-31381-1] updateTexImage: invalid current EGLContext"
"E/CanvasTexture(31381): unexpected error: updateTexImage return -38"
The last two lines keep repeating (even after several minutes) until i close the app manually which results in the following error-message:
"E/libEGL(21636): call to OpenGL ES API with no current context (logged once per thread)"
Full errorlog can be viewed here: http://goo.gl/UkxwfI (Dropbox)
One thing that keeps bugging me is the following message since all "related" problems got that feature turned ON, but i cant find anything to change that.
"D/webcoreglue(31381): netstack: Memory Cache feature is OFF"
Another thing is with: D/CordovaActivity(31381): CB-3064: The errorUrl is null
As of http://goo.gl/qZWc4F (github) there seems to be a problem on resuming in cordova. (just strg+f "cb-3064")
Took me the whole day to use google on terms like:
EGL_BAD_PARAMETER, qeglDrvAPI_eglSetBlobCacheFuncsQCOM, SurfaceTexture, CanvasTexture, updateTexImage, invalid current EGLContext, eglSetBlobCacheFuncsANDROID, error: 0x300c, OpenGL ES API
but without any good hints or helpful info.
additional information:
tested on Sony Xperia J (ST26i) and android 4.1.2
min android SDK 14
android:hardwareAccelerated="true" (default since sdk 14)
landscape + fullscreen
webview used to display the website coming from flambe/haxe
used build phonegap / phonegap 3.3 (using 2.8 - 3.3 = same)
building it myself via cordova doesnt help (so it should not relate to phonegap build)
cordova.inappbrowser only used to open links in external browser, website running in webview
website only shows a html5-game (nothing fancy, just something like bejeweld)
Maybe there's some easy help by juggling with the Android-Manifest. I really dont want to do native since the webapp is shipped to iOS via Phonegap too.
Update 1
Tried to reproduce the error using native android and playing with different manifest-settings. But it seems like this has something to do with how phonegap/cordova handles the drawing of everything on screen.
Next things to do will involve posting on phonegap-forums and checking out iOS errorlog. Maybe i find something while comparing my results to: iOS app crashes on resuming
Update 2
iOS-Errorlog doesn't help. Exception Type 20 and Exception Code 8badf00d. Stack shows errors with graphics too (like in android). Errorlog see Link in Comment...
"Fixed" by building android and ios myself with cordova. Resuming/Restarting feels a bit bumpy but thats the next thing to fix.
Phonegap build just doesn't like the life cycle of apps.
I am new to android app development. I need to develop a very simple app that could detect coloured circles.Can someone please tell me in detail how to use openCV Circle Recognition code in developing a simple android app that can detect coloured circles. I am currently using openCV+ Eclipcs but it does not allow me to compile the source code in to .apk file
I have already done what u are saying...I have done the following
I download the full NVidia tegra-devpack-2.0-windows-2013-02-06-14961502 and installed It contains both android ADT for eclipse and eclipse also. It also contains the open CV and it automatically configures everything.
Everything works fine when I make a new android app like “my first HELLO WORLD app” and I can even install it to my android phone that means everything is working fine and everything is configured correctly.
The problem arises when I try to import the source code of the “CVCamera” an common app for detecting circles in android devices
When I import it in the eclipse as an android project from existing code it gives me following 2 errors
(a) Project 'CVCamera' is missing required Java project: 'OpenCV'
(b) The project cannot be built until build path errors are resolved.
I have tried lots of things I even tried to link the open CV library to the project but of no use
I have a simple objective I want to develop an android application for detection of coloured circles by using the already available code on the web. Please help me in achieving the following
I am using openCV4android . I can even create new apps but the problem arises when i try develop the application by using already made source code. I am using the sources code of the CV camera app which has been already developed by so many developers Following is the link of the error I am getting in the eclipse.
Please see the image
Following is the link of the CV camera application which I am trying to import
stanford.edu/class/ee368/Android/OpenCV/CVCamera_Canny.zip
Here is the link for another error “Build path entry is missing: OpenCV”
image
building workspace problim image
Use opencv4android in eclipse.
Before that you should deploy android ADT for eclipse, or download bundle for android dev.
Then follow the tutorial at opencv official site to start your way to android.
Use hough transform for circle detection.
You must link JavaCV lib into your project rather than OpenCV.
Android app cannot use OpenCV(C++) unless you use JNI, which is far beyond the scope of this question.
So please follow the tutorial for using JavaCV on Android rather than doing the import stuff on your own. Please make some JavaCV examples compile and run on your phone before you try to adapt the code.
Besides I strongly recommend you use openCV4android, which is sure to work, rather than JavaCV + Android, which takes your considerable time coping with configuration and integration.