NatCorder - Video Recording API

It depends on how you implemented your frame commit section.

For instance, if you used to have something like this:

Code (CSharp):

1. AsyncGPUReadback.Request(framebuffer, 0, request => {
2.                  
3.                         NativeArray<byte> frameData = request.GetData<byte>();
4.                         unsafe
5.                         {
6.                             IntPtr ptr = (IntPtr)Unity.Collections.LowLevel.Unsafe.NativeArrayUnsafeUtility.GetUnsafePtr(frameData);
7.                             _natCorder.CommitFrame(ptr, timestamp);
8.                         }
9.                     }
10.                 });

a simple workaround would be something like this:

Code (CSharp):

1. NativeArray<byte> nativeArray = new NativeArray<byte>(framebuffer.width * framebuffer.height * 4, Allocator.TempJob);
2.                 AsyncGPUReadback.RequestIntoNativeArray(ref nativeArray, framebuffer, 0, request =>
3.                 {
4.                     try
5.                     {
6.                         if (_natCorder != null)
7.                         {
8.                             byte[] frameData = new byte[nativeArray.Length];
9.                             nativeArray.CopyTo(frameData);
10.                             _natCorder.CommitFrame(frameData, timestamp);
11.                         }
12.                     }
13.                     finally
14.                     {
15.                         nativeArray.Dispose();
16.                     }
17.                 });

 

Or better yet:

Code (CSharp):

1. using Unity.Collections.LowLevel.Unsafe; // somewhere above
2.  
3. var nativeArray = new NativeArray<byte>(recorder.frameSize.width * recorder.frameSize.height * 4, Allocator.TempJob);
4. AsyncGPUReadback.RequestIntoNativeArray(ref nativeArray, framebuffer, 0, request => {
5.     using (nativeArray)
6.         if (recorder != null)
7.             unsafe {
8.                 var nativeBuffer = (IntPtr)NativeArrayUnsafeUtility.GetUnsafePtr(nativeArray);
9.                 recorder.CommitFrame(nativeBuffer, timestamp);
10.             }
11. });

You could also create the `nativeArray` as a persistent one, hence removing the need for repeated re- and de-allocations. If you're using an unsafe context, you can avoid reading the pixel data into managed memory, then transferring it back to native memory (this roundtrip is pure overhead).

 

Ok thanks! Although I'm not able to add this code as you can't use unsafe in an iterator?

We are using 2019.4 and I read about the iOS issue with this

AsyncGPUReadback.Request

in the CameraInput.cs. So now I'm trying to replace the code there with the

AsyncGPUReadback.RequestIntoNativeArray

one.
Not much luck yet. I tried the code from Menion-Leah, but the Unity Editor crashes when stopping recording.

 

Good to know, thanks!
I must have done something wrong, because when I tried to use an approach similar to the one you suggested I ran into issues with NativeArray being either disposed/inaccessibile/unreleased (also tried Persistent/Temp).

The solution with the round trip prevented those issues.

 

Ah if you're trying to add this to CameraInput then yes you'll run into that error, since `OnFrame` is a coroutine. Can you share the crash logs?

 

Sounds like a Unity bug to me. In any case your code is fine.

 

Hello,

My company bought NatDevice and NatCorder two days ago with the purpose of taking videos mp4 using the phone cameras, but after all the weekend trying, no sound was able to be captured, I tried different Unity versions with no luck, the examples are not working, and not a single test was successful(Tried on WIndows editor and Android).
The examples in a clean project are not working either, is there something I should try?

 

Hi again!
It seems that some Android phones (so far about 30% of tested devices) have trouble with MP4 recording. Specifically, as soon as the recording starts, the whole app crashes.
IOs phones and Editor environments work fine.
The problem seems to be here:

Code (CSharp):

1. 07-06 15:59:25.539  9765  9812 E CRASH   : managed backtrace:
2. 07-06 15:59:25.539  9765  9812 E CRASH   :       #00 (wrapper managed-to-native) NatSuite.Recorders.Internal.Bridge:FrameSize (intptr,int&,int&)
3. 07-06 15:59:25.539  9765  9812 E CRASH   :       #01 NatSuite.Recorders.Internal.NativeRecorder:get_frameSize () <C:\Users\RiccardoVailati\Desktop\IGOODI MOBILE UNITY\IgoodiUnityMobileApp\Assets\External Components\NatSuite\Plugins\Managed\Recorders\Internal\NativeRecorder.cs:19>
4. 07-06 15:59:25.539  9765  9812 E CRASH   :       #02 NatSuite.Recorders.MP4Recorder:get_frameSize () <C:\Users\RiccardoVailati\Desktop\IGOODI MOBILE UNITY\IgoodiUnityMobileApp\Assets\External Components\NatSuite\Plugins\Managed\Recorders\MP4Recorder.cs:21>
5. 07-06 15:59:25.539  9765  9812 E CRASH   :       #03 NatSuite.Recorders.Inputs.CameraInput:.ctor (NatSuite.Recorders.IMediaRecorder,NatSuite.Recorders.Clocks.IClock,UnityEngine.Camera[]) <C:\Users\RiccardoVailati\Desktop\IGOODI MOBILE UNITY\IgoodiUnityMobileApp\Assets\External Components\NatSuite\Plugins\Managed\Recorders\Inputs\CameraInput.cs:43>
6. 07-06 15:59:25.539  9765  9812 E CRASH   :       #04 ScreenshotCameraController:startRecording () <C:\Users\RiccardoVailati\Desktop\IGOODI MOBILE UNITY\IgoodiUnityMobileApp\Assets\Scenes\HomeScreen\ScreenshotCamera\ScreenshotCameraController.cs:61>
7. 07-06 15:59:25.539  9765  9812 E CRASH   :       #05 ScreenshotButtonScript:Update () <C:\Users\RiccardoVailati\Desktop\IGOODI MOBILE UNITY\IgoodiUnityMobileApp\Assets\Scenes\HomeScreen\ScreenshotCamera\ScreenshotButtonScript.cs:32>
8. 07-06 15:59:25.539  9765  9812 E CRASH   :       #06 (wrapper runtime-invoke) object:runtime_invoke_void__this__ (object,intptr,intptr,intptr)

The attached file contains the whole log, if that helps, and this is the code that fails:

Code (CSharp):

1.  
2.         recorder = new MP4Recorder(Screen.width, Screen.height, 24);
3.         RealtimeClock clock = new RealtimeClock();
4.         cameraInput = new CameraInput(recorder, clock, Camera.main);

Any idea on how to fix this?

 

Can you share runtime logs from any of your devices? And can you share your code? I'll need more information to be helpful.

 

Don't record at screen resolution. Use a fixed resolution like 1280x720.

 

exactly, lanare

may be your device resolution smaller then the recording code..like you device resolution is 840x400 and in recording code resolution have 1920x1080.

 

This is the crash log I'm getting:

Code (CSharp):

1. 2020-07-07 13:47:32.002 16471-16519/? A/.voedselsterre: runtime.cc:571] "UnityMain" prio=5 tid=12 Runnable
2. 2020-07-07 13:47:32.002 16471-16519/? A/.voedselsterre: runtime.cc:571]   at com.unity3d.player.UnityPlayer.nativeRender(Native method)
3. 2020-07-07 13:47:32.002 16471-16519/? A/.voedselsterre: runtime.cc:571]   at com.unity3d.player.UnityPlayer.access$300(unavailable:-1)
4. 2020-07-07 13:47:32.002 16471-16519/? A/.voedselsterre: runtime.cc:571]   at com.unity3d.player.UnityPlayer$e$1.handleMessage(unavailable:-1)
5. 2020-07-07 13:47:32.002 16471-16519/? A/.voedselsterre: runtime.cc:571]   at com.unity3d.player.UnityPlayer$e.run(unavailable:-1)
6. 2020-07-07 13:47:32.005 16471-16519/? A/.voedselsterre: runtime.cc:571] "UnityChoreographer" prio=5 tid=14 Native
7. 2020-07-07 13:47:32.009 16471-16519/? A/.voedselsterre: runtime.cc:571]   at com.unity3d.player.UnityPlayer.nativeRender(Native method)
8. 2020-07-07 13:47:32.028 16471-16519/? E/CRASH: pid: 16471, tid: 16519, name: UnityMain  >>> com.samhoud.voedselsterren <<<
9. 2020-07-07 13:47:32.029 16471-16519/? E/CRASH:       #04 NatSuite.Recorders.Inputs.CameraInput/<>c__DisplayClass11_1:<OnFrame>b__0 (UnityEngine.Rendering.AsyncGPUReadbackRequest) <0xd7>
10. 2020-07-07 13:47:32.674 16471-16519/? E/AndroidRuntime: FATAL EXCEPTION: UnityMain
11.     Process: com.company.newproject, PID: 16471
12.     java.lang.Error: *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ***
13.     Version '2019.4.0f1 (0af376155913)', Build type 'Release', Scripting Backend 'mono', CPU 'armeabi-v7a'
14.     Build fingerprint: 'samsung/starltexx/starlte:9/PPR1.180610.011/G960FXXU7CSK1:user/release-keys'
15.     Revision: '26'
16.     ABI: 'arm'
17.     Timestamp: 2020-07-07 13:47:32+0200
18.     pid: 16471, tid: 16519, name: UnityMain  >>> com.company.newproject <<<
19.     uid: 10085
20.     signal 6 (SIGABRT), code -6 (SI_TKILL), fault addr --------
21.         r0  00000000  r1  00004087  r2  00000006  r3  00000008
22.         r4  00004057  r5  00004087  r6  c76be534  r7  0000010c
23.         r8  0000000b  r9  0000005b  r10 e5570384  r11 e5571eb4
24.         ip  c76be4d0  sp  c76be520  lr  e59eeeb1  pc  e59e5eae
25.  
26.     backtrace:
27.           #00 pc 0001ceae  /system/lib/libc.so (abort+58) (BuildId: 231517bb8a85e40761bcb94d2bb2d7a1)
28.           #01 pc 00350fa7  /system/lib/libart.so (art::Runtime::Abort(char const*)+910) (BuildId: 889153dd67b992f69359e6923b98ce10)
29.           #02 pc 000071b3  /system/lib/libbase.so (android::base::LogMessage::~LogMessage()+494) (BuildId: 134fd9b8cd8f2f4ea7304be5857cf27b)
30.           #03 pc 001b952d  /system/lib/libart.so (art::IndirectReferenceTable::AbortIfNoCheckJNI(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&)+172) (BuildId: 889153dd67b992f69359e6923b98ce10)
31.           #04 pc 0023a2f3  /system/lib/libart.so (art::IndirectReferenceTable::GetChecked(void*) const+270) (BuildId: 889153dd67b992f69359e6923b98ce10)
32.           #05 pc 0023668f  /system/lib/libart.so (art::JavaVMExt::DecodeGlobal(void*)+10) (BuildId: 889153dd67b992f69359e6923b98ce10)
33.           #06 pc 00376b65  /system/lib/libart.so (art::Thread::DecodeJObject(_jobject*) const+168) (BuildId: 889153dd67b992f69359e6923b98ce10)
34.           #07 pc 0026b783  /system/lib/libart.so (art::JNI::GetObjectClass(_JNIEnv*, _jobject*)+402) (BuildId: 889153dd67b992f69359e6923b98ce10)
35.           #08 pc 00000da7  /data/app/com.company.newproject-g2f9Wewv3G6KRGfLIPZ9bA==/lib/arm/libNatCorder.so (NCFrameSize+30) (BuildId: 2567ae6a57fad542c249dc2cc296d227ffcfe4c3)
36.           #09 pc 00000e7f  /data/app/com.company.newproject-g2f9Wewv3G6KRGfLIPZ9bA==/lib/arm/libNatCorder.so (NCCommitFrame+50) (BuildId: 2567ae6a57fad542c249dc2cc296d227ffcfe4c3)
37.           #10 pc 0000aa0e  <anonymous:e18c3000>
38.  
39.     managed backtrace:
40.           #00 (wrapper managed-to-native) NatSuite.Recorders.Internal.Bridge:CommitFrame (intptr,intptr,long)
41.           #01 NatSuite.Recorders.Internal.NativeRecorder:CommitFrame (intptr,long) <0x2f>
42.           #02 NatSuite.Recorders.Internal.NativeRecorder:CommitFrame<byte> (byte[],long) <0x4b>
43.           #03 NatSuite.Recorders.MP4Recorder:CommitFrame<byte> (byte[],long) <0x43>
44.           #04 NatSuite.Recorders.Inputs.CameraInput/<>c__DisplayClass11_1:<OnFrame>b__0 (UnityEngine.Rendering.AsyncGPUReadbackRequest) <0xd7>
45.           #05 (wrapper runtime-invoke) <Module>:runtime_invoke_void__this___AsyncGPUReadbackRequest (object,intptr,intptr,intptr)
46.  
47.         at libc.abort(abort:58)
48.         at libart.art::Runtime::Abort(char const*)(Abort:910)
49.         at libbase.android::base::LogMessage::~LogMessage()(~LogMessage:494)
50.         at libart.art::IndirectReferenceTable::AbortIfNoCheckJNI(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&)(AbortIfNoCheckJNI:172)
51.         at libart.art::IndirectReferenceTable::GetChecked(void*) const(GetChecked:270)
52.         at libart.art::JavaVMExt::DecodeGlobal(void*)(DecodeGlobal:10)
53.         at libart.art::Thread::DecodeJObject(_jobject*) const(DecodeJObject:168)
54.         at libart.art::JNI::GetObjectClass(_JNIEnv*, _jobject*)(GetObjectClass:402)
55.         at libNatCorder.NCFrameSize(NCFrameSize:30)
56.         at libNatCorder.NCCommitFrame(NCCommitFrame:50)
57.         at NatSuite.Recorders.Internal.Bridge.CommitFrame (intptr,intptr,long)(Native Method)
58.         at NatSuite.Recorders.Internal.NativeRecorder.CommitFrame (intptr,long)(0x2f:47)
59.         at NatSuite.Recorders.Internal.NativeRecorder.CommitFrame<byte> (byte[],long)(0x4b:75)
60. 2020-07-07 13:47:32.674 16471-16519/? E/AndroidRuntime:     at NatSuite.Recorders.MP4Recorder.CommitFrame<byte> (byte[],long)(0x43:67)
61.         at NatSuite.Recorders.Inputs.CameraInput.<>c__DisplayClass11_1.<OnFrame>b__0 (UnityEngine.Rendering.AsyncGPUReadbackRequest)(0xd7:215)
62.         at <Module>.runtime_invoke_void__this___AsyncGPUReadbackRequest (object,intptr,intptr,intptr)(Native Method)

And this is the code that I changed from (commented) to what Menion-Leah said

Code (CSharp):

1.  
2. if (SystemInfo.supportsAsyncGPUReadback)
3.                 {
4.                     //AsyncGPUReadback.Request(frameBuffer, 0, request => {
5.                     //    if (pixelBuffer != null) {
6.                     //        request.GetData<byte>().CopyTo(pixelBuffer);
7.                     //        recorder.CommitFrame(pixelBuffer, timestamp);
8.                     //    }
9.                     //});
10.                  
11.                      NativeArray<byte> nativeArray = new NativeArray<byte>(frameBuffer.width * frameBuffer.height * 4, Allocator.TempJob);
12.                      AsyncGPUReadback.RequestIntoNativeArray(ref nativeArray, frameBuffer, 0, request =>
13.                      {
14.                          try
15.                          {
16.                              if (recorder != null)
17.                              {
18.                                  pixelBuffer = new byte[nativeArray.Length];
19.                                  nativeArray.CopyTo(pixelBuffer);
20.                                  recorder.CommitFrame(pixelBuffer, timestamp);
21.                              }
22.                          }
23.                          finally
24.                          {
25.                              nativeArray.Dispose();
26.                          }
27.                      });
28.                 }
29.  

 

hii , GJBooij

above crash happending because of not sufficient memory in device...suggestion is stop corountine after your requirment is complete..may execute your corountine that cause crash happening

 

The crash immediately happens when I stop recording, which I implemented as the documentation told me to do. (When releasing the record button) So, I'm not sure how I'm influencing this...

 

The `recorder` will never be `null`; it is a readonly field. So when recording stops and the recorder is invalidated, your readback code will still try to commit a frame, leading to an inevitable crash. See how the original `CameraInput` code handles stopping recording.

 

Hello,

I implemented NatCorder as a replacement for FFmpeg to export videos in my multiplatform project (Windows, iOS, OSX, Android), of which I am very satisfied. It works well in the editor (2019.4.2) but it crashes on the Windows standalone.

I have this crash report:

Code (CSharp):

1. NatCorder.dll caused an Access Violation (0xc0000005)
2.   in module NatCorder.dll at 0033:c67f2cd0
3.  
4. Stack Trace of Crashed Thread 1852:
5. 0x00007FFBC67F2CD0 (NatCorder) NCFinishWriting
6. 0x00007FFBC67F1100 (NatCorder) NCCreateMP4Recorder
7. 0x0000021C8E517ACE (Assembly-CSharp) NatSuite.Recorders.Internal.Bridge.CreateMP4Recorder()
8. 0x0000021C6AEBD7B3 (Assembly-CSharp) <>c__DisplayClass2_0.<.ctor>b__0()
9. 0x0000021C6AEB8F98 (Assembly-CSharp) NatSuite.Recorders.Internal.NativeRecorder..ctor()
10. 0x0000021C6AEB8E0B (Assembly-CSharp) NatSuite.Recorders.MP4Recorder..ctor()
11. 0x0000021C6AEB8BFB (Assembly-CSharp) NatCorderTest.Start()
12. 0x0000021C6AEB6FF0 (mscorlib) System.Object.runtime_invoke_void__this__()
13. 0x00007FFB4911D660 (mono-2.0-bdwgc) mono_get_runtime_build_info
14. 0x00007FFB490A28E2 (mono-2.0-bdwgc) mono_perfcounters_init
15. 0x00007FFB490AB93F (mono-2.0-bdwgc) mono_runtime_invoke
16. 0x00007FFB4C56E21D (UnityPlayer) UnityMain
17. 0x00007FFB4C56B603 (UnityPlayer) UnityMain
18. 0x00007FFB4C556EA9 (UnityPlayer) UnityMain
19. 0x00007FFB4C556F3E (UnityPlayer) UnityMain
20. 0x00007FFB4C554EB9 (UnityPlayer) UnityMain
21. 0x00007FFB4C2EAD7D (UnityPlayer) UnityMain
22. 0x00007FFB4C430567 (UnityPlayer) UnityMain
23. 0x00007FFB4C430603 (UnityPlayer) UnityMain
24. 0x00007FFB4C432A03 (UnityPlayer) UnityMain
25. ERROR: SymGetSymFromAddr64, GetLastError: 'Tentative d’accès à une adresse non valide.' (Address: 00007FFB4C1EB61E)
26. 0x00007FFB4C1EB61E (UnityPlayer) (function-name not available)
27. ERROR: SymGetSymFromAddr64, GetLastError: 'Tentative d’accès à une adresse non valide.' (Address: 00007FFB4C1EA37A)
28. 0x00007FFB4C1EA37A (UnityPlayer) (function-name not available)
29. ERROR: SymGetSymFromAddr64, GetLastError: 'Tentative d’accès à une adresse non valide.' (Address: 00007FFB4C1EE2EC)
30. 0x00007FFB4C1EE2EC (UnityPlayer) (function-name not available)
31. 0x00007FFB4C1F1DDB (UnityPlayer) UnityMain
32. ERROR: SymGetSymFromAddr64, GetLastError: 'Tentative d’accès à une adresse non valide.' (Address: 00007FF795CB11F2)
33. 0x00007FF795CB11F2 (BDNF) (function-name not available)
34. 0x00007FFBCBB07BD4 (KERNEL32) BaseThreadInitThunk
35. 0x00007FFBCCFCCE51 (ntdll) RtlUserThreadStart

Is there any settings that I don't know about? Thank you for your time.

 

Hi there. Can you share your recording code?

 

I created this very simple test class for the editor attached to the main Camera in my scene. Once built, only the recorder and clock are supposed to be instanciated, and the code in Update() is never reached due to record and isRecording both being false.

Code (CSharp):

1. public class NatCorderTest : MonoBehaviour
2. {
3.     private MP4Recorder recorder;
4.     private CameraInput cameraInput;
5.     private FixedIntervalClock clock;
6.  
7.     public bool record = false;
8.  
9.     private bool isRecording = false;
10.  
11.     public string path;
12.  
13.     private void Start()
14.     {
15.         recorder = new MP4Recorder(1920, 1080, 60);
16.         clock = new FixedIntervalClock(60, true);
17.     }
18.  
19.     private void Update()
20.     {
21.         if (record && !isRecording)
22.         {
23.             Debug.Log("Start Recording");
24.             cameraInput = new CameraInput(recorder, clock, Camera.main);
25.             isRecording = true;
26.         }
27.         else if (!record && isRecording)
28.         {
29.             Debug.Log("Stop Recording");
30.             isRecording = false;
31.             cameraInput.Dispose();
32.             GetVideoPath();
33.         }
34.     }
35.  
36.     private async void GetVideoPath()
37.     {
38.         Debug.Log("Start Awaiting");
39.         path = await recorder.FinishWriting();
40.         Debug.Log("Stop Awaiting");
41.         Debug.Log(path);
42.     }
43. }

Thank you for your help!

 

Your code's state management is confusing, as you have a `record` and `isRecording` variable. I recommend rewriting this code to have simpler state management (don't start and stop recording in Update). Follow the included examples for guidance.

 

Sorry for this bad bunch of code, I made it quicly just to figure out why my real project was crashing. I hope this new code snippet is less confusing:

Code (CSharp):

1. public class NatCorderTest : MonoBehaviour
2. {
3.     private MP4Recorder recorder;
4.     private CameraInput cameraInput;
5.     private FixedIntervalClock clock;
6.  
7.     string path;
8.  
9.     private void Start()
10.     {
11.         recorder = new MP4Recorder(1920, 1080, 60);
12.         clock = new FixedIntervalClock(60, true);
13.         StartRecording();
14.     }
15.  
16.     void StartRecording()
17.     {
18.         Debug.Log("Start Recording");
19.         cameraInput = new CameraInput(recorder, clock, Camera.main);
20.         Invoke("StopRecording", 4f);
21.     }
22.  
23.     void StopRecording()
24.     {
25.         Debug.Log("Stop Recording");
26.         cameraInput.Dispose();
27.         GetVideoPath();
28.     }
29.  
30.     private async void GetVideoPath()
31.     {
32.         Debug.Log("Start Awaiting");
33.         path = await recorder.FinishWriting();
34.         Debug.Log("Stop Awaiting");
35.         Debug.Log(path);
36.     }
37. }

It also crashes just after launch. The line causing the crash is

Code (CSharp):

1.  recorder = new MP4Recorder(1920, 1080, 60);

and I think the problem is caused by this part in the MP4Recorder constructor:

Code (CSharp):

1. new NativeRecorder((callback, context) => Bridge.CreateMP4Recorder(width, height, framerate, bitrate, keyframeInterval, sampleRate, channelCount, Utility.GetPath(@".mp4"), callback, context))

If I look at the stacktrace, I also find suspicious that the last function called is NCFinishWriting.

When I run my code with only the Start function, it also crashes and I get the same stack trace.

 

Are you able to run the ReplayCam example? If ReplayCam crashes, post the stack trace. Nothing should be calling `FinishWriting` in the constructor.

 

I can run ReplayCam and it works fine in the editor, but when I build it and run the example, it crashes just like my own test and I get the same stack error:

Code (CSharp):

1. NatCorder.dll caused an Access Violation (0xc0000005)
2.   in module NatCorder.dll at 0033:bf672cd0.
3.  
4. Stack Trace of Crashed Thread 26492:
5. 0x00007FFBBF672CD0 (NatCorder) NCFinishWriting
6. 0x00007FFBBF671100 (NatCorder) NCCreateMP4Recorder
7. 0x0000011A4BD75F0E (Assembly-CSharp) NatSuite.Recorders.Internal.Bridge.CreateMP4Recorder()
8. 0x0000011A4BD5D173 (Assembly-CSharp) <>c__DisplayClass2_0.<.ctor>b__0()
9. 0x0000011A4BD5BE18 (Assembly-CSharp) NatSuite.Recorders.Internal.NativeRecorder..ctor()
10. 0x0000011A4BD5BC8B (Assembly-CSharp) NatSuite.Recorders.MP4Recorder..ctor()
11. 0x0000011A4BD5B7BB (Assembly-CSharp) NatSuite.Examples.ReplayCam.StartRecording()
12. 0x0000011A4BD5B559 (UnityEngine.CoreModule) UnityEngine.Events.InvokableCall.Invoke()
13. 0x0000011A4BD56A2B (UnityEngine.CoreModule) UnityEngine.Events.UnityEvent.Invoke()
14. 0x0000011A4BC2BE93 (Assembly-CSharp) <Countdown>d__10.MoveNext()
15. 0x0000011A4BC2ACEC (UnityEngine.CoreModule) UnityEngine.SetupCoroutine.InvokeMoveNext()
16. 0x0000011A4BC2AE17 (UnityEngine.CoreModule) <Module>.runtime_invoke_void_object_intptr()
17. 0x00007FFB4E35D660 (mono-2.0-bdwgc) mono_get_runtime_build_info
18. 0x00007FFB4E2E28E2 (mono-2.0-bdwgc) mono_perfcounters_init
19. 0x00007FFB4E2EB93F (mono-2.0-bdwgc) mono_runtime_invoke
20. 0x00007FFB3FCBE21D (UnityPlayer) UnityMain
21. 0x00007FFB3FCBB603 (UnityPlayer) UnityMain
22. 0x00007FFB3FCA1BC0 (UnityPlayer) UnityMain
23. 0x00007FFB3FA3AD7D (UnityPlayer) UnityMain
24. 0x00007FFB3FB80567 (UnityPlayer) UnityMain
25. 0x00007FFB3FB80603 (UnityPlayer) UnityMain
26. 0x00007FFB3FB82A03 (UnityPlayer) UnityMain
27. ERROR: SymGetSymFromAddr64, GetLastError: 'Tentative d’accès à une adresse non valide.' (Address: 00007FFB3F93B61E)
28. 0x00007FFB3F93B61E (UnityPlayer) (function-name not available)
29. ERROR: SymGetSymFromAddr64, GetLastError: 'Tentative d’accès à une adresse non valide.' (Address: 00007FFB3F93A37A)
30. 0x00007FFB3F93A37A (UnityPlayer) (function-name not available)
31. ERROR: SymGetSymFromAddr64, GetLastError: 'Tentative d’accès à une adresse non valide.' (Address: 00007FFB3F93E2EC)
32. 0x00007FFB3F93E2EC (UnityPlayer) (function-name not available)
33. 0x00007FFB3F941DDB (UnityPlayer) UnityMain
34. ERROR: SymGetSymFromAddr64, GetLastError: 'Tentative d’accès à une adresse non valide.' (Address: 00007FF73EAB11F2)
35. 0x00007FF73EAB11F2 (BDNF) (function-name not available)
36. 0x00007FFBCBB07BD4 (KERNEL32) BaseThreadInitThunk
37. 0x00007FFBCCFCCE51 (ntdll) RtlUserThreadStart

Also, I tried to build for Mac and it worked well.

 

Okay I'll try to reproduce it on my end and go from there. Thanks for bringing this to my attention.

 

Thank you for looking into it.

 

Hello. @Lanre i need some help.

App crashing sometimes with Fatal Exception.

Fatal Exception: java.lang.Error
FATAL EXCEPTION [MP4Recorder Audio Thread]
(it is not always on the same line of code, every time on different line)

I still have no clue why is this happening. Have you(or others) idea, what can be wrong to get this exception?
Thanks.

 

Can you share your recording code and the full logs? Not sure what's going on here, and I haven't encountered an error like this in a long while.

 

Hello @Lanre

I want to display camera preview in a rect with rounded corners (UI Mask component). I'm using RawImage to display the camera texture, but since the CameraPreview shader doesn't support UI Masking I've added the code for it to work:

Code (CSharp):

1.           // required for UI.Mask
2.          _StencilComp ("Stencil Comparison", Float) = 8
3.          _Stencil ("Stencil ID", Float) = 0
4.          _StencilOp ("Stencil Operation", Float) = 0
5.          _StencilWriteMask ("Stencil Write Mask", Float) = 255
6.          _StencilReadMask ("Stencil Read Mask", Float) = 255
7.          _ColorMask ("Color Mask", Float) = 15

Code (CSharp):

1.         // required for UI.Mask
2.          Stencil
3.          {
4.              Ref [_Stencil]
5.              Comp [_StencilComp]
6.              Pass [_StencilOp]
7.              ReadMask [_StencilReadMask]
8.              WriteMask [_StencilWriteMask]
9.          }
10.           ColorMask [_ColorMask]

However, as soon as I added those snippets, the Rotation and Scale properties stopped working on the shader, resulting in wrongly rotated or flipped displayed camera texture. Could you guide me how to fix this?

 

I'm not sure why this would happen. I recommend creating a thread in the shaders sub-forum to see if anyone more knowledgeable on Unity shaders can help; this is not my area of specialty, and the shader included with the example is purely for demonstration purposes. It isn't built to support more advanced UI effects.

 

Hyy Lanre,

what i m telling you actuly this hair difference is show when plying mp4 recorded video...
help me

 

Hii , Lanre

recording sometime black mp4 ...nothing into vision..only black screen..
so any solutin.??

 

I don't know the inner workings of your hair shader, so I don't know why it renders differently when rendering to texture vs. rendering to the screen.

Can you share more information on this? Can you post your recording code, logs, and anything else that might be useful? Are you able to reproduce this behaviour in the ReplayCam example?

 

its happening with sometime in 4GB RAM device only..

 

Hi, i am using arcore and natCorder. I am trying to get better fps when i recording. It seems like my fps dropping when i hit record. I am using match frame rate unchecked on ar session and i set recording fps on 60. This is a screenshoot from my profiler when i start to record. What can i do to get better, fixed 60 fps ? (my fps looks normal while not recording)

 

I cant share code, im sorry for that. but i can provide more information.

Crash caoused by
java.lang.IllegalStateException
writeSampleData returned an error

com.olokobayusuf.natcorder.MP4Recorder$11.run (MP4Recorder.java:333)

 

It looks like all the time is being spent reading back the pixel data from the GPU. This is currently a limitation when using ARCore because it does not support Vulkan. Vulkan does not have this problem because it supports asynchronous readbacks from the GPU. You can read more on it here.

 

You are using an outdated version of NatCorder. Upgrade to 1.7.2 on the Asset Store.

 

Hi, have some problem with Android 8.1 Phones. Trying to record video in my app but its Crashes after 1-2 seconds What can i do to fix it?
Here is the log

Spoiler: Logcat

07-15 22:56:52.697 29626 29670 V NatSuite: NatCorder: Prepared MP4Recorder video encoder with format: {width=1280, height=720, bitrate=5909760, mime=video/avc, frame-rate=30.0, i-frame-interval=3, color-format=2130708361}
07-15 22:56:52.700 545 545 I MediaPlayerService: MediaPlayerService::getOMX
07-15 22:56:52.701 29626 29914 I OMXClient: MuxOMX ctor
07-15 22:56:52.703 542 4020 I OMXMaster: makeComponentInstance(OMX.google.aac.encoder) in mediacodec process
07-15 22:56:52.716 29626 29670 V NatSuite: NatCorder: Prepared MP4Recorder audio encoder with format: {mime=audio/mp4a-latm, aac-profile=2, channel-count=2, channel-mask=12, max-input-size=16384, bitrate=64000, sample-rate=48000}
07-15 22:56:52.719 542 26821 D GraphicBufferSource: setting dataspace: 0x104
07-15 22:56:52.720 542 3413 E OMXNodeInstance: getParameter(21e0a88:qcom.encoder.avc, ParamConsumerUsageBits(0x6f800004)) ERROR: UnsupportedIndex(0x8000101a)
07-15 22:56:52.720 542 29317 D GraphicBufferSource: requesting color aspects (R:2(Limited), P:1(BT709_5), M:1(BT709_5), T:3(SMPTE170M))
07-15 22:56:53.022 542 4020 D GraphicBufferSource: got buffer with new dataSpace #104
07-15 22:56:53.023 542 29911 I OMX-VENC: open Color conv for RGBA888 W: 1280, H: 720
07-15 22:56:53.023 29626 29905 D ACodec : dataspace changed to 0x10c10000 (R:2(Limited), P:1(BT709_5), M:1(BT709_5), T:3(SMPTE170M)) (R:2(Limited), S:1(BT709), T:3(SMPTE_170M))
07-15 22:56:53.078 29626 29920 I MPEG4Writer: limits: 4294967295/0 bytes/us, bit rate: -1 bps and the estimated moov size 3195 bytes
07-15 22:56:53.080 29626 29925 I MPEG4Writer: setStartTimestampUs: 0
07-15 22:56:53.080 29626 29925 I MPEG4Writer: Earliest track starting time: 0
07-15 22:56:53.080 29626 29924 I MPEG4Writer: setStartTimestampUs: 0
07-15 22:56:53.092 29626 29924 W MPEG4Writer: 0-duration samples found: 1
07-15 22:56:54.192 29626 29702 F art : art/runtime/indirect_reference_table.cc:132] JNI ERROR (app bug): local reference table overflow (max=512)
07-15 22:56:54.192 29626 29702 F art : art/runtime/indirect_reference_table.cc:132] local reference table dump:
07-15 22:56:54.192 29626 29702 F art : art/runtime/indirect_reference_table.cc:132] Last 10 entries (of 512):
07-15 22:56:54.192 29626 29702 F art : art/runtime/indirect_reference_table.cc:132] 511: 0x132ed670 java.lang.Class<api.natsuite.natcorder.MP4Recorder>
07-15 22:56:54.192 29626 29702 F art : art/runtime/indirect_reference_table.cc:132] 510: 0x1357d000 float[] (512 elements)
07-15 22:56:54.192 29626 29702 F art : art/runtime/indirect_reference_table.cc:132] 509: 0x132ed670 java.lang.Class<api.natsuite.natcorder.MP4Recorder>
07-15 22:56:54.192 29626 29702 F art : art/runtime/indirect_reference_table.cc:132] 508: 0x1357c000 float[] (512 elements)
07-15 22:56:54.192 29626 29702 F art : art/runtime/indirect_reference_table.cc:132] 507: 0x132ed670 java.lang.Class<api.natsuite.natcorder.MP4Recorder>
07-15 22:56:54.192 29626 29702 F art : art/runtime/indirect_reference_table.cc:132] 506: 0x13577000 float[] (512 elements)
07-15 22:56:54.192 29626 29702 F art : art/runtime/indirect_reference_table.cc:132] 505: 0x132ed670 java.lang.Class<api.natsuite.natcorder.MP4Recorder>
07-15 22:56:54.192 29626 29702 F art : art/runtime/indirect_reference_table.cc:132] 504: 0x13576000 float[] (512 elements)
07-15 22:56:54.192 29626 29702 F art : art/runtime/indirect_reference_table.cc:132] 503: 0x132ed670 java.lang.Class<api.natsuite.natcorder.MP4Recorder>
07-15 22:56:54.192 29626 29702 F art : art/runtime/indirect_reference_table.cc:132] 502: 0x13574000 float[] (512 elements)
07-15 22:56:54.192 29626 29702 F art : art/runtime/indirect_reference_table.cc:132] Summary:
07-15 22:56:54.192 29626 29702 F art : art/runtime/indirect_reference_table.cc:132] 256 of java.lang.Class (1 unique instances)
07-15 22:56:54.192 29626 29702 F art : art/runtime/indirect_reference_table.cc:132] 256 of float[] (512 elements) (256 unique instances)
07-15 22:56:54.192 29626 29702 F art : art/runtime/indirect_reference_table.cc:132]
07-15 22:56:54.678 29626 29702 F art : art/runtime/runtime.cc:422] Runtime aborting...
07-15 22:56:54.678 29626 29702 F art : art/runtime/runtime.cc:422] Aborting thread:
07-15 22:56:54.678 29626 29702 F art : art/runtime/runtime.cc:422] "Thread-14" prio=10 tid=64 Runnable
07-15 22:56:54.678 29626 29702 F art : art/runtime/runtime.cc:422] | group="" sCount=0 dsCount=0 obj=0x12e37dc0 self=0x71abdba000
07-15 22:56:54.678 29626 29702 F art : art/runtime/runtime.cc:422] | sysTid=29702 nice=-19 cgrp=default sched=0/0 handle=0x71c0605450
07-15 22:56:54.678 29626 29702 F art : art/runtime/runtime.cc:422] | state=R schedstat=( 2348798664 469459003 20061 ) utm=189 stm=44 core=3 HZ=100
07-15 22:56:54.678 29626 29702 F art : art/runtime/runtime.cc:422] | stack=0x71c04ff000-0x71c0501000 stackSize=1053KB
07-15 22:56:54.678 29626 29702 F art : art/runtime/runtime.cc:422] | held mutexes= "abort lock" "mutator lock"(shared held)
07-15 22:56:54.678 29626 29702 F art : art/runtime/runtime.cc:422] native: #00 pc 000000000047ef3c /system/lib64/libart.so (_ZN3art15DumpNativeStackERNSt3__113basic_ostreamIcNS0_11char_traitsIcEEEEiP12BacktraceMapPKcPNS_9ArtMethodEPv+220)
07-15 22:56:54.678 29626 29702 F art : art/runtime/runtime.cc:422] native: #01 pc 000000000047ef38 /system/lib64/libart.so (_ZN3art15DumpNativeStackERNSt3__113basic_ostreamIcNS0_11char_traitsIcEEEEiP12BacktraceMapPKcPNS_9ArtMethodEPv+216)

 

Are you using NatCorder 1.7.2? If not, make sure to upgrade. This should have been fixed recently.

 

@Lanre
Hi i use Wikitude AR plugin, and i need to record video with AR object and real world(background)
For this i have AR camera which draw AR model in depth camera mode, also i create one more camera to draw real world at plane.This is my AR camera http://prntscr.com/tj6s3z
This is camera which draw real world -> http://prntscr.com/tj6rn0
and when i record video, i use both camera.

Code (CSharp):

1.  
2. _clock = new RealtimeClock();
3. _videoRecorder = new MP4Recorder(
4.             1080,
5.             1920,
6.             30,
7.             AudioSettings.outputSampleRate,
8.             2
9.         );
10. _cameraInput = new CameraInput(_videoRecorder,_clock,Camera.allCameras);
11. _audioInput = new AudioInput(_videoRecorder,_clock,Camera.main.GetComponent<AudioListener>());
12.  

my video at android(perfect) ->

my video at ios(problems, the video seems to be cropped) ->

How to fix it, please help
i use latest natcorder, unity 2019.3.15 and wkitude 9.2

 

You'll have to reach out to Wikitude's support for this one. CameraInput is simply rendering the camera view to a texture. It is likely the case that Wikitude is overriding how the camera renders, which is why it is offset (cropped). This would have nothing to do with NatCorder, and you can verify this by looking at the CameraInput sources; all it is doing is re-rendering the camera to a RenderTexture by calling `camera.Render`. The camera is expected to render normally, as it did when rendering to the screen.

 

my project is on Unity 2018.4.23. Natcorder v1.7.2 is for 2019.3.11 or newer.
will it be big problem? i cant update app. there is a lot of plugins which is conflicting somehow.

 

Our application mixes game audio with microphone input, and because of device-specific mic latency we mix in memory and commit all the audio samples at the end of the recording, calling CommitSamples() with a timestamp of 0.

This works great in Natcorder 1.5.1 but we're trying to upgrade to 1.7.2 so we can use AsyncGPUReadback, and we find that CommitFrame() starts to take more than 1 second to return if we don't commit audio samples regularly.

I was always surprised that our technique worked, but it was useful. Has there been a specification change?

 

NatCorder will work on Unity 2018.3+. You can use Unity 2019 to download the package, then copy it to your 2018 project.

 

Are you on Windows? I have no clue why, but on Windows if you don't commit audio frames in step with video frames, then it starts to lag terribly. I have absolutely no clue why this happens.

 

Hey,
we are using Unity 2018.4.24f, NatCorder 1.7.0 and are recording WebCamTextureInput.

When the user puts the app in background or the app loses focus, we are pausing the recording.

Thats pretty much how we are doing it:

Code (CSharp):

1. protected override void OnApplicationPause(bool pauseStatus)
2.         {
3.             if (_recording)
4.             {
5.                 if (pauseStatus)
6.                 {
7.                     DisposeInputs();
8.                 }
9.                 else
10.                 {
11.                     StartInputs();
12.                 }
13.             }
14.         }
15.  
16.         protected override void OnApplicationFocus(bool hasFocus)
17.         {
18.             if (_recording)
19.             {
20.                 if (hasFocus)
21.                 {
22.                     StartInputs();
23.                 }
24.                 else
25.                 {
26.                     DisposeInputs();
27.                 }
28.             }
29.         }
30.  
31. private void StartInputs()
32. {
33.    _recordingClock.Paused = false;
34.    _webCamTextureInput = new WebCamTextureInput(_videoRecorder, _recordingClock, _activeCameraTexture);
35.    _audioInput = new AudioInput(_videoRecorder, _recordingClock, microphoneSource, true);
36. }
37.  
38. private void DisposeInputs()
39. {
40.    _audioInput.Dispose();
41.    _webCamTextureInput.Dispose();
42.    _recordingClock.Paused = true;
43. }
44.  
45.  
46.  

That works on Android, but not on iOs.

When leaving the app on iOs and reopneing the app from the background, we are getting spammed with this:

Code (CSharp):

1. -> applicationWillResignActive()
2. -> applicationDidEnterBackground()
3. 2020-07-20 16:38:22.819321+0200 app[5112:1324024] [Snapshotting] Snapshotting a view (0x10d0a5800, UIKeyboardImpl) that has not been rendered at least once requires afterScreenUpdates:YES.
4. -> applicationWillEnterForeground()
5. 2020-07-20 16:38:27.352353+0200 app[5112:1325830] [] nw_read_request_report [C9] Receive failed with error "Software caused connection abort"
6. 2020-07-20 16:38:27.353032+0200 app[5112:1325830] [] nw_read_request_report [C9] Receive failed with error "Software caused connection abort"
7. -> applicationDidBecomeActive()
8. 2020-07-20 16:38:27.645525+0200 app[5112:1324024] NatCorder Error: MP4Recorder failed to record video frame for time 2619768600 due to invalid recorder state: 3
9. 2020-07-20 16:38:27.782399+0200 app[5112:1324024] NatCorder Error: MP4Recorder failed to record video frame for time 2756612200 due to invalid recorder state: 3
10. 2020-07-20 16:38:27.813959+0200 app[5112:1324024] NatCorder Error: MP4Recorder failed to record video frame for time 2788182700 due to invalid recorder state: 3
11. 2020-07-20 16:38:27.848963+0200 app[5112:1324024] NatCorder Error: MP4Recorder failed to record video frame for time 2823182900 due to invalid recorder state: 3
12. 2020-07-20 16:38:27.880851+0200 app[5112:1324024] NatCorder Error: MP4Recorder failed to record video frame for time 2855077900 due to invalid recorder state: 3
13. 2020-07-20 16:38:27.915029+0200 app[5112:1324024] NatCorder Error: MP4Recorder failed to record video frame for time 2889262900 due to invalid recorder state: 3
14. 2020-07-20 16:38:28.014888+0200 app[5112:1324024] NatCorder Error: MP4Recorder failed to record video frame for time 2989118800 due to invalid recorder state: 3
15. 2020-07-20 16:38:28.048782+0200 app[5112:1324024] NatCorder Error: MP4Recorder failed to record video frame for time 3023013500 due to invalid recorder state: 3
16. 2020-07-20 16:38:28.079793+0200 app[5112:1324024] NatCorder Error: MP4Recorder failed to record video frame for time 3054017900 due to invalid recorder state: 3
17. 2020-07-20 16:38:28.115865+0200 app[5112:1324024] NatCorder Error: MP4Recorder failed to record video frame for time 3090077500 due to invalid recorder state: 3
18. 2020-07-20 16:38:28.146360+0200 app[5112:1324024] NatCorder Error: MP4Recorder failed to record video frame for time 3120600100 due to invalid recorder state: 3
19. 2020-07-20 16:38:28.180375+0200 app[5112:1324024] NatCorder Error: MP4Recorder failed to record video frame for time 3154616900 due to invalid recorder state: 3

I mean I know what the error is telling me, however we are doing everything we can do to prevent that.

 

If the app loses focus, the native encoder will stop. This is an OS restriction on both iOS and Android. As such, you cannot pause recording then suspend the app--recording will fail.

 

Yes this is on Windows. But I'm puzzled why it only happens after upgrading from 1.5.1, if the native encoder is presumably the same?

 

Do you know if this has an issue number in Unity's Issue Tracker? This is another problem we face in upgrading from 1.5.1 to 1.7.2

 

I remember this behaviour going all the way back. Nothing has changed in the native Windows implementation since 1.5.1.

 

No I don't think this was posted on Issue Tracker.