Next Gen Recorder - News & Support

You are welcome

 

To two separate video files? Unfortunately that is not supported. If that would be supported, other encoding process would probably be done in software (slower performance) instead of using the hardware encoder.

You could render both cameras to a texture and record both cameras to a single video file by recording the texture.

 

Hi! Long time no see. First thing that comes to my mind is that are you sure are running it on Metal? Those examples are only for Metal (not for OpenGL). I haven't tried the examples with 2020 so not sure if something has changed in Unity. I'll try to give it a spin at some point.

Also please remember that if you add more than one Camera to the scene, you should start using the last example from the integration documentation or you might be sending multiple render passes to the encoder which will result an interesting video

EDIT: Did a quick test with 2020.1.7f1, not exactly the same version but it worked fine. Are you using the same Forward Renderer on Mac? If not, are you sure that you have applied the Recording Pass to Renderer Features of both Forward Renderers? And are you sure you have UrpRecorder in scene hierarchy on both platforms?

 

It has not been removed since it has never been inside the Asset Store package. You should still find it from this thread

 

Could you contact me at "support [at] pmjo [dot] org".

 

Hi! So you want to render something different to the video than you are rendering on to the screen, did I understand correctly?

For that you should have multiple cameras. One camera (or more) for rendering to the offscreen texture and one or more cameras rendering to the screen. It somehow sounds like there is something wrong with your camera setup but without knowing more about your project I don't know how to help. Feel free to contact me at "support [at] pmjo [dot] org" and tell a bit more about your project and how you would like it to work.

If you just want to record the screen, you don't need any of the above. Just add Metal Screen Recorder component on your main camera.

 

1) You can change video scale (http://www.pmjo.org/nextgenrecorder/usage.html#downscaling_the_video) but if you want to change both width and height you must change the source texture size. Keep in mind that you cannot change it during recording but you can change the source texture size when not recording.
2) No, with Metal Screen Recorder you can only change the scale

For your case I would use two textures. One is the full screen texture that you want to render on to the screen and the other texture is the actual recording texture to which you blit the area that you want to record.

 

You cannot apply AudioSourceRecorder on multiple Audio Sources. That won't work. So you want to record some audio sources but not all? If you want to record all audio sources, don't just add AudioSourceRecorder to anywhere and all mixer output is recorded. For recording only some audio sources I cannot think anything else than using the Audio Mixer Recorder or mixing the audio sources manually in c# and feeding that to NGR like AudioSourceRecorder does.

 

I'm not sure if I understood correctly. So you would always want to have volume 1.0 in the video? even someone has lowered volume?

 

And the problem with the Audio Mixer Recorder is that it did not work when built inside a library? So Playback and Record are otherwise identical but other one is with lower volume? We could probably create a duplicate of the default audio recorder and before passing the data to NGR we could multiply the samples with the inverse of the volume.

 

Oh yeah, I think I remember when we tried to make it work from external library at some point.

Instead of using default NGR audio recording you could use the component below to first capture the audio and then set the volume to the Audio Listener manually. So instead of setting AudioListener.volume you would set the volume for this component:

Code (CSharp):

1. using UnityEngine;
2. using System.Runtime.InteropServices;
3.  
4. namespace pmjo.NextGenRecorder
5. {
6.     [AddComponentMenu("Next Gen Recorder/Audio Recorder With Volume")]
7.     public class AudioRecorderWithVolume : Recorder.AudioRecorderBase
8.     {
9.         [Range(0, 1.0f)]
10.         public float volume = 1.0f;
11.         private int m_SampleRate = 0;
12.  
13.         void Awake()
14.         {
15.             AudioConfiguration config = AudioSettings.GetConfiguration();
16.  
17.             if (config.sampleRate >= 8000)
18.                 m_SampleRate = config.sampleRate;
19.             else
20.                 Debug.Log("Only sample rate >= 8000 is supported");
21.         }
22.  
23.         void OnAudioFilterRead(float[] data, int channels)
24.         {
25.             // Capture audio before volume changed
26.             if (m_SampleRate > 0)
27.             {
28.                 GCHandle handle = GCHandle.Alloc(data, GCHandleType.Pinned);
29.  
30.                 if (handle.IsAllocated)
31.                 {
32.                     AppendInterleavedAudio(handle.AddrOfPinnedObject(), data.Length, channels, m_SampleRate);
33.                     handle.Free();
34.                 }
35.             }
36.  
37.             // Apply volume to audio listener
38.             for (int i = 0; i < data.Length; i++)
39.             {
40.                 data[i] = data[i] * volume;
41.             }
42.         }
43.     }
44. }
45.  

Just drop this script on your AudioListener (usually the main camera). Only tested it on a Mac

 

Nice! You are welcome

 

Yeah, Unity Audio Mixer is quite limited and many AAA titles use WWise instead.

 

Audio Mixer Recorder only works with Next Gen Recorder. If you are using Next Gen Recorder, you can contact me through support@pmjo.org

 

By default the Simple Recorder only playbacks the recorded file on iOS. If you look at the actual recording component in Simple Recorder scene you can see that the saving of the file is commented out for iOS and a video player is opened instead. Just uncomment the video saving part and comment out the video playback.

Keep in mind that when doing a more complex recording like an offscreen recorder the integration method depends on the used Rendering Pipeline (Built-in Rendering Pipeline, Scriptable Render Pipeline). You can find some examples at the end of the Integration guide.

 

Hmm. Sounds strange. I think there has been people using NGR and ARKit successfully but I don't remember if that required some kind of custom integration. What Render Pipeline you are using? Do you have any other graphics plugins in your project? Best would be if you could contact me through email (support [at] pmjo.org) and provide a small test project for replicating the issue.

EDIT: On second page of this thread I have actually posted that I tried ARKit and it worked fine. That was in 2020 so it can be that it has changed somehow. What ARKit library or plugin you are using?

 

Hi again. I was able to replicate your "Failed to capture render target." issue with AR Foundation 4.2. Earlier it has just worked out of the box so something has definitely changed in it. I will try to investigate it there would be se some way to record it, for example with a custom integration.

EDIT: The AR Foundation plugin seems to somehow mess up the native rendering command buffer but I already found a workaround to make the recording work with it. The key is to render the AR Camera to a texture and then blit that to the screen. Then NGR is able to capture the screen output. I will try to create some kind of custom integration script for this case, hopefully during next weekend.

 

Here is one way to go around the problem. The key is to record a texture instead of the screen. So, first render everything to a texture, record the texture and then blit it on to the screen.

Code (CSharp):

1. using UnityEngine;
2. using UnityEngine.Rendering;
3. using pmjo.NextGenRecorder;
4.  
5. [RequireComponent(typeof(Camera))]
6. public class ArCameraRecorder : Recorder.VideoRecorderBase
7. {
8.     public Camera arCamera;
9.  
10.     private Camera m_Camera;
11.     private RenderTexture m_RenderTexture;
12.     private CommandBuffer m_CommandBuffer;
13.  
14.     void Awake()
15.     {
16.         if (arCamera == null)
17.         {
18.             Debug.LogError("No AR Camera set. Please set AR Camera for ArCameraRecorder");
19.             return;
20.         }
21.  
22.         if (!Recorder.IsSupported)
23.         {
24.             Debug.LogWarning("Next Gen Recorder not supported on this platform");
25.             return;
26.         }
27.  
28.         m_RenderTexture = new RenderTexture(Screen.width, Screen.height, 32, RenderTextureFormat.Default);
29.         m_RenderTexture.Create();
30.  
31.         arCamera.targetTexture = m_RenderTexture;
32.  
33.         RecordingTexture = m_RenderTexture;
34.  
35.         m_Camera = GetComponent<Camera>();
36.         m_Camera.depth = arCamera.depth + 1;
37.         m_Camera.clearFlags = CameraClearFlags.Nothing;
38.         m_Camera.cullingMask = 0;
39.  
40.         m_CommandBuffer = new CommandBuffer();
41.         m_CommandBuffer.Blit(m_RenderTexture, BuiltinRenderTextureType.CameraTarget);
42.         CommandBufferBlitRecordingTexture(m_CommandBuffer);
43.         m_Camera.AddCommandBuffer(CameraEvent.AfterEverything, m_CommandBuffer);
44.     }
45. }
46.  

I tested this script in Unity AR Foundation 4.2 SimpleAR scene and was able to record it successfully.

Usage:
1. First I dropped the SimpleRecorder prefab to the scene for being able to test the recording easily
2. Then I created a new gameobject called "RecordingCamera" and dropped the ArCameraRecorder script on it
3. Then I drag and dropped the "AR Camera" from "AR Session Origin" to the "Ar Camera" slot on the "RecordingCamera". That's it.

So, it seems the latest AR Foundation does not work directly with NGR but with custom integration you can go around it. The above example custom integration works for the mentioned test scene but for some other project you might need a different kind of integration. I hope this helps!

 

Hi! The problem is that you try to run the code with URP. The description of the example code mentions that it is for Built-In render pipeline The provided custom integration snippets in the documentation are just examples how you can feed different data to the encoder (instead of just screen). With custom integrations one needs full understanding what happens under the hood and how to hook to Unity command buffers on that pipeline. Unfortunately I don't have an offscreen camera example for you at the moment but I could try to create one at some point. Meanwhile you could try to investigate the URP examples and try to think if they could be modified to record only some specific "offscreen" camera. The key is that you only do context.ExecuteCommandBuffer(m_CommandBuffer) for the camera you wish to record and only once per frame.

 

Then it might be a bit challenging. Can you explain me what you exactly want to do do? for example, how many cameras you would be using? one for screen, one for offscreen? would the offscreen camera have post processing?

 

This should be easy. Does anything need to be hidden from the square are (like UI) or is it just a partial screen capture?

 

A bit but did not get far, just a flickering AR project even there was not recording yet. I think the AR Foundation I tried is not compatible with URP. What AR Foundation version you are using? Haven't played with AR kit for a long time.

 

On the second page of this thread there is an additional plugin that I created for recording a specific node in Audio Mixer. Never totally finished it so that I would put it inside NGR but at least it worked then. For that I needed to write a native audio plugin in c (or maybe it was c++).