Next Gen Recorder - News & Support

May I ask why you need to access the audio samples? And on what platforms?

 

A customer just reported that Next Gen Recorder is giving an error "failed to capture render target" with URP 14. I have confirmed this to happen and it seems URP has changed so radically from 12 that I don't have a fast fix or any fix for this at the moment.

I will investigate the issue but for now I advice not to buy NGR if you are planning to use URP14. And please always try the FREE version before buying to make sure it works with your rendering setup.

 

Hmm. Might actually be a bug in the Unity version you are using. At least such problem seems to be fixed in 2019.4.3f1. See here: https://unity.com/releases/editor/whats-new/2019.4.3

Maybe you could try that version version? If that does not help, let's check more.

EDIT :Hmm.. but you mentioned that your version is slightly newer. What kind of integration are you using and does the problem happen in editor? You could try to see if it is audio related by disabling audio, set Recorder.RecordAudio = false and be sure you haven't added the audio recorder component (usually added automatically but setting RecordAudio false disables it)

 

Automatic integration with URP 14 is still unsupported but I have created a custom integration guide for it. It handles two different cases, a) Recording the screen, b) Recording the screen without the UI.

You can find it from here.

 

Unfortunately no. Audio Mixer Recorder passes the mixer channel data directly to NGR in native code so the data never goes to C#.

 

Actually after thinking it a bit more it might actually work. You just need to change the ngr_AudioMixerRecorderConnect's first parameter in AudioMixerRecorder.cs to point to a static C# function instead of the native one.

Code (CSharp):

1. delegate void AudioCallbackDelegate(System.IntPtr ptr, int bufferLength, int numChannels, int sampleRate);
2.  
3. [MonoPInvokeCallback(typeof(AudioCallbackDelegate))]
4. private static void AudioCallback(IntPtr ptr, int bufferLength, int numChannels, int sampleRate) {
5.     float[] channelData = new float[bufferLength];
6.     Marshal.Copy(ptr, channelData, 0, bufferLength);
7. }
8.  
9. void OnEnable()
10. {
11.     if (s_Instance == this)
12.     {
13.         ngr_AudioMixerRecorderConnect(AudioCallback);
14.     }
15. }

This would not be very optimal though because you need to do the marshal copy but you would get the data. You must also keep in mind that the function will be called in audio thread so you should not block it in any way or the audio will stutter. You should probably instead send the channelData to some queue or a circular buffer and process it in some Update() so everything you do will be called in main thread. Almost all Unity calls will fail or crash when called from an arbitrary thread.

 

The problem in your code is that it only defines what texture to record but it does not tell when to record it. So you are basically missing the recording texture blit part which actually copies the recording texture to a video frame. Please view the custom integration part at the end of this documentation: http://www.pmjo.org/nextgenrecorder/integration.html

Keep in mind that the way of integration depends on the used rendering pipeline. By default Unity uses built in render pipeline. There are some example scripts for different pipelines in the above mentioned integration guide.

 

It would not be black because there would be no frames to store to the video, video would be impossible to export.