Does gst-nvdsmetamux support mux output metas from NvDsFrameMeta and NvDsAudioFrameMeta?

[dev_37]

When I refer to deepstream_parallel_inference_app, I see that the app only support multiple models inference with nvinfer(TensorRT) or nvinferserver(Triton) in parallel. So, I have idea to inference multiple models with nvinfer and nvinferaudio in parallel. I want to use gst-nvdsmetamux (MetaMux in the below image) to mux output meta from audio model and video model, Is it possible? I know that NvDsAudioFrameMeta used to store metadata of audio and NvDsFrameMeta used to store metadata of video.
Thank you very much.

• Hardware Platform (Jetson / GPU) Jetson NX Xavier
• DeepStream Version 6.3
• JetPack Version (valid for Jetson only) 5.1.2
• TensorRT Version 8.5
• Issue Type (questions, new requirements, bugs) questions

[Fiona.Chen]

@dev_37

So, I have idea to inference multiple models with nvinfer and nvinferaudio in parallel. I want to use gst-nvdsmetamux (MetaMux in the below image) to mux output meta from audio model and video model, Is it possible?

As you known, “NvDsAudioFrameMeta used to store metadata of audio and NvDsFrameMeta used to store metadata of video”, the audio and video are separated data. What do you mean by “mux output meta from audio model and video model”?

[dev_37]

@Fiona.Chen

What do you mean by “mux output meta from audio model and video model”?

Thank you for your answer,
I checked again, it’s not necessary to “mux output meta from audio model and video model”.