Control video record at runtime

Healthcare Holoscan SDK
1

Is there possible to use ops::VideoStreamRecorderOp operator to achieve the following goals at runtime:

  1. Start and stop when some conditions met, such as user clicks buttons.
  2. Record to different file name for each recording.
  3. Split the record file name for every N size, such as a_1.mp4, a_2.mp4 with size 10MB each.
  4. H265 encoding.
2

Hi there,

  1. There currently isn’t a supported and tested way to implement this functionality.
  2. Looking at path in video_stream_recorder.cpp, it seems like the currently implementation of video recorder can only write to one output file from one application run, with a fixed name (predetermined basename or receiver name), and different file names is also not currently supported.
  3. Same as 2, it does not seem implemented in the current recorder.
  4. There isn’t H265 only H264 decoding and recording as illustrated in applications/h264_endoscopy_tool_tracking/main.cpp.

If you’d like to, could you tell us a bit more about your use case and application needs? We’re always looking for feedback for improving the SDK.

3

Thanks for your answer.
Here are our requirements:

  1. User can click start recording and stop recording at any time to start or stop recording microscope video stream.
  2. When recording video stream, user can ALSO click screenshot to capture current frame and save as jpg picture.
  3. The recording file can be splitted by file size such as every 1GB (can be configured when start recording) for each file, just like gstreamer multifilesink element.
  4. Supporting H265 is better!
    Thanks!