Added simple RTSP support and custom video sink option #40
Conversation
- Implement RTSP source handling in the GStreamer pipeline - Add command-line argument for specifying custom video sink - Update pipeline string generation to include RTSP source - Enhance flexibility by allowing user to choose video sink - Improve error handling for invalid network types - Refactor pipeline string generation for better readability
|
@chm10 Thanks for your input. The current code is hardcoded to support only H.265 videos. Please convert it to use auto-pluggers like decodebin. It should look something like this: rtspsrc location=rtsp://your_rtsp_stream_url ! application/x-rtp,media=video ! decodebin ! videoconvert ! autovideosink Once that's done, I'll be able to merge it. Thanks a lot for your contributions; they are highly appreciated. I'd be happy to merge your code here. |
I will change the code. I only have devices using H.265 rtsp for now. |
…into feature/rtsp-compatibility-39
…lo-ai#40) Update the GStreamer pipeline to use decodebin for RTSP video sources Remove hardcoded H.265 video decoding elements Automatically select the appropriate decoder based on the incoming video stream's codec Ensure compatibility with various video codecs without explicit specification Maintain existing functionality for other video sources (RPI, USB, file)
| @@ -92,6 +92,7 @@ class GStreamerDetectionApp(GStreamerApp): | |||
| def __init__(self, args, user_data): | |||
| # Call the parent class constructor | |||
| super().__init__(args, user_data) | |||
| self.video_sink = args.video_sink if hasattr(args, 'video_sink') else 'autovideosink' | |||
| @@ -216,6 +225,8 @@ def get_pipeline_string(self): | |||
| default=None, | |||
| help="Path to costume labels JSON file", | |||
| ) | |||
| parser.add_argument('--video-sink', default='autovideosink', | |||
| source_element = ( | ||
| f"rtspsrc location={self.video_source} name=src_0 ! " | ||
| "application/x-rtp,media=video ! " | ||
| "decodebin ! " |
There was a problem hiding this comment.
You'll need to add some queues between decodebin, videoconvert and videoscale. These are big tasks which can benefit if separated to multiple threads.
New feature #39
Disclaimer:
I am not the owner or maintainer of this project. This pull request is a suggested solution based on
the information available in issue #39. It is provided "as is", without warranty of any kind, express
or implied. The project maintainers should review and test these changes thoroughly before considering
implementation. I'm open to feedback and further modifications if needed.