Skip to content

[DOCS] Corrections in Multi-Stream GStreamer README #524

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
tbujewsk merged 1 commit into from
Jan 13, 2026
Merged

[DOCS] Corrections in Multi-Stream GStreamer README #524

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
22 changes: 11 additions & 11 deletions samples/gstreamer/gst_launch/multi_stream/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,21 +1,21 @@
# Video streams from multiple cameeras (gst-launch command line)
# Video streams from multiple cameras (gst-launch command line)

This sample demonstrates how to construct multi-stream pipeline via `gst-launch-1.0` command-line utility using detection and classification models.
It combines four pipelines. By default, the first streams run on NPU [Intel® Core™ Ultra processors] and the other two using the GPU device.
It combines four pipelines. By default, the first two streams run on NPU [Intel® Core™ Ultra processors] and the other two using the GPU device.

## How It Works
This sample utilizes GStreamer command-line tool `gst-launch-1.0` which can build and run GStreamer pipeline described in a string format.
The string contains a list of GStreamer elements separated by exclamation mark `!`, each element may have properties specified in the format `property`=`value`.

> **NOTE**: Before run please download yolov8s model to `$MODELS_PATH/public/yolov8s/FP16/` location.
> **NOTE**: Before running, download the yolov8s model to `$MODELS_PATH/public/yolov8s/FP16/`.
Please follow instruction: [Detection with Yolo](./gst_launch/detection_with_yolo/README.md) how to download Yolov8s model.

This sample builds four GStreamer pipeline of the following elements
This sample builds for GStreamer a pipeline of the following elements:
* `filesrc`
* `decodebin3` for video decoding
* `videoconvert` for converting video frame into different color formats
* [gvadetect](../../../../docs/source/elements/gvadetect.md) uses for full-frame object detection and marking objects with labels
* [gvawatermark](../../../../docs/source/elements/gvawatermark.md) for points and theirs connections visualization
* [gvawatermark](../../../../docs/source/elements/gvawatermark.md) for points and visualization of their connections
* `autovideosink` for rendering output video into screen
> **NOTE**: Each of the two pipelines can run on CPU or GPU or NPU.
> **NOTE**: `sync=false` property in `autovideosink` element disables real-time synchronization so pipeline runs as fast as possible
Expand Down Expand Up @@ -57,17 +57,17 @@ gvawatermark ! gvafpscounter ! vaapih264enc ! h264parse ! mp4mux ! filesink loca
filesrc location=${INPUT_VIDEO_FILE_2} ! decodebin3 ! vaapipostproc ! video/x-raw(memory:VASurface) ! \
gvadetect model=${DETECTION_MODEL} device=NPU pre-process-backend=ie nireq=4 model-instance-id=inf0 ! queue ! \
gvawatermark ! gvafpscounter ! vaapih264enc ! h264parse ! mp4mux ! filesink location=${OUTPUT_VIDEO_FILE_2} \
filesrc location=${INPUT_VIDEO_FILE} ! decodebin3 vaapipostproc ! video/x-raw(memory:VASurface) !
gvadetect model={$DETECTION_MODEL_3} device=GPU pre-process-backend=vaapi-surface-sharing nireq=4 model-instance-id=inf1 ! queue ! \
filesrc location=${INPUT_VIDEO_FILE} ! decodebin3 ! vaapipostproc ! video/x-raw(memory:VASurface) !
gvadetect model=${$DETECTION_MODEL_3} device=GPU pre-process-backend=vaapi-surface-sharing nireq=4 model-instance-id=inf1 ! queue ! \
gvawatermark ! gvafpscounter ! vaapih264enc ! h264parse ! mp4mux ! filesink location=${OUTPUT_VIDEO_FILE_3} \
filesrc location=${INPUT_VIDEO_FILE_4} ! decodebin3 vaapipostproc ! video/x-raw(memory:VASurface) !
gvadetect model=${DETECTION_MODEL} device=GPU pre-process-backend=vaapi-surface-sharing nireq=4 model-instance-id=inf1 ! queue ! \
gvawatermark ! gvafpscounter ! vaapih264enc ! h264parse ! mp4mux ! filesink location=${OUTPUT_VIDEO_FILE_4}
```

The next pipeline illustrates how to construct a plipeline with multiple AI models and a single video stream.
In addition, this pipeline runs AI inference every 3 frames ('inference-inteval=3') and uses 'gvatrack' to keep analytics results for non-inferenced frames.
The example also batches inference requests ('batch-size=8') to maximize AI model throughput at the expense of single-reqeust lantency.
The next pipeline illustrates how to construct a pipeline with multiple AI models and a single video stream.
In addition, this pipeline runs AI inference every 3 frames ('inference-interval=3') and uses 'gvatrack' to keep analytics results for non-inferenced frames.
The example also batches inference requests ('batch-size=8') to maximize AI model throughput at the expense of single-request latency.

```sh
gst-launch-1.0 \
Expand Down Expand Up @@ -96,4 +96,4 @@ gvawatermark ! gvafpscounter ! vaapih264enc ! h264parse ! mp4mux ! filesink loca
```

## See also
* [Samples overview](../../README.md)
* [Samples overview](../../README.md)
Loading