PyStreamer

A FFmpeg-based tool that helps you push video and audio to RTMP streaming server in real time. It can be easily integrated into your python DIP workflow (e.g. NumPy, OpenCV, and PyTorch) to create live streaming applications.

This repository is currently under development, and the code may be unstable. If you encounter any issues or bugs, please feel free to open an issue.

Preparation

• Install FFmpeg.
• Deploy an RTMP server, like NGINX-based Media Streaming Server.

Usage

First, clone this respository under your workdir:

$ cd /your/workdir
$ git clone https://github.com/JianghaoLan/pystreamer.git

# /your/workdir
# ├── pystreamer
# │   ├── ...
# ├── your_python_file.py
# └── ...

Next, import Streamer in your python file:

from pystreamer import Streamer

Then, you can refer to the following example to push your video and audio to RTMP server:

from pystreamer import Streamer

if __name__ == '__main__':

    # Use opencv to read video frames and librosa to read audio frames. You can get these data by any other ways.
    import cv2
    import librosa

    cap = cv2.VideoCapture("test-video.mp4")
    fps = int(cap.get(cv2.CAP_PROP_FPS))
    width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
    height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
    print(f"Video info: {width}x{height}, {fps}fps.")

    audio_frames, sr = librosa.load("test_audio.wav")
    print(f"Audio sample rate: {sr}")

    ### 1. create `Streamer` object ###
    rtmp_url = "rtmp://localhost:1935/live/app"
    streamer = Streamer(rtmp_url, (width, height), fps, sr)

    ### 2. invoke `start()` to open ffmpeg process ###
    streamer.start()

    i = 0
    audio_frame_size = sr // fps
    while (cap.isOpened()):
        ret, video_frame = cap.read()
        if not ret:
            break

        audio_frame = audio_frames[i:i+audio_frame_size]

        ### 3. use `push()` to push your media data to RTMP server ###
        # `video_frame`: numpy array (h, w, 3) or (n, h, w, 3), uint8
        # `audio_frame`: 1d numpy array, float32
        streamer.push(video_frame, audio_frame)

        i += audio_frame_size

    ### 4. `stop()` when you finish your stream ###
    streamer.stop()

    print("Finish.")

Options

Streamer's constructor and push() method accept some other parameters you may need. For example, you may want to hide ffmpeg's console output (set show_log=False), or don't want to push audio (set no_audio=True). You can refer to method's docstring for more detail.

For some other customizations, you may need to modify my code. For example, if you are using other streaming protocols (such as HLS, DASH, etc.) instead of RTMP, check out FfmpegProcess.py and modify the FFmpeg command.