Recherche avancée

Sur d’autres sites (10789)

• Issues with Publishing and Subscribing Rates for H.264 Video Streaming over RabbitMQ

  7 octobre 2024, par Luis

  I am working on a project to stream an H.264 video file using RabbitMQ (AMQP protocol) and display it in a web application. The setup involves capturing video frames, encoding them, sending them to RabbitMQ, and then consuming and decoding them on the web application side using Flask and Flask-SocketIO.

  


  However, I am encountering performance issues with the publishing and subscribing rates in RabbitMQ. I cannot seem to achieve more than 10 messages per second. This is not sufficient for smooth video streaming.
I need help to diagnose and resolve these performance bottlenecks.

  


  Here is my code :

  


  

  • Video Capture and Publishing Script :

  


  


  # RabbitMQ setup
RABBITMQ_HOST = 'localhost'
EXCHANGE = 'DRONE'
CAM_LOCATION = 'Out_Front'
KEY = f'DRONE_{CAM_LOCATION}'
QUEUE_NAME = f'DRONE_{CAM_LOCATION}_video_queue'

# Path to the H.264 video file
VIDEO_FILE_PATH = 'videos/FPV.h264'

# Configure logging
logging.basicConfig(level=logging.INFO)

@contextmanager
def rabbitmq_channel(host):
    """Context manager to handle RabbitMQ channel setup and teardown."""
    connection = pika.BlockingConnection(pika.ConnectionParameters(host))
    channel = connection.channel()
    try:
        yield channel
    finally:
        connection.close()

def initialize_rabbitmq(channel):
    """Initialize RabbitMQ exchange and queue, and bind them together."""
    channel.exchange_declare(exchange=EXCHANGE, exchange_type='direct')
    channel.queue_declare(queue=QUEUE_NAME)
    channel.queue_bind(exchange=EXCHANGE, queue=QUEUE_NAME, routing_key=KEY)

def send_frame(channel, frame):
    """Encode the video frame using FFmpeg and send it to RabbitMQ."""
    ffmpeg_path = 'ffmpeg/bin/ffmpeg.exe'
    cmd = [
        ffmpeg_path,
        '-f', 'rawvideo',
        '-pix_fmt', 'rgb24',
        '-s', '{}x{}'.format(frame.shape[1], frame.shape[0]),
        '-i', 'pipe:0',
        '-f', 'h264',
        '-vcodec', 'libx264',
        '-pix_fmt', 'yuv420p',
        '-preset', 'ultrafast',
        'pipe:1'
    ]
    
    start_time = time.time()
    process = subprocess.Popen(cmd, stdin=subprocess.PIPE, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
    out, err = process.communicate(input=frame.tobytes())
    encoding_time = time.time() - start_time
    
    if process.returncode != 0:
        logging.error("ffmpeg error: %s", err.decode())
        raise RuntimeError("ffmpeg error")
    
    frame_size = len(out)
    logging.info("Sending frame with shape: %s, size: %d bytes", frame.shape, frame_size)
    timestamp = time.time()
    formatted_timestamp = datetime.fromtimestamp(timestamp).strftime('%H:%M:%S.%f')
    logging.info(f"Timestamp: {timestamp}") 
    logging.info(f"Formatted Timestamp: {formatted_timestamp[:-3]}")
    timestamp_bytes = struct.pack('d', timestamp)
    message_body = timestamp_bytes + out
    channel.basic_publish(exchange=EXCHANGE, routing_key=KEY, body=message_body)
    logging.info(f"Encoding time: {encoding_time:.4f} seconds")

def capture_video(channel):
    """Read video from the file, encode frames, and send them to RabbitMQ."""
    if not os.path.exists(VIDEO_FILE_PATH):
        logging.error("Error: Video file does not exist.")
        return
    cap = cv2.VideoCapture(VIDEO_FILE_PATH)
    if not cap.isOpened():
        logging.error("Error: Could not open video file.")
        return
    try:
        while True:
            start_time = time.time()
            ret, frame = cap.read()
            read_time = time.time() - start_time
            if not ret:
                break
            frame_rgb = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
            frame_rgb = np.ascontiguousarray(frame_rgb) # Ensure the frame is contiguous
            send_frame(channel, frame_rgb)
            cv2.imshow('Video', frame)
            if cv2.waitKey(1) & 0xFF == ord('q'):
                break
            logging.info(f"Read time: {read_time:.4f} seconds")
    finally:
        cap.release()
        cv2.destroyAllWindows()


  


  

  • the backend (flask) :

  


  


  app = Flask(__name__)
CORS(app)
socketio = SocketIO(app, cors_allowed_origins="*")

RABBITMQ_HOST = 'localhost'
EXCHANGE = 'DRONE'
CAM_LOCATION = 'Out_Front'
QUEUE_NAME = f'DRONE_{CAM_LOCATION}_video_queue'

def initialize_rabbitmq():
    connection = pika.BlockingConnection(pika.ConnectionParameters(RABBITMQ_HOST))
    channel = connection.channel()
    channel.exchange_declare(exchange=EXCHANGE, exchange_type='direct')
    channel.queue_declare(queue=QUEUE_NAME)
    channel.queue_bind(exchange=EXCHANGE, queue=QUEUE_NAME, routing_key=f'DRONE_{CAM_LOCATION}')
    return connection, channel

def decode_frame(frame_data):
    # FFmpeg command to decode H.264 frame data
    ffmpeg_path = 'ffmpeg/bin/ffmpeg.exe'
    cmd = [
        ffmpeg_path,
        '-f', 'h264',
        '-i', 'pipe:0',
        '-pix_fmt', 'bgr24',
        '-vcodec', 'rawvideo',
        '-an', '-sn',
        '-f', 'rawvideo',
        'pipe:1'
    ]
    process = subprocess.Popen(cmd, stdin=subprocess.PIPE, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
    start_time = time.time()  # Start timing the decoding process
    out, err = process.communicate(input=frame_data)
    decoding_time = time.time() - start_time  # Calculate decoding time
    
    if process.returncode != 0:
        print("ffmpeg error: ", err.decode())
        return None
    frame_size = (960, 1280, 3)  # frame dimensions expected by the frontend
    frame = np.frombuffer(out, np.uint8).reshape(frame_size)
    print(f"Decoding time: {decoding_time:.4f} seconds")
    return frame

def format_timestamp(ts):
    dt = datetime.fromtimestamp(ts)
    return dt.strftime('%H:%M:%S.%f')[:-3]

def rabbitmq_consumer():
    connection, channel = initialize_rabbitmq()
    for method_frame, properties, body in channel.consume(QUEUE_NAME):
        message_receive_time = time.time()  # Time when the message is received

        # Extract the timestamp from the message body
        timestamp_bytes = body[:8]
        frame_data = body[8:]
        publish_timestamp = struct.unpack('d', timestamp_bytes)[0]

        print(f"Message Receive Time: {message_receive_time:.4f} ({format_timestamp(message_receive_time)})")
        print(f"Publish Time: {publish_timestamp:.4f} ({format_timestamp(publish_timestamp)})")

        frame = decode_frame(frame_data)
        decode_time = time.time() - message_receive_time  # Calculate decode time

        if frame is not None:
            _, buffer = cv2.imencode('.jpg', frame)
            frame_data = buffer.tobytes()
            socketio.emit('video_frame', {'frame': frame_data, 'timestamp': publish_timestamp}, namespace='/')
            emit_time = time.time()  # Time after emitting the frame

            # Log the time taken to emit the frame and its size
            rtt = emit_time - publish_timestamp  # Calculate RTT from publish to emit
            print(f"Current Time: {emit_time:.4f} ({format_timestamp(emit_time)})")
            print(f"RTT: {rtt:.4f} seconds")
            print(f"Emit time: {emit_time - message_receive_time:.4f} seconds, Frame size: {len(frame_data)} bytes")
        channel.basic_ack(method_frame.delivery_tag)

@app.route('/')
def index():
    return render_template('index.html')

@socketio.on('connect')
def handle_connect():
    print('Client connected')

@socketio.on('disconnect')
def handle_disconnect():
    print('Client disconnected')

if __name__ == '__main__':
    consumer_thread = threading.Thread(target=rabbitmq_consumer)
    consumer_thread.daemon = True
    consumer_thread.start()
    socketio.run(app, host='0.0.0.0', port=5000)



  


  How can I optimize the publishing and subscribing rates to handle a higher number of messages per second ?

  


  Any help or suggestions would be greatly appreciated !

  


  I attempted to use threading and multiprocessing to handle multiple frames concurrently and I tried to optimize the frame decoding function to make it faster but with no success.

  


• Revision 122868 : Toujours sur ces vues de saisies fieldset. Champ extra se base sur le ...

  7 mars 2020, par Maïeul Rouquette — Log

  Toujours sur ces vues de saisies fieldset.
  Champ extra se base sur le classe vide pour ne pas afficher les saisies
  sans réponse. Le masquage est en css, contrairement à ce que je pensais
  (suis-je bête !)
  On ajoute donc la classe vide à la vie d’une saisie si l’ensemble de ses
  sous saisies sont vide.
  Permet d’éviter avec champ extra d’avoir un titre de fieldset qui se
  balade sans aucune réponse en dessous.

• Evolution #2040 (Nouveau) : Accueil inistial sur un site nouvellement installé

  1er mai 2011, par Suske -

  Première arrivée dans le privé : contenu/accueil.html est vide => il y a mieux que le vide :-)

  un lien vers la config du site

  une phrase "Créez maintenant une rubrique puis publiez-y un premier article... bla."

  ... Un page "premiers pas en quelques sortes (...)