Recherche avancée

Sur d’autres sites (10861)

• ffmpeg + AWS Lambda issues. Won't compress full video

  7 juillet 2022, par Joesph Stah Lynn

  So I followed this tutorial to set everything up, and changed the function a bit to compress video, but no matter what I try, on larger videos (basically anything over 50-100MB), the output file will always be cut short, and depending on the encoding settings I'm using, will be cut by different amounts. I tried using the solution found here, adding a -nostdin flag to my ffmpeg command, but that also didn't seem to fix the issue.
  
Another odd thing, is no matter what I try, if I remove the '-f mpegts' flag, the output video will be 0B.
  
My Lambda function is set up with 3008MB of Memory (submitted a ticket to get my limit upped so I can use the full 10240MB available), and 2048MB of Ephemeral storage (I honestly am not sure if I need anything more than the minimum 512, but I upped it to try and fix the issue). When I check my cloudwatch logs, on really large files, it will occasionally time out, but other than that, I will get no error messages, just the standard start, end, and billable time messages.

  


  This is the code for my lambda function.

  


  import json
import os
import subprocess
import shlex
import boto3

S3_DESTINATION_BUCKET = "rw-video-out"
SIGNED_URL_TIMEOUT = 600

def lambda_handler(event, context):

    s3_source_bucket = event['Records'][0]['s3']['bucket']['name']
    s3_source_key = event['Records'][0]['s3']['object']['key']

    s3_source_basename = os.path.splitext(os.path.basename(s3_source_key))[0]
    s3_destination_filename = s3_source_basename + "-comp.mp4"

    s3_client = boto3.client('s3')
    s3_source_signed_url = s3_client.generate_presigned_url('get_object',
        Params={'Bucket': s3_source_bucket, 'Key': s3_source_key},
        ExpiresIn=SIGNED_URL_TIMEOUT)

    ffmpeg_cmd = f"/opt/bin/ffmpeg -nostdin -i {s3_source_signed_url} -f mpegts libx264 -preset fast -crf 28 -c:a copy - "
    command1 = shlex.split(ffmpeg_cmd)
    p1 = subprocess.run(command1, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
    resp = s3_client.put_object(Body=p1.stdout, Bucket=S3_DESTINATION_BUCKET, Key=s3_destination_filename)
    s3 = boto3.resource('s3')
    s3.Object(s3_source_bucket,s3_source_key).delete()

    return {
        'statusCode': 200,
        'body': json.dumps('Processing complete successfully')
    }


  


  This is the code from the solution I mentioned, but when I try using this code, I get output.mp4 not found errors

  


  def lambda_handler(event, context):
    print(event)
    os.chdir('/tmp')
    s3_source_bucket = event['Records'][0]['s3']['bucket']['name']
    s3_source_key = event['Records'][0]['s3']['object']['key']

    s3_source_basename = os.path.splitext(os.path.basename(s3_source_key))[0]
    s3_destination_filename = s3_source_basename + ".mp4"

    s3_client = boto3.client('s3')
    s3_source_signed_url = s3_client.generate_presigned_url('get_object',
        Params={'Bucket': s3_source_bucket, 'Key': s3_source_key},
        ExpiresIn=SIGNED_URL_TIMEOUT)
    print(s3_source_signed_url)
    s3_client.download_file(s3_source_bucket,s3_source_key,s3_source_key)
    # ffmpeg_cmd = "/opt/bin/ffmpeg -framerate 25 -i \"" + s3_source_signed_url + "\" output.mp4 "
    ffmpeg_cmd = f"/opt/bin/ffmpeg -framerate 25 -i {s3_source_key} output.mp4 "
    # command1 = shlex.split(ffmpeg_cmd)
    # print(command1)
    os.system(ffmpeg_cmd)
    # os.system('ls')
    # p1 = subprocess.run(command1, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
    file = 'output.mp4'
    resp = s3_client.put_object(Body=open(file,"rb"), Bucket=S3_DESTINATION_BUCKET, Key=s3_destination_filename)
    # resp = s3_client.put_object(Body=p1.stdout, Bucket=S3_DESTINATION_BUCKET, Key=s3_destination_filename)
    s3 = boto3.resource('s3')
    s3.Object(s3_source_bucket,s3_source_key).delete()
    return {
        'statusCode': 200,
        'body': json.dumps('Processing complete successfully')
    }


  


  Any help would be greatly appreciated.

  


• Facing Severe Buffering Issues in FFmpeg Pipeline for Simultaneous Video and Audio Live Streaming via SRT Protocol

  7 novembre 2023, par Johnny

  I'm currently working on a project where I'm using FFmpeg to livestream via Cloudflare using the SRT protocol. My main goal is to make sure I can stream both video and audio at the same time. Things have been going well with video streaming, but I've run into a bit of a problem when it comes to adding audio to the mix.

  


  The issue I'm facing is that when I try to connect the audio to the video stream, I'm having some serious buffering problems. The video with audio does work, but all that buffering is messing up the quality of the livestream.

  


  I have a feeling the problem might have something to do with my FFmpeg pipeline, but I'm still pretty new to using FFmpeg. Could someone please lend a hand and help me figure out how to fix this buffering issue in my FFmpeg setup so I can livestream both video and audio at the same time ? Any advice or suggestions you have would be a huge help.

  


  Here's the code I'm using for my FFmpeg pipeline :

  


  ffmpeg_stream = [
            "ffmpeg",
            "-re",
            "-flags", "+global_header",
            "-fflags", "+genpts",
            "-threads", "8",
            "-thread_queue_size", "1024",
            "-loglevel", "debug",
            "-f", "rawvideo",
            "-pix_fmt", "bgr24",
            "-s", "1280x720",
            "-i", "-",
            "-channel_layout", "mono",
            "-f", "alsa",
            "-thread_queue_size", "32768",
            "-ac", "1",
            "-i", "hw:4,0",
            "-pix_fmt", "yuv420p",
            "-c:v", "libx264",
            "-max_muxing_queue_size", "9999",
            "-b:v:0", "3000k",
            "-g", "25",
            "-c:a", "aac",
            "-ar", "44100",
            "-b:a", "128k",
            "-map", "0:v:0",
            "-map", "1:a:0",
            "-preset", "veryfast",
            "-tune", "zerolatency",
            "-f", "mpegts",
            my_srt_url
        ]


  


  Let me know if any other info is needed. Thanks !

  


• 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.