Recherche avancée

Sur d’autres sites (11978)

• matplotlib 3D linecollection animation gets slower over time

  15 juin 2021, par Vignesh Desmond

  I'm trying to animate a 3d line plot for attractors, using Line3DCollection. The animation is initally fast but it gets progressively slower over time. A minimal example of my code :

  


  def generate_video(nframes):

    fig = plt.figure(figsize=(16, 9), dpi=120)
    canvas_width, canvas_height = fig.canvas.get_width_height()
    ax = fig.add_axes([0, 0, 1, 1], projection='3d')

    X = np.random.random(nframes)
    Y = np.random.random(nframes)
    Z = np.random.random(nframes)

    cmap = plt.cm.get_cmap("hsv")
    line = Line3DCollection([], cmap=cmap)
    ax.add_collection3d(line)
    line.set_segments([])

    def update(frame):
        i = frame % len(vect.X)
        points = np.array([vect.X[:i], vect.Y[:i], vect.Z[:i]]).transpose().reshape(-1,1,3)
        segs = np.concatenate([points[:-1],points[1:]],axis=1)
        line.set_segments(segs)
        line.set_array(np.array(vect.Y)) # Color gradient
        ax.elev += 0.0001
        ax.azim += 0.1

    outf = 'test.mp4'
    cmdstring = ('ffmpeg', 
                    '-y', '-r', '60', # overwrite, 1fps
                    '-s', '%dx%d' % (canvas_width, canvas_height),
                    '-pix_fmt', 'argb',
                    '-f', 'rawvideo',  '-i', '-',
                    '-b:v', '5000k','-vcodec', 'mpeg4', outf)
    p = subprocess.Popen(cmdstring, stdin=subprocess.PIPE)

    for frame in range(nframes):
        update(frame)
        fig.canvas.draw()
        string = fig.canvas.tostring_argb()
        p.stdin.write(string)

    p.communicate()

generate_video(nframes=10000)


  


  I used the code from this answer to save the animation to mp4 using ffmpeg instead of anim.FuncAnimation as its much faster for me. But both methods get slower over time and I'm not sure how to make the animation not become slower. Any advice is welcome.

  


  Versions :
Matplotlib : 3.4.2
FFMpeg : 4.2.4-1ubuntu0.1

  


• Playing RTP stream on Android 4.1.2 (Jelly Bean) [closed]

  27 décembre 2024, par Homie_Tomie

  I'll try to keep it quick. Using FFMPEG I started a stream on my PC. Here is the code :

  


  import subprocess

def start_stream():
    command = [
        'ffmpeg',
        '-f', 'gdigrab',  # Desktop capture (Windows)
        '-framerate', '15',  # Low framerate for higher performance
        '-i', 'desktop',  # Capture desktop
        '-c:v', 'libx264',  # Video codec (H.264)
        '-preset', 'ultrafast',  # Ultra-fast encoding preset for minimal latency
        '-tune', 'zerolatency',  # Zero latency for real-time streaming
        '-x264opts', 'keyint=15:min-keyint=15:no-scenecut',  # Frequent keyframes
        '-b:v', '500k',  # Low bitrate to minimize data usage and reduce latency
        '-s', '800x480',  # Resolution fits phone screen and helps performance
        '-max_delay', '0',  # No buffering, instant frame output
        '-flush_packets', '1',  # Flush packets immediately after encoding
        '-f', 'rtp',  # Use mpegts as the container for RTP stream
        'rtp://192.168.72.26:1234',  # Stream over UDP to localhost on port 1234
        '-sdp_file', 'stream.sdp'  # Create SDP file
    ]
    
    try:
        print("Starting stream...")
        subprocess.run(command, check=True)
    except subprocess.CalledProcessError as e:
        print(f"Error occurred: {e}")
    except KeyboardInterrupt:
        print("\nStream interrupted")

if __name__ == "__main__":
    print("Starting screen capture...")
    start_stream()


  


  Now, when I start the stream I can connect to it in VLC when I open up the stream.sdp file. Using the same method I can open up the stream on my iPhone, but when I try to open it on my old Android phone the stream connects but the screen is black. However, when I turn the screen I can see the first frame that was sent to the phone. Why does the stream not work ?

  


  I will be thankful for any and all advice :)

  


• iOS Radio App : Need to extract and stream audio-only from HLS streams with video content

  20 décembre 2024, par Bader Alghamdi

  I'm developing an iOS radio app that plays various HLS streams. The challenge is that some stations broadcast HLS streams containing both audio and video (example : https://svs.itworkscdn.net/smcwatarlive/smcwatar/chunks.m3u8), but I want to :

  


  Extract and play only the audio track
Support AirPlay for audio-only streaming
Minimize data usage by not downloading video content
Technical Details :

  


  iOS 17+
Swift 6
Using AVFoundation for playback
Current implementation uses AVPlayer with AVPlayerItem
Current Code Structure :

  


  class StreamPlayer: ObservableObject { @Published var isPlaying = false private var player: AVPlayer? private var playerItem: AVPlayerItem?

func playStream(url: URL) {
    let asset = AVURLAsset(url: url)
    playerItem = AVPlayerItem(asset: asset)
    player = AVPlayer(playerItem: playerItem)
    player?.play()
}



  


  Stream Analysis : When analyzing the video stream using FFmpeg :

  


  CopyInput #0, hls, from 'https://svs.itworkscdn.net/smcwatarlive/smcwatar/chunks.m3u8':
  Stream #0:0: Video: h264, yuv420p(tv, bt709), 1920x1080 [SAR 1:1 DAR 16:9], 25 fps
  Stream #0:1: Audio: aac, 44100 Hz, stereo, fltp



  


  Attempted Solutions :

  


  Using MobileFFmpeg :

  


  let command = [
    "-i", streamUrl,
    "-vn",
    "-acodec", "aac",
    "-ac", "2",
    "-ar", "44100",
    "-b:a", "128k",
    "-f", "mpegts",
    "udp://127.0.0.1:12345"
].joined(separator: " ")

ffmpegProcess = MobileFFmpeg.execute(command)

I


  


  ssue : While FFmpeg successfully extracts audio, playback through AVPlayer doesn't work reliably.

  


  Tried using HLS output :

  


  let command = [
    "-i", streamUrl,
    "-vn",
    "-acodec", "aac",
    "-ac", "2",
    "-ar", "44100",
    "-b:a", "128k",
    "-f", "hls",
    "-hls_time", "2",
    "-hls_list_size", "3",
    outputUrl.path
]



  


  Issue : Creates temporary files but faces synchronization issues with live streams.

  


  ---

  


  Testing URLs :

  


  Audio+Video : https://svs.itworkscdn.net/smcwatarlive/smcwatar/chunks.m3u8
Audio Only : https://mbcfm-radio.mbc.net/mbcfm-radio.m3u8

  


  ---

  


  Requirements :

  


  

  • Real-time audio extraction from HLS stream

  


  • Maintain live streaming capabilities

  


  • Full AirPlay support

  


  • Minimal data usage (avoid downloading video content)

  


  • Handle network interruptions gracefully

  


  


  Questions :

  


  

  • What's the most efficient way to extract only audio from an HLS stream in real-time ?

  


  • Is there a way to tell AVPlayer to ignore video tracks completely ?

  


  • Are there better alternatives to FFmpeg for this specific use case ?

  


  • What's the recommended approach for handling AirPlay with modified streams ?

  


  


  Any guidance or alternative approaches would be greatly appreciated. Thank you !

  


  What I Tried :

  


  

  1. Direct AVPlayer Implementation :

  


  


  

  • Used standard AVPlayer to play HLS stream

  


  • Expected it to allow selecting audio-only tracks

  


  • Result : Always downloads both video and audio, consuming unnecessary bandwidth

  


  


  ---

  


  

  2. FFmpeg Audio Extraction :

  


  


  let command = [
    "-i", "https://svs.itworkscdn.net/smcwatarlive/smcwatar/chunks.m3u8",
    "-vn",                    // Remove video
    "-acodec", "aac",        // Audio codec
    "-ac", "2",              // 2 channels
    "-ar", "44100",          // Sample rate
    "-b:a", "128k",          // Bitrate
    "-f", "mpegts",          // Output format
    "udp://127.0.0.1:12345"  // Local stream
]
ffmpegProcess = MobileFFmpeg.execute(command)



  


  Expected : Clean audio stream that AVPlayer could play
Result : FFmpeg extracts audio but AVPlayer can't play the UDP stream

  


  ---

  


  

  3. HLS Segmented Approach :

  


  


  swiftCopylet command = [
    "-i", streamUrl,
    "-vn",
    "-acodec", "aac",
    "-f", "hls",
    "-hls_time", "2",
    "-hls_list_size", "3",
    outputUrl.path
]



  


  Expected : Create local HLS playlist with audio-only segments
Result : Creates files but faces sync issues with live stream

  


  ---

  


  Expected Behavior :

  


  

  • Stream plays audio only

  


  • Minimal data usage (no video download)

  


  • Working AirPlay support

  


  • Real-time playback without delays

  


  


  Actual Results :

  


  

  • Either downloads full video stream (wasteful)

  


  • Or fails to play extracted audio

  


  • AirPlay issues with modified streams

  


  • Sync problems with live content