Recherche avancée

Sur d’autres sites (10598)

• Publishing an RTSP stream from my IP camera to a remote MediaMTX / RTSP Simple Server using FFmpeg and Azure [closed]

  28 janvier 2024, par cmd

  Using FFmpeg, I am trying to publish the RTSP stream from my IP camera to an Azure VM running a MediaMTX instance https://github.com/bluenviron/mediamtx. My intention is for the stream to be accessible from other remote connections by connecting to the VM.

  


  I was able to setup a local MediaMTX server on my laptopm, and publish the stream using this FFmpeg command :

  


  ffmpeg -i rtsp://<camera username="username">:<camera password="password">@<camera local="local" ip="ip">:554 -c:v copy -c:a copy -f rtsp rtsp://<laptop local="local" ip="ip">:8554/stream/mystream&#xA;</laptop></camera></camera></camera>

  


  This worked fine, as I was able to connect to the MediaMTX server in VLC player to view the stream. However, I have setup an Azure VM running Windows 10, and run the same MediaMTX instance on it. I have attempted to publish the RTSP stream from the IP camera using the same command, but with the Laptop IP replaced with the address of the VM, but this doesn't work.

  


  I have disabled the Windows Firewall on my laptop and the VM, but the stream can still not be published. I have also tried the same approach by running the MediaMTX server on my friend's computer on his network, and adding the necessary port forwarding rules to his router. I am unable to edit any rules on the router where I am living however.

  


  What else might be causing this issue, and is there any way to publish the IP camera's RTSP to a remote server where it could then be read ?

  


• FFMPEG : How do I maintain the aspect ratio of various images used in a slideshow ?

  7 mars 2019, par Kimberly W

  I’m just starting out with FFMPEG and trying to use it to make a slideshow. Ideally, I’d like to get it where I can input an arbitrary number of images, and each image is shown for (example) 2 seconds and them moves on to the next. Each image also maintains it’s original aspect ratio and is not stretched in any way (they can of course be scaled up/down to fit the resolution of the output video).

  I started off with a basic command like the following.

  ffmpeg -r 1/9 -pattern_type glob -i "*.jpg" -c:v libx264 -y -pix_fmt yuv420p -vf scale="720:trunc(ow/a/2)*2" out.mp4

  In this example, the first image is a wide (landscape) image and some of the others are tall (portrait) images. The portrait ones get squished in output video. Also the images aren’t displayed for equal amount of times.

  A couple of problems I’ve been running into :

  1. There’s an error for width not divisible by 2, because the images can be literally any random width. To resolve that, I’ve been trying various -vf options I’ve found through googling (like the one above). They take care of the error, but don’t solve my aspect ratio issue.
  2. All the images seem to be stretch or squished to fit the dimensions of the first input image. In reality, the images are of various different dimensions (like frames in a comic book). There’s no pattern to them.

  Is there a ffmpeg command for just taking images and creating as slideshow, while preserving their original aspect ratios ?

• av_read_frame() in ffmpeg returns -5 while receiving UDP data [closed]

  26 février 2024, par zttang

  I'm working on an app which is using ffmpeg lib. There is a living source keep sending TS stream thru udp. Then I allocated a thread use ffmpeg receive the TS packets. the code for this thread is like :

  


  void* av_source_thread(void *data) {
    char udp_url[50];
    snprintf(udp_url, sizeof(udp_url), "udp://127.0.0.1:12345");
    AVDictionary* options = NULL;
    av_dict_set(&options, "timeout", "500000", 0);  // timeout=0.5s
    av_dict_set(&options, "overrun_nonfatal", "1", 0);
    av_dict_set(&options, "fifo_size", "278876", 0);  //50MB

    AVFormatContext *ffmpeg_source = avformat_alloc_context();
    while (running) {
        if (avformat_open_input(&ffmpeg_source, udp_url, NULL, &options) != 0) {
            continue;
        } else {
            break;
        }
    }
    // Some code fill codec type and alloc context here

    av_format_inject_global_side_data(ffmpeg_source);
    AVPacket *packet = av_packet_alloc();
    while (running) {
        ret = av_read_frame(ffmpeg_source, packet);
        if (ret < 0) {
            char errbuf[AV_ERROR_MAX_STRING_SIZE];
            av_strerror(ret, errbuf, AV_ERROR_MAX_STRING_SIZE);
            usleep(10000);
            log("av_read_frame failed, Exit, %s, %d", errbuf, ret);
            continue;
        }
        putPkt2Q(packet);  
    }
    av_packet_free(&packet);
    avformat_close_input(&ffmpeg_source);
    av_dict_free(&options);
    return nullptr;
}


  


  Then the packets will be sent to Q and other thread will process them. but when I run the program, the av_read_frame will return -5(I/O error) from time to time. and once it returned -5, it can not recover unless I restart this thread.

  


  since this is a real time source, so lost some frames are acceptable, I just want to know how to avoid this kind of issue or how to recover without restart the whole thread. I tried to increase the fifo_size in options, but it does not work.