Recherche avancée

Sur d’autres sites (8812)

• Handling RTSP Camera Offline with Image showing issue in GStreamer

  3 janvier, par AD Thumar

  I am using below list of version.

  


  


  Gstreamer-1.24.10, windows 10.

  


  


  I’m working on a GStreamer pipeline to handle RTSP camera streams. When the RTSP camera goes offline, I want to display a image instead of stopping the pipeline. Once the RTSP stream is back online, the pipeline should automatically switch back to the live stream.

  


  Do let me know If you can solve this issue with ffmpeg.

  


  I have already tried pipeline belowed :

  


  gst-launch-1.0 compositor name=comp sink_0::xpos=0 sink_0::ypos=0 sink_1::xpos=640 sink_1::ypos=0 sink_2::xpos=1280 sink_2::ypos=0 ! videoconvert ! autovideosink input-selector name=selector1 ! queue ! videoconvert ! videoscale ! video/x-raw,width=640,height=360 ! queue ! comp. rtspsrc location=rtsp://?? latency=0 ! rtph265depay ! h265parse ! avdec_h265 ! videoconvert ! queue ! selector1.sink_0 uridecodebin uri=file:///img.jpg ! imagefreeze ! videoconvert ! videoscale ! video/x-raw,width=640,height=360 ! queue ! selector1.sink_1 input-selector name=selector2 ! queue ! videoconvert ! videoscale ! video/x-raw,width=640,height=360 ! queue ! comp. rtspsrc location=rtsp://?? latency=0 ! rtph265depay ! h265parse ! avdec_h265 ! videoconvert ! queue ! selector2.sink_0 uridecodebin uri=file:///img.jpg ! imagefreeze ! videoconvert ! videoscale ! video/x-raw,width=640,height=360 ! queue ! selector2.sink_1 input-selector name=selector3 ! queue ! videoconvert ! videoscale ! video/x-raw,width=640,height=360 ! queue ! comp. rtspsrc location=rtsp://?? latency=0 ! rtph265depay ! h265parse ! avdec_h265 ! videoconvert ! queue ! selector3.sink_0 uridecodebin uri=file:///img.jpg ! imagefreeze ! videoconvert ! videoscale ! video/x-raw,width=640,height=360 ! queue ! selector3.sink_1

  


  Can you please let me know.How to deal with this issue ??

  


• Using ffmpeg.wasm to convert an image with V360 filter without installing node.js

  6 mars 2024, par mcgyver

  I can't figure out if the problem is in my setup of ffmpeg.wasm, or in ffmpeg processing of my image, or somewhere else.

  


  I have this command line which I know does work :

  


  ffmpeg -i input.jpg -vf v360=fisheye:e:ih_fov=180:iv_fov=180 -y output.jpg


  


  I would like to implement this conversion in a web page unsing ffmpeg.wasm, so I forked this repo into mine and started playing : I edited the original transcode.html ending up in something completely standalone and functioning in github pages out of the box : transcode.html

  


  Starting from there, I split the original code in 3 parts :

  


  

  • setupFFmpeg : preloads ffmpeg stuff - ok

  


  • setName : store the data of the file selected by the user - ok

  


  • transcode : the remaining of the original code, the parts which executes the ffmpeg command, which I changed from video prcoessing to image processing. - not working

  


  


  My current file :

  


  https://github.com/jumpjack/ffmpeg.wasm-gh-pages/blob/main/public/f2e-001.html

  


  The core part, the not working one :

  


  const transcode   = async () => { 
        await ffmpeg.exec(['-hide_banner', '-i', filename, '-filter_complex', filterString,  'output.jpg']); //Execute ffmpeg command line 
        const data = await ffmpeg.readFile('output.jpg'); // Read output file from ffmpeg virtual filesystem
console.log("=== DATA=",data);
        const imgOutput = document.getElementById('myimage');
        imgOutput.src = URL.createObjectURL(new Blob([data.buffer], { type: 'image/jpeg' })); // put output into html page
      }


  


  (filterString is v360=fisheye:e:ih_fov=180:iv_fov=180 )

  


  the console.log("=== DATA=",data) ; line is never reached, I see this output :

  


  f2e.html:41 Input #0, png_pipe, from 'fisheye-mono.png':
f2e.html:41   Duration: N/A, bitrate: N/A
f2e.html:41   Stream #0:0: Video: png, rgb24(pc), 540x541 [SAR 1:1 DAR 540:541], 25 fps, 25 tbr, 25 tbn
f2e.html:41 Stream mapping:
f2e.html:41   Stream #0:0 (png) -> v360:default
f2e.html:41   v360:default -> Stream #0:0 (mjpeg)
f2e.html:41 [swscaler @ 0xe0a340] deprecated pixel format used, make sure you did set range correctly
f2e.html:41     Last message repeated 1 times
    Last message repeated 2 times
    Last message repeated 3 times
    Last message repeated 3 times
f2e.html:41 Output #0, image2, to 'output.jpg':
f2e.html:41   Metadata:
f2e.html:41     encoder         : Lavf59.27.100
f2e.html:41   Stream #0:0: Video: mjpeg, yuvj444p(pc, progressive), 1080x541 [SAR 1:1 DAR 1080:541], q=2-31, 200 kb/s, 25 fps, 25 tbn
f2e.html:41     Metadata:
f2e.html:41       encoder         : Lavc59.37.100 mjpeg
f2e.html:41     Side data:
f2e.html:41       cpb: bitrate max/min/avg: 0/0/200000 buffer size: 0 vbv_delay: N/A


  


  I don't undrestand if this is an issue of the web page or of ffmpeg : why do I get all these data about video converting, if I use a .png file as input ?

  


  Sometimes this list also finishes with a "memory error" :

  


  Uncaught (in promise) RuntimeError : memory access out of bounds

  


  Please note that I don't want to install node.js or anything else on my PC, the page must work standalone ; I am currently trying it online, I wonder if it could work also locally, but I doubt, due to cors restrictions.

  


• hardware conversion of image pixel format in ffmpeg ?

  18 janvier 2024, par dongrixinyu

  I am trying to decode an online rtmp video stream into RGB format frames, and then encoding RGB frames into an online stream.

  


  Task

  


  Here is what I do now :

  


  


  decoding a video stream to get images(RGB) ---> ai model process ---> encoding frames(RGB) to form a video stream in H264

  


  


  My scheme

  


  All my code in written in C with FFmpeg dependencies. The detailed steps are :

  


  


  rtmp/rtsp video stream ---> AVPacket ---(nvidia cuda)---> AVFrame(nv12 pix fmt) ---> AVFrame(RGB pix fmt) ---> AI process.

  


  


  


  AVFrame(RGB pix fmt) ---> AVFrame(nv12 pix fmt) ---(nvidia cuda)---> AVPacket ---> rtmp/rtsp video stream

  


  


  Now, the decoding and encoding part are run on NVIDIA GPU, which is quite fast.

  


  But the conversion of pixel format between AV_PIX_FMT_NV12 and AV_PIX_FMT_RGB is run on CPU, which is astonishingly CPU-consuming cause the size of video frame is 2k.

  


  My question

  


  So, is there any off-the-shelf method to fulfill the conversion of image pixel format on GPU (especially via cuda) directly ?