Recherche avancée

Sur d’autres sites (11947)

• Revision bcaf1d69ec : Moved call to vp9_clear_system_state() to a proper location The commit moved a

  22 juillet 2014, par Yaowu Xu

  Changed Paths :
   Modify /vp9/encoder/vp9_encoder.c

  
  
  Moved call to vp9_clear_system_state() to a proper location

  The commit moved a call to vp9_clear_system_state() to a correct
  location, i.e. prior function calls using floating point numbers.
  This was to fix a mismatch mmx code and sse2 version, where a
  floating point number used in adjust_frame_rate(cpi) gets NAN due
  to mmx registers being in wrong state.

  Change-Id : I40e0a6de98812000ccee6a729badb630604fd7e6

• How to resolve the source of an RTSP server unstable problem ?

  5 mars 2021, par EricLu

  I have already build an application by gstreamer rtsp-server with its appsrc.
It's workingflow like test-appsrc. And I implement static void need_data (GstElement * appsrc, guint unused, MyContext * ctx) function using read from a IP camera's RTSP stream cv::VideoCapture in need_data.

  


  All this function is working well and it's have worked already for a day.

  


  However, I must use some algorithm deal with the IP camera source by OpenCV. When I do this algorithm at
need_data. The RTSP Server seems unstable it always print something like the below :

  


  (opencv_rtsp_server:175041): GStreamer-CRITICAL **: 15:08:38.738: gst_mini_object_add_parent: assertion 'object != NULL' failed

(opencv_rtsp_server:175041): GStreamer-CRITICAL **: 15:08:38.739: gst_memory_get_sizes: assertion 'mem != NULL' failed

(opencv_rtsp_server:175041): GStreamer-CRITICAL **: 15:08:38.739: gst_memory_get_sizes: assertion 'mem != NULL' failed

(opencv_rtsp_server:175041): GStreamer-CRITICAL **: 15:08:38.739: gst_memory_get_sizes: assertion 'mem != NULL' failed

(opencv_rtsp_server:175041): GStreamer-CRITICAL **: 15:08:38.739: gst_mini_object_ref: assertion 'mini_object != NULL' failed


  


  It seems the RTSP-server have losing connect appsrc.

  


  One idea to reslove this problem is imporving the algorithm as it can respond as soon as framerate is.
Maybe there are some idea about caching the output of algorithm can let the appsrc element stable.
Can some one give me some advice or trahave already build a application by gstreamer rtsp-server with its appsrc.
It's workingflow like test-appsrc. And I implement static void need_data (GstElement * appsrc, guint unused, MyContext * ctx) function using read from a IP camera's RTSP stream cv::VideoCapture in need_data.

  


  All this function is working well and it's have worked already for a day.

  


  However, I must use some algorithm deal with the IP camera source by OpenCV. When I do this algorithm at
need_data. The RTSP Server seems unstable it always print something like below :

  


  (opencv_rtsp_server:175041): GStreamer-CRITICAL **: 15:08:38.738: gst_mini_object_add_parent: assertion 'object != NULL' failed

(opencv_rtsp_server:175041): GStreamer-CRITICAL **: 15:08:38.739: gst_memory_get_sizes: assertion 'mem != NULL' failed

(opencv_rtsp_server:175041): GStreamer-CRITICAL **: 15:08:38.739: gst_memory_get_sizes: assertion 'mem != NULL' failed

(opencv_rtsp_server:175041): GStreamer-CRITICAL **: 15:08:38.739: gst_memory_get_sizes: assertion 'mem != NULL' failed

(opencv_rtsp_server:175041): GStreamer-CRITICAL **: 15:08:38.739: gst_mini_object_ref: assertion 'mini_object != NULL' failed


  


  It seems the RTSP-server has losing connect appsrc.

  


  Or h264 encoder give me error :

  


  valid fps=23 [h264 @ 0x5598f7f2ba40] illegal short term buffer state detected
[h264 @ 0x5598f7f2ba40] Missing reference picture, default is 18
valid fps=24 Invalid UE golomb code
[h264 @ 0x5598f8011580] cbp too large (3199971767) at 8 95
[h264 @ 0x5598f8011580] error while decoding MB 8 95
valid fps=24 [h264 @ 0x5598f7eab3c0] corrupted macroblock 2 51 (total_coeff=-1)
[h264 @ 0x5598f7eab3c0] error while decoding MB 2 51
valid fps=25 [h264 @ 0x5598f8011580] illegal short term buffer state detected
[h264 @ 0x5598f8011580] Missing reference picture, default is 10
valid fps=24 [h264 @ 0x5598f7f2ba40] Invalid level prefix
[h264 @ 0x5598f7f2ba40] error while decoding MB 17 49
valid fps=24 Invalid UE golomb code
[h264 @ 0x5598f7fe1280] cbp too large (3199971767) at 3 163
[h264 @ 0x5598f7fe1280] error while decoding MB 3 163
valid fps=22 [h264 @ 0x5598f8011580] illegal short term buffer state detected
[h264 @ 0x5598f8011580] Missing reference picture, default is 40
valid fps=24 [h264 @ 0x5598f7fe1280] illegal short term buffer state detected
[h264 @ 0x5598f7fe1280] Missing reference picture, default is 40
valid fps=25 [h264 @ 0x5598f7eab3c0] illegal short term buffer state detected
[h264 @ 0x5598f7eab3c0] Missing reference picture, default is 36


  


  One idea to resolve this problem is improving the algorithm as it can respond as soon as framerate is.
Maybe there is some ideas about caching the output of the algorithm that can let the appsrc element stable.
Can someone give me some advice or strategy about caching it or resolve its idea ?

  


  Appreciate it so much...

  


• ffmpeg not working if run from py2app

  28 juin 2018, par dprogramz

  I’m trying to build a simple app that concats 2 mp4 files. It works fine if I run it from the command line, but if I run it from py2app app it doesn’t work. If I run the app from within the py2app in the console (eg ’dist/addTag.app/Contents/MacOS/addTag’), it works fine. It only doesn’t work if i run the app by double clicking on it. Any ideas ? code below

  #! /usr/bin/python
  
  import argparse
  
  import ffmpeg
  
  import os
  
  import shutil
  
  import sys
  
  from Tkinter import *
  
  import time
  
  
  
  fields = 'Input Video', 'Tag Video', 'Output Name'
  
  
  
  def fetch(entries, bu, lb, rt):
  
      bu['state'] = 'disabled'
  
      lb['text'] = 'working'
  
      rt.update()
  
      ffmpeg.concat(ffmpeg.input(entries[0][1].get()), ffmpeg.input(entries[1][1].get())).output(os.path.expanduser("~/desktop/")+entries[2][1].get()).run()
  
      bu['state'] = 'normal'
  
      lb['text'] = 'Ready'
  
      rt.update()
  
  
  
  
  
  def makeform(root, fields):
  
     entries = []
  
     for field in fields:
  
        row = Frame(root)
  
        lab = Label(row, width=15, text=field, anchor='w')
  
        ent = Entry(row)
  
        row.pack(side=TOP, fill=X, padx=5, pady=5)
  
        lab.pack(side=LEFT)
  
        ent.pack(side=RIGHT, expand=YES, fill=X)
  
        entries.append((field, ent))
  
     return entries
  
  
  
  if __name__ == '__main__':
  
     root = Tk()
  
     root.title("Video Maker")
  
     ents = makeform(root, fields)
  
     root.bind('<return>', (lambda event, e=ents: fetch(e))) 
  
     label = Label(root, text="Ready")
  
     label.pack(side=LEFT)  
  
     b1 = Button(root, text='Make Video',
  
            command=(lambda e=ents: fetch(e, b1, label, root)))
  
     b1.pack(side=LEFT, padx=5, pady=5)
  
     b2 = Button(root, text='Quit', command=root.quit)
  
     b2.pack(side=LEFT, padx=5, pady=5)
  
  
  
     root.mainloop()
  
  </return>