Recherche avancée

Sur d’autres sites (10787)

• Video streaming- control frame rate x264

  3 mars 2017, par H.A

  I’m new to video coding, I found this https://github.com/RafaelPalomar/H264LiveStreamer application code very useful to my project. I want to live stream a webcam over RTP . However, I’m having problem selecting the Frame Per Seconds (FPS). below is the function that controls the encoding param.

  void x264Encoder::initilize()
  
  {
  
    x264_param_default_preset(&parameters, "veryfast", "zerolatency");
  
    parameters.i_log_level =  X264_LOG_DEBUG;
  
    parameters.i_threads = 1;
  
    parameters.i_width = 640;
  
    parameters.i_height = 480;
  
    parameters.i_fps_num = 25;
  
    parameters.i_fps_den = 1;
  
    parameters.i_keyint_max = 25;
  
    parameters.b_intra_refresh = 1;
  
    parameters.rc.i_rc_method = X264_RC_CRF;
  
    parameters.rc.i_vbv_buffer_size = 1000000;
  
    parameters.rc.i_vbv_max_bitrate = 90000;
  
    parameters.rc.f_rf_constant = 25;
  
    parameters.rc.f_rf_constant_max = 35;
  
    parameters.i_sps_id = 7;
  
    // the following two value you should keep 1
  
    parameters.b_repeat_headers = 1;    // to get header before every I-Frame
  
    parameters.b_annexb = 1; // put start code in front of nal. we will remove start code later
  
    x264_param_apply_profile(&parameters, "baseline");
  
  
  
    encoder = x264_encoder_open(&parameters);
  
    x264_picture_alloc(&picture_in, X264_CSP_I420, parameters.i_width, parameters.i_height);
  
    picture_in.i_type = X264_TYPE_AUTO;
  
    picture_in.img.i_csp = X264_CSP_I420;
  
    // i have initilized my color space converter for BGR24 to YUV420 because my opencv video capture gives BGR24 image. You can initilize according to your input pixelFormat
  
    convertContext = sws_getContext(parameters.i_width,parameters.i_height, PIX_FMT_BGR24, parameters.i_width,parameters.i_height,PIX_FMT_YUV420P, SWS_FAST_BILINEAR, NULL, NULL, NULL);
  
  }

  I’m using VLC as client whatever number i set i_fps_num to, I always get 12-14 FPS. fore example, in VLC Codec tab shows 25 FPS but in Stream statistics tab it shows it only display 13 FPS on average. here is the X264 encoder output

  x264 [debug]: frame=  72 QP=22.24 NAL=2 Slice:P Poc:144 I:97   P:116  SKIP:987  size=1464 bytes
  
  x264 [info]: frame I:1     Avg QP:20.06  size: 17848
  
  x264 [info]: frame P:72    Avg QP:22.12  size:  1337
  
  x264 [info]: mb I  I16..4: 40.2%  0.0% 59.8%
  
  x264 [info]: mb P  I16..4:  3.1%  0.0%  2.3%  P16..4:  8.3%  2.0%  0.6%  0.0%  0.0%    skip:83.7%
  
  x264 [info]: coded y,uvDC,uvAC intra: 53.6% 46.4% 7.4% inter: 3.0% 1.6% 0.0%
  
  x264 [info]: i16 v,h,dc,p: 11% 65% 12% 12%
  
  x264 [info]: i4 v,h,dc,ddl,ddr,vr,hd,vl,hu: 15% 46% 16%  3%  4%  3%  5%  3%  6%
  
  x264 [info]: i8c dc,h,v,p: 46% 42% 10%  2%
  
  x264 [info]: kb/s:312.60

  the last frame number that was encoded is 72 while in VLC stream statistics it shows that 133 frames are Displayed and 66 blocks are decoded for 10 seconds streaming in addition the video playback is bursty.

  my question is how to change the x264 parameters to get 25 FPS ?

• Video streaming- control frame rate x264

  3 mars 2017, par H.A

  I’m new to video coding, I found this https://github.com/RafaelPalomar/H264LiveStreamer application code very useful to my project. I want to live stream a webcam over RTP . However, I’m having problem selecting the Frame Per Seconds (FPS). below is the function that controls the encoding param.

  void x264Encoder::initilize()
  
  {
  
    x264_param_default_preset(&parameters, "veryfast", "zerolatency");
  
    parameters.i_log_level =  X264_LOG_DEBUG;
  
    parameters.i_threads = 1;
  
    parameters.i_width = 640;
  
    parameters.i_height = 480;
  
    parameters.i_fps_num = 25;
  
    parameters.i_fps_den = 1;
  
    parameters.i_keyint_max = 25;
  
    parameters.b_intra_refresh = 1;
  
    parameters.rc.i_rc_method = X264_RC_CRF;
  
    parameters.rc.i_vbv_buffer_size = 1000000;
  
    parameters.rc.i_vbv_max_bitrate = 90000;
  
    parameters.rc.f_rf_constant = 25;
  
    parameters.rc.f_rf_constant_max = 35;
  
    parameters.i_sps_id = 7;
  
    // the following two value you should keep 1
  
    parameters.b_repeat_headers = 1;    // to get header before every I-Frame
  
    parameters.b_annexb = 1; // put start code in front of nal. we will remove start code later
  
    x264_param_apply_profile(&parameters, "baseline");
  
  
  
    encoder = x264_encoder_open(&parameters);
  
    x264_picture_alloc(&picture_in, X264_CSP_I420, parameters.i_width, parameters.i_height);
  
    picture_in.i_type = X264_TYPE_AUTO;
  
    picture_in.img.i_csp = X264_CSP_I420;
  
    // i have initilized my color space converter for BGR24 to YUV420 because my opencv video capture gives BGR24 image. You can initilize according to your input pixelFormat
  
    convertContext = sws_getContext(parameters.i_width,parameters.i_height, PIX_FMT_BGR24, parameters.i_width,parameters.i_height,PIX_FMT_YUV420P, SWS_FAST_BILINEAR, NULL, NULL, NULL);
  
  }

  I’m using VLC as client whatever number i set i_fps_num to, I always get 12-14 FPS. fore example, in VLC Codec tab shows 25 FPS but in Stream statistics tab it shows it only display 13 FPS on average. here is the X264 encoder output

  x264 [debug]: frame=  72 QP=22.24 NAL=2 Slice:P Poc:144 I:97   P:116  SKIP:987  size=1464 bytes
  
  x264 [info]: frame I:1     Avg QP:20.06  size: 17848
  
  x264 [info]: frame P:72    Avg QP:22.12  size:  1337
  
  x264 [info]: mb I  I16..4: 40.2%  0.0% 59.8%
  
  x264 [info]: mb P  I16..4:  3.1%  0.0%  2.3%  P16..4:  8.3%  2.0%  0.6%  0.0%  0.0%    skip:83.7%
  
  x264 [info]: coded y,uvDC,uvAC intra: 53.6% 46.4% 7.4% inter: 3.0% 1.6% 0.0%
  
  x264 [info]: i16 v,h,dc,p: 11% 65% 12% 12%
  
  x264 [info]: i4 v,h,dc,ddl,ddr,vr,hd,vl,hu: 15% 46% 16%  3%  4%  3%  5%  3%  6%
  
  x264 [info]: i8c dc,h,v,p: 46% 42% 10%  2%
  
  x264 [info]: kb/s:312.60

  the last frame number that was encoded is 72 while in VLC stream statistics it shows that 133 frames are Displayed and 66 blocks are decoded for 10 seconds streaming in addition the video playback is bursty.

  my question is how to change the x264 parameters to get 25 FPS ?

• Django api returns Gif as JPG despite a function to add it as video

  7 septembre 2023, par Earthling

  I'm trying to upload a .gif to my django 3.2 api. I have already ran troubleshoots through Postman and came to the conclusion that my flutter app sends it as a .gif and it gets returned as a .jpg. The problem is on the backend. Here is my relevant code which checks for file_meme subtype and then the function should convert the incoming .gif to a video :

  


  


  


  def add_media(self, file, order=None):
    check_can_add_media(post=self)

    is_in_memory_file = isinstance(file, InMemoryUploadedFile) or isinstance(file, SimpleUploadedFile)

    if is_in_memory_file:
        file_mime = magic.from_buffer(file.read())
    elif isinstance(file, TemporaryUploadedFile):
        file_mime = magic.from_file(file.temporary_file_path())
    else:
        file_mime = magic.from_file(file.name)

    check_mimetype_is_supported_media_mimetypes(file_mime)
    # Mime check moved pointer
    file.seek(0)

    file_mime_types = file_mime.split('/')

    file_mime_type = file_mime_types[0]
    file_mime_subtype = file_mime_types[1]

    temp_files_to_close = []

    if file_mime_subtype == 'gif':
                if is_in_memory_file:
                    file = write_in_memory_file_to_disk(file)

                temp_dir = tempfile.gettempdir()
                converted_gif_file_name = os.path.join(temp_dir, str(uuid.uuid4()) + '.mp4')

                ff = ffmpy.FFmpeg(
                    inputs={file.temporary_file_path() if hasattr(file, 'temporary_file_path') else file.name: None},
                    outputs={converted_gif_file_name: None})
                ff.run()
                converted_gif_file = open(converted_gif_file_name, 'rb')
                temp_files_to_close.append(converted_gif_file)
                file = File(file=converted_gif_file)
                file_mime_type = 'video'

            has_other_media = self.media.exists()
    
    if file_mime_type == 'image':
        post_image = self._add_media_image(image=file, order=order)
        if not has_other_media:
            self.media_width = post_image.width
            self.media_height = post_image.height
            self.media_thumbnail = file

    elif file_mime_type == 'video':
        post_video = self._add_media_video(video=file, order=order)
        if not has_other_media:
            self.media_width = post_video.width
            self.media_height = post_video.height
            self.media_thumbnail = post_video.thumbnail.file
    else:
        raise ValidationError(
            _('Unsupported media file type')
        )

    for file_to_close in temp_files_to_close:
        file_to_close.close()
            
            
    self.save()        

  


  


  



  def _add_media_image(self, image, order):
    return PostImage.create_post_media_image(image=image, post_id=self.pk, order=order)

def _add_media_video(self, video, order):
    return PostVideo.create_post_media_video(file=video, post_id=self.pk, order=order)


@classmethod
    def create_post_media_image(cls, image, post_id, order):
        hash = sha256sum(file=image.file)
        post_image = cls.objects.create(image=image, post_id=post_id, hash=hash, thumbnail=image)
        PostMedia.create_post_media(type=PostMedia.MEDIA_TYPE_IMAGE,
                                    content_object=post_image,
                                    post_id=post_id, order=order)
        return post_image


@classmethod
    def create_post_media_video(cls, file, post_id, order):
        hash = sha256sum(file=file.file)
        video_backend = get_backend()

        if isinstance(file, InMemoryUploadedFile):
            # If its in memory, doing read shouldn't be an issue as the file should be small.
            in_disk_file = write_in_memory_file_to_disk(file)
            thumbnail_path = video_backend.get_thumbnail(video_path=in_disk_file.name, at_time=0.0)
        else:
            thumbnail_path = video_backend.get_thumbnail(video_path=file.file.name, at_time=0.0)

        with open(thumbnail_path, 'rb+') as thumbnail_file:
            post_video = cls.objects.create(file=file, post_id=post_id, hash=hash, thumbnail=File(thumbnail_file), )
        PostMedia.create_post_media(type=PostMedia.MEDIA_TYPE_VIDEO,
                                    content_object=post_video,
                                    post_id=post_id, order=order)
        return post_video
        
    


  


  I'm not sure where the problem is. From my limited understanding, it is taking only the first frame of the .gif and uploading it as an image.