Recherche avancée

Sur d’autres sites (10462)

• Build Opencv3.1 with ffmpeg

  10 mai 2018, par batuman

  I tried to build Opencv3.1 at Ubuntu16.04 with ffmpeg.
  I follow the instructions how to install ffmpeg at here.

  According to this discussion ,I checked.

  >> mediainfo ~/Desktop/grb_2.mp4 
  
  General
  
  Complete name                            : /home/nyan/Desktop/grb_2.mp4
  
  Format                                   : MPEG-4
  
  Format profile                           : Base Media
  
  Codec ID                                 : isom (isom/iso2/avc1/mp41)
  
  File size                                : 445 KiB
  
  Duration                                 : 27s 862ms
  
  Overall bit rate                         : 131 Kbps
  
  Encoded date                             : UTC 1904-01-01 00:00:00
  
  Tagged date                              : UTC 1904-01-01 00:00:00
  
  Writing application                      : Lavf57.66.105
  
  
  
  Video
  
  ID                                       : 1
  
  Format                                   : AVC
  
  Format/Info                              : Advanced Video Codec
  
  Format profile                           : High@L3
  
  Format settings, CABAC                   : Yes
  
  Format settings, ReFrames                : 4 frames
  
  Codec ID                                 : avc1
  
  Codec ID/Info                            : Advanced Video Coding
  
  Duration                                 : 27s 862ms
  
  Bit rate                                 : 128 Kbps
  
  Width                                    : 720 pixels
  
  Height                                   : 480 pixels
  
  Display aspect ratio                     : 4:3
  
  Original display aspect ratio            : 4:3
  
  Frame rate mode                          : Constant
  
  Frame rate                               : 29.970 (30000/1001) fps
  
  Standard                                 : NTSC
  
  Color space                              : YUV
  
  Chroma subsampling                       : 4:2:0
  
  Bit depth                                : 8 bits
  
  Scan type                                : Progressive
  
  Bits/(Pixel*Frame)                       : 0.012
  
  Stream size                              : 435 KiB (98%)
  
  Writing library                          : x264 core 148
  
  Encoding settings                        : cabac=1 / ref=3 / deblock=1:0:0 / analyse=0x3:0x113 / me=hex / subme=7 / psy=1 / psy_rd=1.00:0.00 / mixed_ref=1 / me_range=16 / chroma_me=1 / trellis=1 / 8x8dct=1 / cqm=0 / deadzone=21,11 / fast_pskip=1 / chroma_qp_offset=-2 / threads=12 / lookahead_threads=2 / sliced_threads=0 / nr=0 / decimate=1 / interlaced=0 / bluray_compat=0 / constrained_intra=0 / bframes=3 / b_pyramid=2 / b_adapt=1 / b_bias=0 / direct=1 / weightb=1 / open_gop=0 / weightp=2 / keyint=250 / keyint_min=25 / scenecut=40 / intra_refresh=0 / rc_lookahead=40 / rc=crf / mbtree=1 / crf=23.0 / qcomp=0.60 / qpmin=0 / qpmax=69 / qpstep=4 / ip_ratio=1.40 / aq=1:1.00
  
  Encoded date                             : UTC 1904-01-01 00:00:00
  
  Tagged date                              : UTC 1904-01-01 00:00:00 
  
  
  
  
  
  
  
  
  
  
  
  ffmpeg -codecs | grep -i avc
  
  ffmpeg version N-90982-gb995ec0 Copyright (c) 2000-2018 the FFmpeg developers
  
    built with gcc 5.4.0 (Ubuntu 5.4.0-6ubuntu1~16.04.9) 20160609
  
    configuration: --prefix=/home/nyan/ffmpeg_build --enable-shared --extra-cflags=-I/home/nyan/ffmpeg_build/include --extra-ldflags=-L/home/nyan/ffmpeg_build/lib --extra-libs='-lpthread -lm' --bindir=/home/nyan/bin --enable-gpl --enable-libass --enable-libfdk-aac --enable-libfreetype --enable-libmp3lame --enable-libopus --enable-libtheora --enable-libvorbis --enable-libvpx --enable-libx264 --enable-libx265 --enable-nonfree
  
    libavutil      56. 18.100 / 56. 18.100
  
    libavcodec     58. 19.100 / 58. 19.100
  
    libavformat    58. 13.101 / 58. 13.101
  
    libavdevice    58.  4.100 / 58.  4.100
  
    libavfilter     7. 21.100 /  7. 21.100
  
    libswscale      5.  2.100 /  5.  2.100
  
    libswresample   3.  2.100 /  3.  2.100
  
    libpostproc    55.  2.100 / 55.  2.100
  
   DEV.LS h264                 H.264 / AVC / MPEG-4 AVC / MPEG-4 part 10 (decoders: h264 h264_v4l2m2m ) (encoders: libx264 libx264rgb h264_v4l2m2m h264_vaapi )
  
   D.A.L. avc                  On2 Audio for Video Codec (decoders: on2avc )

  So my ffmpeg installation is ok.

  When I cmake to my Opencv, I have set up to ffmpeg as

  

  But when I make to compile, I have error as

  make[2]: *** No rule to make target '/home/nyan/ffmpeg_build/lib/libavresample.a', needed by 'lib/libopencv_videoio.so.3.1.0'.  Stop.
  
  CMakeFiles/Makefile2:8709: recipe for target 'modules/videoio/CMakeFiles/opencv_videoio.dir/all' failed
  
  make[1]: *** [modules/videoio/CMakeFiles/opencv_videoio.dir/all] Error 2
  
  Makefile:160: recipe for target 'all' failed
  
  make: *** [all] Error 2

  How can I fix the problem ?

  I install libavresample manually and

  sudo apt-get install libavresample-dev
  
  [sudo] password for nyan: 
  
  Reading package lists... Done
  
  Building dependency tree       
  
  Reading state information... Done
  
  libavresample-dev is already the newest version (7:2.8.14-0ubuntu0.16.04.1).

  It is newest version.

  EDIT:

  Actually libavresample at ffmpeg is deprecated already.
  So I tried to install Opencv3.4.1 higher version.
  Compile, build and installation are fine.
  But VideoCapture cap(0); does not read from the device.
  So I checked ffmpeg at Opencv3.4.1 installation as

  python -c "import cv2; print(cv2.getBuildInformation())" | grep -i ffmpeg

  FFMMPEG IS NO. So build with ffmpeg is not successful even though there was no error at build for Opencv3.4.1.

• FFmpeg decoding subtitle (delphi) decoding resutl > 0 but values are zero/null

  27 janvier 2020, par coban

  while True do begin
  
  if vst.sub_q.size = 0
  
  then Continue;
  
  
  
  ret := mpl.packet_queue_get(vst.sub_q, pkt, 0);
  
  if(pkt.data = mpl.flush_pkt.data) then
  
  begin
  
    avcodec_flush_buffers(vst.sub_st.codec);
  
    continue;
  
  end;
  
  
  
  
  
  ret := avcodec_decode_subtitle2(vst.sub_codec_ctx, @subt, @got_sub, @pkt);
  
  if ret >= 0
  
  then begin
  
    if got_sub = 1
  
    then begin
  
      for i := 0 to subt.num_rects-1
  
      do begin
  
         s := subt.rects[i].text;
  
         if (subt.rects[i].text <> '') or
  
            (subt.rects[i].ass <> '')
  
         then  raise Exception.Create(s);
  
      end;
  
  
  
    end;
  
  end;
  
  
  
  av_packet_unref(pkt);
  
  end;

  I’m trying to build a video player using ffmpeg + sdl2 with Delphi. I found an example for this which did play video without problems but had problems with audio. Audio is solved by inspecting the ffplay example in c/c++ and translating the audio part to Delphi. I tried to translate complete ffplay but I had strange behavior on video part, it starts playing but after a few seconds audio is going on without problems, while video is getting green screen.

  Something is not going well while calling "decoder_decode_frame" function or "video_refresh" in case of video shomehow ?
  However, I am, so far, able to decode/play audio and video streams synchronized. At the moment picture/video quality can be better, but this is for later.

  The problem :
  I am using the code above to decode the subtitle stream.
  initializing and opening the codec is succeeds "avcodec_open2", queueing packets is succeeds, retrieving packet succeeds at the right moment when it should be displayed.

  "avcodec_decode_subtitle2" function succeeds, result is > 0, got_frame variable is assigned "1", avsubtitle(subt) is being assigned.
  I do not get any errors or warnings everyting seems to be correct except the value of avsubtitle.

  subt format is 1 this seems to be the problem, inspecting ffplay.c
  subt start_display_time is 0
  subt end_display_time is 0
  subt num_rects is 1

  rects ppAVSubtitleRect is assigned with values zero/nil
  rects type_ is subtitle_none

  Playing a *.mkv file, subtitle information from codec is
  codec name ’srt’
  codec long_name ’SubRip subtitle’

  ---

  If this type of subtitle is not supported, why can we decode it without any warnings or errors ?
  If it is possible to get the subtitle text, how can I achieve this ?

  ---

  AVSubtitle = record
  
  format: uint16_t;             (* 0 = graphics *)
  
  start_display_time: uint32_t; (* relative to packet pts, in ms *)
  
  end_display_time: uint32_t;   (* relative to packet pts, in ms *)
  
  num_rects: unsigned;
  
  rects: ppAVSubtitleRect;
  
  pts: int64_t; // < Same as packet pts, in AV_TIME_BASE
  
  end;
  
  
  
  AVSubtitleType = (                //
  
  SUBTITLE_NONE, SUBTITLE_BITMAP, // < A bitmap, pict will be set
  
  (*
  
    * Plain text, the text field must be set by the decoder and is
  
    * authoritative. ass and pict fields may contain approximations.
  
  *)
  
  SUBTITLE_TEXT,
  
  (*
  
    * Formatted text, the ass field must be set by the decoder and is
  
    * authoritative. pict and text fields may contain approximations.
  
  *)
  
  SUBTITLE_ASS);
  
  
  
  AVSubtitleRect = record
  
    x: int;         // < top left corner  of pict, undefined when pict is not set
  
    y: int;         // < top left corner  of pict, undefined when pict is not set
  
    w: int;         // < width            of pict, undefined when pict is not set
  
    h: int;         // < height           of pict, undefined when pict is not set
  
    nb_colors: int; // < number of colors in pict, undefined when pict is not set
  
  
  
   {$IFDEF FF_API_AVPICTURE}
  
   (*
  
    * @deprecated unused
  
   *)
  
   // attribute_deprecated
  
   pict: AVPicture deprecated;
  
   {$ENDIF}
  
   (*
  
    * data+linesize for the bitmap of this subtitle.
  
    * Can be set for text/ass as well once they are rendered.
  
   *)
  
   data: puint8_t_array_4;
  
   linesize: Tint_array_4;
  
  
  
   _type: AVSubtitleType;
  
  
  
   text: pAnsiChar; // < 0 terminated plain UTF-8 text
  
  
  
   (*
  
    * 0 terminated ASS/SSA compatible event line.
  
    * The presentation of this is unaffected by the other values in this
  
    * struct.
  
   *)
  
   ass: pAnsiChar;
  
  
  
   flags: int;
  
  end;

  -----------edit----------------------

  SetLength(buf, 0);
  
  for i := 0 to pkt.size-1
  
  do begin
  
    Insert(pkt.data[i], buf, length(buf));
  
  end;
  
  
  
  s := TEncoding.UTF8.GetString(buf);

  It seems that in this case the text is stored in pkt.data, I don’t think this is the right way to get the subtitle text but I wonder why avsubtitle doesn’t get the values. avpacket values seem like, (pts, dts, duration) etc, to be correct.
  If I should use such construction to get the subtitle (text)value, how can I detect if it is utf8 encoded or how to determine the encoding ?
  There are more questions to be answered like, the position of the subtitle. The bad thing about ffmpeg is it’s almost impossible to get answers.

  ------------------edi2----------------------------------
  Ok how I am solving this, is creating in memory a png image with transparency and writing text over it, converting the image to PAVRAME and show at the given pts time + duration. For the positioning is chosen horizontal center for each line and vertically the bottom minus X pixels. Looks like it’s working so far.

  Next ; the formatting of the text, like bold italic etc. Can anybody give a hint if there is a quick solution(function/method/example), I will try to create one, but a quick solution is welcome. As long as there is no nested formatting is used, it doesn’t sound tricky, but I’ve read that nested formatting can be used.

  Should it be that this(formatted text) the reason is why ffmpeg is not assigning any values to avsubtitle ?

• lavu/x86 : add FFT assembly

  10 avril 2021, par Lynne

  lavu/x86 : add FFT assembly
  This commit adds a pure x86 assembly SIMD version of the FFT in libavutil/tx.
  
  The design of this pure assembly FFT is pretty unconventional.
  On the lowest level, instead of splitting the complex numbers into
  
  real and imaginary parts, we keep complex numbers together but split
  
  them in terms of parity. This saves a number of shuffles in each transform,
  
  but more importantly, it splits each transform into two independent
  
  paths, which we process using separate registers in parallel.
  
  This allows us to keep all units saturated and lets us use all available
  
  registers to avoid dependencies.
  
  Moreover, it allows us to double the granularity of our per-load permutation,
  
  skipping many expensive lookups and allowing us to use just 4 loads per register,
  
  rather than 8, or in case FMA3 (and by extension, AVX2), use the vgatherdpd
  
  instruction, which is at least as fast as 4 separate loads on old hardware,
  
  and quite a bit faster on modern CPUs).
  Higher up, we go for a bottom-up construction of large transforms, foregoing
  
  the traditional per-transform call-return recursion chains. Instead, we always
  
  start at the bottom-most basis transform (in this case, a 32-point transform),
  
  and continue constructing larger and larger transforms until we return to the
  
  top-most transform.
  
  This way, we only touch the stack 3 times per a complete target transform :
  
  once for the 1/2 length transform and two times for the 1/4 length transform.
  The combination algorithm we use is a standard Split-Radix algorithm,
  
  as used in our C code. Although a version with less operations exists
  
  (Steven G. Johnson and Matteo Frigo's "A modified split-radix FFT with fewer
  
  arithmetic operations", IEEE Trans. Signal Process. 55 (1), 111–119 (2007),
  
  which is the one FFTW uses), it only has 2% less operations and requires at least 4x
  
  the binary code (due to it needing 4 different paths to do a single transform).
  
  That version also has other issues which prevent it from being implemented
  
  with SIMD code as efficiently, which makes it lose the marginal gains it offered,
  
  and cannot be performed bottom-up, requiring many recursive call-return chains,
  
  whose overhead adds up.
  We go through a lot of effort to minimize load/stores by keeping as much in
  
  registers in between construcring transforms. This saves us around 32 cycles,
  
  on paper, but in reality a lot more due to load/store aliasing (a load from a
  
  memory location cannot be issued while there's a store pending, and there are
  
  only so many (2 for Zen 3) load/store units in a CPU).
  
  Also, we interleave coefficients during the last stage to save on a store+load
  
  per register.
  Each of the smallest, basis transforms (4, 8 and 16-point in our case)
  
  has been extremely optimized. Our 8-point transform is barely 20 instructions
  
  in total, beating our old implementation 8-point transform by 1 instruction.
  
  Our 2x8-point transform is 23 instructions, beating our old implementation by
  
  6 instruction and needing 50% less cycles. Our 16-point transform's combination
  
  code takes slightly more instructions than our old implementation, but makes up
  
  for it by requiring a lot less arithmetic operations.
  Overall, the transform was optimized for the timings of Zen 3, which at the
  
  time of writing has the most IPC from all documented CPUs. Shuffles were
  
  preferred over arithmetic operations due to their 1/0.5 latency/throughput.
  On average, this code is 30% faster than our old libavcodec implementation.
  
  It's able to trade blows with the previously-untouchable FFTW on small transforms,
  
  and due to its tiny size and better prediction, outdoes FFTW on larger transforms
  
  by 11% on the largest currently supported size.
  

  • [DH] libavutil/tx.c
  • [DH] libavutil/tx_priv.h
  • [DH] libavutil/x86/Makefile
  • [DH] libavutil/x86/tx_float.asm
  • [DH] libavutil/x86/tx_float_init.c