Recherche avancée

Sur d’autres sites (8540)

• how to creating V210 encoder in pure c/c++ code [closed]

  18 mai 2024, par mans

  I am trying to implement V210 video encoding by writing the encoder by myself in c/c++ and not using any library.

  


  To achieve this, I am trying to see how I can create frame data and then use FFMPEG to put the frames in a video container using this command

  


  ffmpeg -s 1280x720 -f v210 -i frames.bin -c:v copy sample_video.mkv


  


  to create the frames.bin, I am creating a list of blocks based on the specifications that I found above.

  


  The code that I am using to pack YUV to the block is as follows :

  


  #pragma pack (push,1)
class V210Block
{
public:
    void Init(uint16_t y[], uint16_t u[], uint16_t v[])
    {
        Init(y[0], u[0], v[0],
            y[1], u[1], v[1],
            y[2], u[2], v[2],
            y[3], u[3], v[3],
            y[4], u[4], v[4],
            y[5], u[5], v[5]
            );
    }

    void Init(uint16_t y0, uint16_t u0, uint16_t v0,
        uint16_t y1, uint16_t u1, uint16_t v1,
        uint16_t y2, uint16_t u2, uint16_t v2,
        uint16_t y3, uint16_t u3, uint16_t v3,
        uint16_t y4, uint16_t u4, uint16_t v4,
        uint16_t y5, uint16_t u5, uint16_t v5
    )
    {
        PackValuesToBlockbe(u0, y0, v0, 0);
        PackValuesToBlockbe(y1, u2, y2, 1);
        PackValuesToBlockbe(v2, y3, u4, 2);
        PackValuesToBlockbe(y4, v4, y5, 3);
    }
private:
    uint32_t block[4] = { 0 };
    inline void PackValuesToBlockle(uint16_t value1, uint16_t value2, uint16_t value3, int blockNo)
    {
        // Little-endian packing (least significant bits to lower memory addresses)
        block[blockNo] = (value1 << 20) | ((value2 & 0x3FF) << 10) | (value3 & 0x3FF);
    }
    inline void PackValuesToBlockbe(uint16_t value1, uint16_t value2, uint16_t value3, int blockNo)
    {
        // big-endian packing (most significant bits to lower memory addresses)
        block[blockNo] = (value3 << 20) | ((value2 & 0x3FF) << 10) | (value1 & 0x3FF);
    }

};
#pragma pack (pop)


  


  I tested both versions of PackValuesToBlock (big endian and little endian) and I got the same result.

  


  In my test, I put Y=128 and u and V to zero for all pixels in a frame and all frames in the video. When I play the video, I can see that all pixels have values of R=0, G=173, B=0

  


  Why is this happening ?

  


  Is there can I can extract a file similar to frames.bin from a video that is already encoded in this format, so I can check the binary data and find what is wrong with my encoding ?

  


  Is there any sample c code that tries to encode one or a series of images into this format ?

  


  How can I do this using OpenCV then I can check the binary data or generated bin foe with mine to find what is the problem with my code.

  


• FFMPEG - Scrambled Output converting from VOB with ffmpeg

  24 décembre 2013, par MattD

  I've concatenated a series of VOB files from a DVD into a single VOB file and I am trying to convert it to MP4 or other similar format. I see a lot of errors when converting and the output appears scrambled.

  >ffmpeg.exe" -i file.vob -sameq file.mp4
  
  FFmpeg version git-N-29181-ga304071, Copyright (c) 2000-2011 the FFmpeg develope
  
  rs
  
    built on Apr 18 2011 21:24:03 with gcc 4.5.2
  
    configuration: --enable-gpl --enable-version3 --enable-runtime-cpudetect --ena
  
  ble-memalign-hack --enable-avisynth --enable-bzlib --enable-frei0r --enable-libo
  
  pencore-amrnb --enable-libopencore-amrwb --enable-libfreetype --enable-libgsm --
  
  enable-libmp3lame --enable-libopenjpeg --enable-librtmp --enable-libschroedinger
  
   --enable-libspeex --enable-libtheora --enable-libvorbis --enable-libvpx --enabl
  
  e-libx264 --enable-libxavs --enable-libxvid --enable-zlib --cross-prefix=i686-w6
  
  4-mingw32- --target-os=mingw32 --arch=x86_32 --extra-cflags=-I/home/kyle/softwar
  
  e/ffmpeg/external-libraries/win32/include --extra-ldflags=-L/home/kyle/software/
  
  ffmpeg/external-libraries/win32/lib --pkg-config=pkg-config
  
    libavutil    50. 40. 1 / 50. 40. 1
  
    libavcodec   52.120. 0 / 52.120. 0
  
    libavformat  52.108. 0 / 52.108. 0
  
    libavdevice  52.  4. 0 / 52.  4. 0
  
    libavfilter   1. 79. 0 /  1. 79. 0
  
    libswscale    0. 13. 0 /  0. 13. 0
  
  [mpeg2video @ 01751A90] ac-tex damaged at 5 16
  
  [mpeg2video @ 01751A90] invalid mb type in I Frame at 0 1
  
  [mpeg2video @ 01751A90] invalid mb type in I Frame at 0 2
  
  [mpeg2video @ 01751A90] invalid mb type in I Frame at 0 3
  
  [mpeg2video @ 01751A90] invalid mb type in I Frame at 0 4
  
  
  
  [...]

  I'm guessing this is CSS scrambling and I need to do some sort of DeCSS. Does FFMpeg have an option for this ? Cringe if you'd like, but is there C# source to achieve this ?

  My end goal is really just to get some DVDs that I own onto my media server. I've tried a few demo-ware products with limited success including Acala DVD Ripper, Click to Disk, AVIDemux to name a few. I even paid for a couple of them to get the full version, but Acala only works for about half of my DVDs and Click to Disk can decode the VOB, but I need to use FFMpeg to convert. I'd like to have it all in one app that works and I am willing to write some code for it.

• Incorrect length of video produced with ffmpeg libraries [closed]

  28 mars, par ivan.ukr

  I'm writing a C program that takes series of PNG images and converts them into a video. Video consists of the initial black screen and then each of those images, shown for the same constant amount of time, say 200 ms. I'm using libx264 as codec and mp4 as output format. I'm compiling my program with GCC 12 on Ubuntu 22.04 LTS. I'm using ffmpeg version from Ubuntu repositories.
In order to achieve above behavior I've set time base to 1/5 in the both stream and codec.

  


  // assume imageDuration = 200
AVRational timeBase;
av_reduce(&timeBase.num, &timeBase.den, imageDuration, 1000, INT_MAX);

const AVCodec *c = avcodec_find_encoder(codecId);
AVStream *s = avformat_new_stream(fc, c);
s->time_base = timeBase;
s->nb_frames = numImages + 1; // inital black screen + images
s->duration = numImages + 1;

AVCodecContext *cc = avcodec_alloc_context3(c);
cc->width = width;
cc->height = height;
cc->pix_fmt = pixelFormat;
cc->time_base = timeBase;

// ffmpeg headers suggest: Set to time_base ticks per frame.
// Default 1, e.g., H.264/MPEG-2 set it to 2.
cc->ticks_per_frame = 2;

cc->framerate = av_inv_q(timeBase);
if (fc->oformat->flags & AVFMT_GLOBALHEADER) {
    cc->flags |= AV_CODEC_FLAG_GLOBAL_HEADER;
}



  


  Then I'm encoding 11 frames.

  


  Finally, I'm getting video with the following characteristics :

  


  $ ffprobe v.mp4

....

Input #0, mov,mp4,m4a,3gp,3g2,mj2, from 'v.mp4':
  Metadata:
    major_brand     : isom
    minor_version   : 512
    compatible_brands: isomiso2avc1mp41
    encoder         : Lavf58.76.100
  Duration: 00:00:00.01, start: 0.000000, bitrate: 68414 kb/s
  Stream #0:0(und): Video: h264 (High) (avc1 / 0x31637661), yuv420p, 640x360, 69073 kb/s, 10333.94 fps, 10240 tbr, 10240 tbn, 10 tbc (default)
    Metadata:
      handler_name    : VideoHandler
      vendor_id       : [0][0][0][0]



  


  Please pay attention to :

  


  Duration: 00:00:00.01


  


  and

  


  10333.94 fps


  


  That's totally NOT what I've expected (2.2s video length and 5 fps frame rate).

  


  Note : The content of video is correct, this can be verified by looking into the generated video file frame-by-frame in some program like Avidemux. But video length and frame rate are incorrect.

  


  Please advise, how to fix this ?