Recherche avancée

Sur d’autres sites (6821)

• ffmpeg 1FPS extraction call spits out infinite images until disk is bricked on poorly-encoded movie

  12 juillet 2017, par John Allard

  This is one of the oddest things I’ve ever seen while using ffmpeg.

  This is the scenario - we are getting video footage from a IONODE ION-E100 encoder, the footage is encoded in h264. I can view the footage fine in VLC, and I can copy it over to another video using ffmpeg and -c:v copy. The issue arises when I try and extract one FPS from the movie using the following command :

  ffmpeg -i testmovie.mp4 -r 1 -vf 640:-2 -q:v 18 -loglevel error /tmp/tmp/extraction_%04d.jpeg

  What this should do is go through the movie and extract one frame per second, then dump the frames to the files /tmp/tmp/extraction_0001.jpeg, /tmp/tmp/extraction_0002.jpeg, ..., /tmp/tmp/extraction_000X.jpeg. It should end when it reaches the end of the video.

  Now the odd thing that happens when I run this command is that it will run in a tight loop, writing tens or hundreds of thousands of images to the disk until the disk is full at which point the command fails. All of the extracted images are the exact same as one another.

  Here is the ffmpeg stats output on the video

  ffmpeg version 3.2.2 Copyright (c) 2000-2016 the FFmpeg developers
  
    built with Apple LLVM version 8.0.0 (clang-800.0.42.1)
  
    configuration: --prefix=/usr/local/Cellar/ffmpeg/3.2.2 --enable-shared --enable-pthreads --enable-gpl --enable-version3 --enable-hardcoded-tables --enable-avresample --cc=clang --host-cflags= --host-ldflags= --enable-libmp3lame --enable-libx264 --enable-libxvid --enable-opencl --disable-lzma --enable-vda
  
    libavutil      55. 34.100 / 55. 34.100
  
    libavcodec     57. 64.101 / 57. 64.101
  
    libavformat    57. 56.100 / 57. 56.100
  
    libavdevice    57.  1.100 / 57.  1.100
  
    libavfilter     6. 65.100 /  6. 65.100
  
    libavresample   3.  1.  0 /  3.  1.  0
  
    libswscale      4.  2.100 /  4.  2.100
  
    libswresample   2.  3.100 /  2.  3.100
  
    libpostproc    54.  1.100 / 54.  1.100
  
  Input #0, mov,mp4,m4a,3gp,3g2,mj2, from '/Users/john/ionode/test2.mp4':
  
    Metadata:
  
      major_brand     : isom
  
      minor_version   : 512
  
      compatible_brands: isomiso2avc1mp41
  
      title           : /videoinput_1:0/h264_1/media.stm
  
      encoder         : Lavf57.56.100
  
    Duration: 00:00:16.52, start: -9791.427750, bitrate: 2165 kb/s
  
      Stream #0:0(und): Video: h264 (High) (avc1 / 0x31637661), yuv420p, 1920x1080 [SAR 1:1 DAR 16:9], 2119 kb/s, 25 fps, 25 tbr, 12800 tbn, 50 tbc (default)
  
      Metadata:
  
        handler_name    : VideoHandler
  
      Stream #0:1(und): Audio: aac (LC) (mp4a / 0x6134706D), 8000 Hz, mono, fltp, 236 kb/s (default)
  
      Metadata:
  
        handler_name    : SoundHandler

  (notice the start: -9791.427750 value, what is this ?)

  Here are the encoding details on the footage :

  {
  
      "programs": [
  
  
  
      ],
  
      "streams": [
  
          {
  
              "index": 0,
  
              "codec_name": "h264",
  
              "codec_long_name": "H.264 / AVC / MPEG-4 AVC / MPEG-4 part 10",
  
              "profile": "High",
  
              "codec_type": "video",
  
              "codec_time_base": "1/50",
  
              "codec_tag_string": "avc1",
  
              "codec_tag": "0x31637661",
  
              "width": 1920,
  
              "height": 1080,
  
              "coded_width": 1920,
  
              "coded_height": 1080,
  
              "has_b_frames": 2,
  
              "sample_aspect_ratio": "1:1",
  
              "display_aspect_ratio": "16:9",
  
              "pix_fmt": "yuv420p",
  
              "level": 40,
  
              "chroma_location": "left",
  
              "refs": 1,
  
              "is_avc": "true",
  
              "nal_length_size": "4",
  
              "r_frame_rate": "25/1",
  
              "avg_frame_rate": "25/1",
  
              "time_base": "1/12800",
  
              "start_pts": 0,
  
              "start_time": "0.000000",
  
              "duration_ts": 211456,
  
              "duration": "16.520000",
  
              "bit_rate": "2119813",
  
              "bits_per_raw_sample": "8",
  
              "nb_frames": "413",
  
              "disposition": {
  
                  "default": 1,
  
                  "dub": 0,
  
                  "original": 0,
  
                  "comment": 0,
  
                  "lyrics": 0,
  
                  "karaoke": 0,
  
                  "forced": 0,
  
                  "hearing_impaired": 0,
  
                  "visual_impaired": 0,
  
                  "clean_effects": 0,
  
                  "attached_pic": 0,
  
                  "timed_thumbnails": 0
  
              },
  
              "tags": {
  
                  "language": "und",
  
                  "handler_name": "VideoHandler"
  
              }
  
          }
  
      ]
  
  }

  Here a link to the sample video to use : http://s000.tinyupload.com/?file_id=03169189167771012515
  (it’s a 5MB video file)

  The only clues I have as to what is actually going on is the error output
  from the command that grabs the video from the encoder :

  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823640, current: -630275919; changing to -629823639. This may result in incorrect timestamps in the output file.
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823639, current: -630274895; changing to -629823638. This may result in incorrect timestamps in the output file.
  
  frame=  124 fps= 21 q=28.0 size=      41kB time=00:00:02.60 bitrate= 128.8kbits/s dup=26 drop=4 spee[mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823638, current: -630273871; changing to -629823637. This may result in incorrect timestamps in the output file.
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823637, current: -630272847; changing to -629823636. This may result in incorrect timestamps in the output file.
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823636, current: -630271824; changing to -629823635. This may result in incorrect timestamps in the output file.
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823635, current: -630270800; changing to -629823634. This may result in incorrect timestamps in the output file.
  
  frame=  124 fps= 19 q=28.0 size=      43kB time=00:00:02.60 bitrate= 137.0kbits/s dup=26 drop=4 spee[mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823634, current: -630269776; changing to -629823633. This may result in incorrect timestamps in the output file.
  
  [aac @ 0x7faa1a832000] Queue input is backward in time
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823633, current: -630268752; changing to -629823632. This may result in incorrect timestamps in the output file.
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823632, current: -630267728; changing to -629823631. This may result in incorrect timestamps in the output file.
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823631, current: -630513833; changing to -629823630. This may result in incorrect timestamps in the output file.
  
  frame=  124 fps= 18 q=28.0 size=      46kB time=00:00:02.60 bitrate= 145.4kbits/s dup=26 drop=4 spee[mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823630, current: -630512810; changing to -629823629. This may result in incorrect timestamps in the output file.
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823629, current: -630511786; changing to -629823628. This may result in incorrect timestamps in the output file.
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823628, current: -630510762; changing to -629823627. This may result in incorrect timestamps in the output file.
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823627, current: -630509738; changing to -629823626. This may result in incorrect timestamps in the output file.
  
  frame=  124 fps= 17 q=28.0 size=      49kB time=00:00:02.60 bitrate= 153.6kbits/s dup=26 drop=4 spee[mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823626, current: -630508714; changing to -629823625. This may result in incorrect timestamps in the output file.
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823625, current: -630507690; changing to -629823624. This may result in incorrect timestamps in the output file.
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823624, current: -630506667; changing to -629823623. This may result in incorrect timestamps in the output file.
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823623, current: -630505643; changing to -629823622. This may result in incorrect timestamps in the output file.
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823622, current: -630504619; changing to -629823621. This may result in incorrect timestamps in the output file.
  
  frame=  124 fps= 16 q=28.0 size=      52kB time=00:00:02.60 bitrate= 163.5kbits/s dup=26 drop=4 spee[mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823621, current: -630503595; changing to -629823620. This may result in incorrect timestamps in the output file.
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823620, current: -630502571; changing to -629823619. This may result in incorrect timestamps in the output file.
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823619, current: -630501548; changing to -629823618. This may result in incorrect timestamps in the output file.
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823618, current: -630500524; changing to -629823617. This may result in incorrect timestamps in the output file.
  
  frame=  124 fps= 15 q=28.0 size=      54kB time=00:00:02.60 bitrate= 171.4kbits/s dup=26 drop=4 spee[mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823617, current: -630499500; changing to -629823616. This may result in incorrect timestamps in the output file.
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823616, current: -630498476; changing to -629823615. This may result in incorrect timestamps in the output file.
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823615, current: -630497452; changing to -629823614. This may result in incorrect timestamps in the output file.
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823614, current: -630496428; changing to -629823613. This may result in incorrect timestamps in the output file.
  
  frame=  124 fps= 14 q=28.0 size=      57kB time=00:00:02.60 bitrate= 179.3kbits/s dup=26 drop=4 spee[mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823613, current: -630495405; changing to -629823612. This may result in incorrect timestamps in the output file.
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823612, current: -630494381; changing to -629823611. This may result in incorrect timestamps in the output file.
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823611, current: -630493357; changing to -629823610. This may result in incorrect timestamps in the output file.
  
  frame=  124 fps= 13 q=28.0 size=      59kB time=00:00:02.60 bitrate= 185.6kbits/s dup=26 drop=4 spee[aac @ 0x7faa1a832000] Queue input is backward in time
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823610, current: -630492333; changing to -629823609. This may result in incorrect timestamps in the output file.
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823609, current: -630491309; changing to -629823608. This may result in incorrect timestamps in the output file.
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823608, current: -630736398; changing to -629823607. This may result in incorrect timestamps in the output file.
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823607, current: -630735375; changing to -629823606. This may result in incorrect timestamps in the output file.
  
  frame=  124 fps= 12 q=28.0 size=      61kB time=00:00:02.60 bitrate= 193.5kbits/s dup=26 drop=4 spee[mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823606, current: -630734351; changing to -629823605. This may result in incorrect timestamps in the output file.
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823605, current: -630733327; changing to -629823604. This may result in incorrect timestamps in the output file.
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823604, current: -630732303; changing to -629823603. This may result in incorrect timestamps in the output file.
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823603, current: -630731279; changing to -629823602. This may result in incorrect timestamps in the output file.
  
  [mp4 @ 0x7faa1a83ba00] Non-monotonous DTS in output stream 0:1; previous: -629823602, current: -630730256; changing to -629823601. This may result in incorrect timestamps in the output file.
  
  frame=  124 fps= 12 q=-1.0 Lsize=     507kB time=00:00:04.84 bitrate= 858.7kbits/s dup=26 drop=4 speed=0.453x

  So my question is, why does this video break ffmpeg ? Why does ffmpeg get caught up in a tight loop extracting images without ever making any progres towards the end of the video ?

• FFMPEG and libavfilter

  21 mars 2012, par Sergio

  With the currently installed ffmpeg I can't post watermark on converted videos. Can the reason for that be that I have not installed libavfilter library ?

  Currently # ffmpeg -v looks like :

  FFmpeg version SVN-r20374, Copyright (c) 2000-2009 Fabrice Bellard, et al.
  
  built on Oct 26 2009 22:47:01 with gcc 4.1.2 20080704 (Red Hat 4.1.2-46)
  
  configuration: --enable-libmp3lame --enable-libopencore-amrnb --enable-version3 -   enable-libopencore-amrwb --enable-version3 --enable-nonfree --enable-libfaad --enable-gpl - -disable-mmx --enable-shared --enable-libfaac --enable-libvorbis
  
  libavutil     50. 3. 0 / 50. 3. 0
  
  libavcodec    52.37. 1 / 52.37. 1
  
  libavformat   52.39. 2 / 52.39. 2
  
  libavdevice   52. 2. 0 / 52. 2. 0

  As you can see I'm missing libavfilter ? How can I add that library to FFMPEG and successfully add watermark on videos ?

  Thanks for any help.

• Bit-field badness

  30 janvier 2010, par Mans — Compilers, Optimisation

  Consider the following C code which is based on an real-world situation.

  struct bf1_31 
      unsigned a:1 ;
      unsigned b:31 ;
   ;
  void func(struct bf1_31 *p, int n, int a)
  
  
  
      int i = 0 ;
  
      do 
  
          if (p[i].a)
  
              p[i].b += a ;
  
       while (++i < n) ;
  
  
  

  How would we best write this in ARM assembler ? This is how I would do it :
  

  func :
          ldr     r3,  [r0], #4
          tst     r3,  #1
          add     r3,  r3,  r2,  lsl #1
          strne   r3,  [r0, #-4]
          subs    r1,  r1,  #1
          bgt     func
          bx      lr
  

  The add instruction is unconditional to avoid a dependency on the comparison. Unrolling the loop would mask the latency of the ldr instruction as well, but that is outside the scope of this experiment.

  Now compile this code with gcc -march=armv5te -O3 and watch in horror :

  func :
          push    r4
          mov     ip, #0
          mov     r4, r2
  loop :
          ldrb    r3, [r0]
          add     ip, ip, #1
          tst     r3, #1
          ldrne   r3, [r0]
          andne   r2, r3, #1
          addne   r3, r4, r3, lsr #1
          orrne   r2, r2, r3, lsl #1
          strne   r2, [r0]
          cmp     ip, r1
          add     r0, r0, #4
          blt     loop
          pop     r4
          bx      lr
  

  This is nothing short of awful :

  • The same value is loaded from memory twice.
  • A complicated mask/shift/or operation is used where a simple shifted add would suffice.
  • Write-back addressing is not used.
  • The loop control counts up and compares instead of counting down.
  • Useless mov in the prologue ; swapping the roles or r2 and r4 would avoid this.
  • Using lr in place of r4 would allow the return to be done with pop {pc}, saving one instruction (ignoring for the moment that no callee-saved registers are needed at all).

  Even for this trivial function the gcc-generated code is more than twice the optimal size and slower by approximately the same factor.

  The main issue I wanted to illustrate is the poor handling of bit-fields by gcc. When accessing bitfields from memory, gcc issues a separate load for each field even when they are contained in the same aligned memory word. Although each load after the first will most likely hit L1 cache, this is still bad for several reasons :

  • Loads have typically two or three cycles result latency compared to one cycle for data processing instructions. Any bit-field can be extracted from a register with two shifts, and on ARM the second of these can generally be achieved using a shifted second operand to a following instruction. The ARMv6T2 instruction set also adds the SBFX and UBFX instructions for extracting any signed or unsigned bit-field in one cycle.
  • Most CPUs have more data processing units than load/store units. It is thus more likely for an ALU instruction than a load/store to issue without delay on a superscalar processor.
  • Redundant memory accesses can trigger early flushing of store buffers rendering these less efficient.

  No gcc bashing is complete without a comparison with another compiler, so without further ado, here is the ARM RVCT output (armcc --cpu 5te -O3) :

  func :
          mov     r3, #0
          push    r4, lr
  loop :
          ldr     ip, [r0, r3, lsl #2]
          tst     ip, #1
          addne   ip, ip, r2, lsl #1
          strne   ip, [r0, r3, lsl #2]
          add     r3, r3, #1
          cmp     r3, r1
          blt     loop
          pop     r4, pc
  

  This is much better, the core loop using only one instruction more than my version. The loop control is counting up, but at least this register is reused as offset for the memory accesses. More remarkable is the push/pop of two registers that are never used. I had not expected to see this from RVCT.

  Even the best compilers are still no match for a human.