Newest 'x264' Questions - Stack Overflow

Les articles publiés sur le site

• Does H.264 encoding profile affect encoding performance ?

  3 novembre 2017, par user8154565

  I want to encode in real-time using as little CPU and memory\disk I\O as possible, keeping a low CRF (currently 14). Will changing the encoding profile from Baseline to something else affect encoding CPU usage?

  My current settings are:

  Profile\Level: Baseline@L5.2

  Resolution and frame rate: 1920x1080 @ ~144.001 FPS (variable frame rate mode)

  Average overall data rate: 174 mbit\s (audio is ~320kbit\s VBR AAC)

  x264 settings: cabac=0 / ref=1 / deblock=0:0:0 / analyse=0:0 / me=dia / subme=0 / psy=1 / psy_rd=1.00:0.00 / mixed_ref=0 / me_range=16 / chroma_me=1 / trellis=0 / 8x8dct=0 / cqm=0 / deadzone=21,11 / fast_pskip=1 / chroma_qp_offset=0 / threads=24 / lookahead_threads=4 / sliced_threads=0 / nr=0 / decimate=1 / interlaced=0 / bluray_compat=0 / constrained_intra=0 / bframes=0 / weightp=0 / keyint=250 / keyint_min=25 / scenecut=0 / intra_refresh=0 / rc=crf / mbtree=0 / crf=14.0 / qcomp=0.60 / qpmin=0 / qpmax=69 / qpstep=4 / ip_ratio=1.40 / aq=0

• How to modify x264 source code to get motion search time [closed]

  21 octobre 2017, par Yicheng Li

  I am a new member of stack overflow, I hope you guys can help me solve this problem. I am now using x264 with ffmpeg to encode some raw video, I was asked how much time the motion search would cost in libx264 video coding. Therefore, I need to modify the code, I tried put the clock around the switch sentence in the encoder/me.c clock start

  clock end After compiling the source code and run a test, it shows that the time is very small and it seems I am actually getting the time each macro block spent on motion search.

  So, how should I modify the x264 source code to get the time duration in motion search.

  Big Thanks.

• x264 segmentation error :in ff_rgb24ToY_avx ()

  20 octobre 2017, par helloliu

  On my Ubunutu 16.04,I compiled the ffmpeg 、nasm and x264 code as follows:

  cd ffmeg/
  ./configure &&make &&make install 
  git clone http://git.videolan.org/git/x264.git
  cd x264/
  ./configure &&make &&make install 
  

  And then, I run in the follow way :

  ./x264 --input-res=1080x1920 --input-csp=rgb --output ~/x264/raw_framews.264  ~/x264/raw_framews
  

  Unfortunately,it reports with the errors as follows:

    lavf [error]: could not open input file
    avs [error]: failed to load avisynth
    raw [info]: 1080x1920p 0:0 @ 25/1 fps (cfr)
    resize [warning]: converting from rgb24 to yuv420p
    x264 [info]: using cpu capabilities: MMX2 SSE2Fast SSSE3 SSE4.2 AVX FMA3 BMI2 AVX2
    x264 [info]: profile High, level 4.0
    [swscaler @ 0x3f31540] Warning: data is not aligned! This can lead to a speedloss
    Thread 1 "x264encode" received signal SIGSEGV, Segmentation fault.
    0x0000000000534777 in ff_rgb24ToY_avx ()
  

  Can anyone help us with this error? many thanks.

• X264 input frame persistence

  18 octobre 2017, par martin_

  Issue: X264 input frame persistence/ validity - duration.

  Description: I set up an encoder using X264, and created some test cases. 1. Simple linear encode 2. Parallel simultaneous encodes 3. Threaded Encoding 4. Threaded time delayed start encoding.

  All worked well, and I was happy with the results. Then I threw one more test at the encoder. I started nullifying (memset(0)) input frames after they had been passed to the encoder, which is what I would expect to happen in the final application; the results were that 'blank' (green - I420) frames started getting passed into the output stream.

  Hence my question - how many old frames do I have to keep in memory for the encoder to function properly? Is this a configurable parameter? Or moreover which parameter controls this behaviour. I was using 'high' as the profile. Baseline didn't exhibit these issues.

  Thanks

  Martin

• What's the relationship between AVPacket in FFMpeg and H.264 NAL Uint ?

  10 octobre 2017, par Ello

  I use FFMpeg's api to encode images to H.264 stream, the code is like this encode code. When I called the function avcodec_encode_video2 I got one AVPacket's data, then I saved it to file. I use several images to get several files for test. Based on the startcode 00 00 00 01 or 00 00 01, I found that the 1st and 2nd files' data both contain several H.264 NAL Uints. But start from the third file, I can't found the startcode. Like followed image.

  I feel confused. Base on the FFMpeg's code I think one AVPacket's data should contain at least one NAL Uint, but the result is unexpected. Is that means one NAL Uint may be divided to two AVPackets' data? What's the relationship between AVPacket and H.264 NAL Uint?