Screen Casting using ffmpeg (too fast)

9

4

I can use ffmpeg to make screen casts:

ffmpeg -f x11grab -s 1280x800 -i :0.0 -c:v libx264 -framerate 30 -r 30 -crf 18 out.mkv

However the output comes out to be too fast paced. It also happens with GTK RecordMyDesktop if I enable the encode on the fly. So, the questions is how to get a normal video pace. Also in order to capture the sound with ffmpeg what option should be used?

FFmpeg Output:

    ffmpeg -f x11grab -s 1280x800 -r 30 -i :0.0 -c:v libx264 -framerate 30 -r 30 -crf 18 out.mkv
ffmpeg version N-35162-g87244c8 Copyright (c) 2000-2012 the FFmpeg developers
  built on Oct  7 2012 15:56:19 with gcc 4.6 (Ubuntu/Linaro 4.6.3-1ubuntu5)
  configuration: --enable-gpl --enable-libfaac --enable-libfdk-aac --enable-libmp3lame --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-librtmp --enable-libtheora --enable-libvorbis --enable-libvpx --enable-x11grab --enable-libx264 --enable-nonfree --enable-version3
  libavutil      51. 73.102 / 51. 73.102
  libavcodec     54. 64.100 / 54. 64.100
  libavformat    54. 29.105 / 54. 29.105
  libavdevice    54.  3.100 / 54.  3.100
  libavfilter     3. 19.102 /  3. 19.102
  libswscale      2.  1.101 /  2.  1.101
  libswresample   0. 16.100 /  0. 16.100
  libpostproc    52.  1.100 / 52.  1.100
[x11grab @ 0xab896a0] device: :0.0 -> display: :0.0 x: 0 y: 0 width: 1280 height: 800
[x11grab @ 0xab896a0] shared memory extension found
[x11grab @ 0xab896a0] Estimating duration from bitrate, this may be inaccurate
Input #0, x11grab, from ':0.0':
  Duration: N/A, start: 1350136942.608988, bitrate: 983040 kb/s
    Stream #0:0: Video: rawvideo (BGR[0] / 0x524742), bgr0, 1280x800, 983040 kb/s, 30 tbr, 1000k tbn, 30 tbc
[libx264 @ 0xab87320] using cpu capabilities: MMX2 SSE2Fast SSSE3 Cache64 SlowCTZ SlowAtom
[libx264 @ 0xab87320] profile High 4:4:4 Predictive, level 3.2, 4:4:4 8-bit
[libx264 @ 0xab87320] 264 - core 128 r2 198a7ea - H.264/MPEG-4 AVC codec - Copyleft 2003-2012 - http://www.videolan.org/x264.html - options: 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=4 threads=6 lookahead_threads=1 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=18.0 qcomp=0.60 qpmin=0 qpmax=69 qpstep=4 ip_ratio=1.40 aq=1:1.00
Output #0, matroska, to 'out.mkv':
  Metadata:
    encoder         : Lavf54.29.105
    Stream #0:0: Video: h264, yuv444p, 1280x800, q=-1--1, 1k tbn, 30 tbc
Stream mapping:
  Stream #0:0 -> #0:0 (rawvideo -> libx264)
Press [q] to stop, [?] for help
frame=   10 fps=0.0 q=0.0 size=       1kB time=00:00:00.00 bitrate=   0.0kbits/sframe=   19 fps= 17 q=0.0 size=       1kB time=00:00:00.00 bitrate=   0.0kbits/sframe=   28 fps= 17 q=0.0 size=       1kB time=00:00:00.00 bitrate=   0.0kbits/sframe=   37 fps= 17 q=0.0 size=       1kB time=00:00:00.00 bitrate=   0.0kbits/sframe=   45 fps= 16 q=0.0 size=       1kB time=00:00:00.00 bitrate=   0.0kbits/sframe=   47 fps= 14 q=0.0 size=       1kB time=00:00:00.00 bitrate=   0.0kbits/sframe=   52 fps= 13 q=24.0 size=     257kB time=00:00:00.00 bitrate=2101632.0kbiframe=   55 fps= 12 q=24.0 size=     257kB time=00:00:00.10 bitrate=20808.2kbitsframe=   59 fps= 11 q=24.0 size=     289kB time=00:00:00.23 bitrate=10145.0kbitsframe=   64 fps= 11 q=24.0 size=     289kB time=00:00:00.40 bitrate=5894.7kbits/frame=   70 fps= 11 q=24.0 size=     289kB time=00:00:00.60 bitrate=3933.1kbits/frame=   72 fps= 10 q=24.0 size=     289kB time=00:00:00.66 bitrate=3549.2kbits/frame=   77 fps=9.8 q=24.0 size=     289kB time=00:00:00.83 bitrate=2837.7kbits/frame=   80 fps=9.6 q=24.0 size=     289kB time=00:00:00.93 bitrate=2533.5kbits/frame=   85 fps=9.3 q=24.0 size=     289kB time=00:00:01.10 bitrate=2146.9kbits/frame=   89 fps=9.3 q=24.0 size=     289kB time=00:00:01.23 bitrate=1917.1kbits/frame=   92 fps=9.1 q=24.0 size=     289kB time=00:00:01.33 bitrate=1773.3kbits/frame=   96 fps=9.0 q=24.0 size=     289kB time=00:00:01.46 bitrate=1612.4kbits/frame=   99 fps=8.8 q=24.0 size=     321kB time=00:00:01.56 bitrate=1676.8kbits/frame=  104 fps=8.7 q=24.0 size=     321kB time=00:00:01.73 bitrate=1515.2kbits/frame=  109 fps=5.3 q=24.0 Lsize=    1093kB time=00:00:03.56 bitrate=2511.5kbits/s    
video:1092kB audio:0kB subtitle:0 global headers:0kB muxing overhead 0.120198%
[libx264 @ 0xab87320] frame I:3     Avg QP:18.93  size:142610
[libx264 @ 0xab87320] frame P:43    Avg QP:20.79  size: 15751
[libx264 @ 0xab87320] frame B:63    Avg QP:23.75  size:   195
[libx264 @ 0xab87320] consecutive B-frames: 21.1%  1.8% 11.0% 66.1%
[libx264 @ 0xab87320] mb I  I16..4: 50.0% 21.1% 28.9%
[libx264 @ 0xab87320] mb P  I16..4:  6.1%  0.9%  3.2%  P16..4:  5.5%  1.2%  0.6%  0.0%  0.0%    skip:82.5%
[libx264 @ 0xab87320] mb B  I16..4:  0.4%  0.1%  0.0%  B16..8:  2.9%  0.1%  0.0%  direct: 0.0%  skip:96.5%  L0:40.7% L1:57.0% BI: 2.3%
[libx264 @ 0xab87320] 8x8 transform intra:14.5% inter:46.1%
[libx264 @ 0xab87320] coded y,u,v intra: 33.5% 24.1% 25.4% inter: 0.9% 0.4% 0.4%
[libx264 @ 0xab87320] i16 v,h,dc,p: 70% 26%  1%  3%
[libx264 @ 0xab87320] i8 v,h,dc,ddl,ddr,vr,hd,vl,hu: 11% 21% 30%  5%  7%  5%  7%  4% 10%
[libx264 @ 0xab87320] i4 v,h,dc,ddl,ddr,vr,hd,vl,hu: 32% 35% 12%  2%  4%  3%  4%  3%  5%
[libx264 @ 0xab87320] Weighted P-Frames: Y:0.0% UV:0.0%
[libx264 @ 0xab87320] ref P L0: 57.0%  5.6% 26.8% 10.6%
[libx264 @ 0xab87320] ref B L0: 69.4% 22.6%  8.0%
[libx264 @ 0xab87320] ref B L1: 93.7%  6.3%
[libx264 @ 0xab87320] kb/s:2460.40

rowman

Posted 2012-10-13T09:47:14.793

Reputation: 335

I believe the problem is the first -r 30 this means "ignore the incoming data timestamps and treat it as if it were exactly 30 fps" try "-framerate 30" up there instead – rogerdpack – 2015-10-12T21:22:22.003

-f alsa -i pulse should get you audio input. Can you give us the complete, uncut command line output as well? – slhck – 2012-10-13T09:53:52.260

1

I see x11grab has a -framerate option. It defaults to NTSC, so you could maybe use -framerate 30 and -r 30 for the output in combination?

– slhck – 2012-10-13T10:07:52.607

@slhck, thanks. the post is updated according to your suggestion, however the same problem. My computer is not that fast either. – rowman – 2012-10-13T14:07:30.820

@slhck, I think I have some clue what happens. It seems that it misses some shots in between while doing the encoding at the same time. That is why it seems faster. Specially when the load is higher the rate of frame loss is much higher and the video just jumps. Is there a method to just capture without encoding and encode the video when the capturing is done, as done by GTK RecordMyDesktop? – rowman – 2012-10-13T21:41:28.227

Answers

5

Try using a lossless encoder to capture the screen, and then re-encode the output when you are finished to create a smaller file if desired. The advantage of this method is often a less intensive capturing process which can result in a "faster" capture frame rate. Of course results can vary.

ffmpeg -f x11grab -s 1280x800 -i :0.0 -c:v libx264 -preset ultrafast -crf 0 out.mkv

It is possible that your CPU simply does not have the capability to encode at your declared frame rate at a particular screen capture size. In that case you can try a smaller -s value. It might be worth experimenting with other lossless encoders such huffyuv, ffv1, or utvideo but I personally haven't tried these for screencasts.

More info:

llogan

Posted 2012-10-13T09:47:14.793

Reputation: 31 929

It seems that the other lossless codecs that you mentioned are less resource intensive compared to x264. Will comment more accurately on that later on. – rowman – 2012-10-14T07:28:45.767

1@rowman Almost any encoder is less resource intensive than x264 (or generally, h.264 encoders), which is simply a matter of quality vs. time taken to encode. This is why many users still stick to XviD or similar when it comes to real-time encoding. – slhck – 2012-10-14T08:43:23.277

|@slhck Good point. Does container also have any effect on the resources? and Is there any literature on comparison of different lossless video codecs in terms of resources? Almost all of them claim to be the fastest one. – rowman – 2012-10-14T14:21:20.387

@rowman Did you try my example? Using x264 to create a lossless output should give somewhat similar performance compared to the other encoders I listed. Maybe a little slower; maybe a little faster, but I imagine the differences should not be considerable. – llogan – 2012-10-14T17:15:50.363

@LordNeckbeard, yes I did. x264 in your example has 60-70% overload on my cpu while huffyuv, utvideo, ffv1 have 25-35% in average. I have a solid intel atom ;) – rowman – 2012-10-14T18:11:53.267

By the way, can you tell me output bitrate of your tested video? (I know it's depended on video content and is not constant, but I'd like to know an example value) – Jet – 2014-04-29T16:38:38.803

Asked: 2012-10-13T09:47:14.793

Viewed: 6 451 times

Active: 2015-01-20T16:34:08.673