HOWTO: Install and use the latest FFmpeg and x264

[Gigalo]

FakeOutdoorsman, your assumptions are all correct. Although I changed from 32 bit to 64 bit Ubuntu when I switched, doing a complete OS wipe to achieve a consistent result. That said, I have now recloned both x264 and ffmpeg. After this, no problem except that it took two compiles to ffmpeg to complete. Wierd.

At any rate, does anyone have any idea where the '-vf' command option went?

... As of my previous compile (~2mo ago) this worked.

ffmpeg -i $inFile -pass 1 -an -sn -vf crop=8:0:704:480 \
-vcodec libx264 -vpre slow_firstpass -b 2048k -bt 2048k \
-threads 0 -y -flags2 +dct8x8+bpyramid -flags +loop \
-deblockalpha 2 -deblockbeta 5 -r 30000/1001 -y \
-f matroska /dev/null

FFmpeg version SVN-r19327
... bunch of output ...

[NULL @ 0x21274b0]Unknown option 'vf'
[swscaler @ 0x21284e0]Unknown option 'vf'
[NULL @ 0x2126010]Unknown option 'f'
ffmpeg: unrecognized option '-vf'

Also, any critiques regarding my settings? This is being used to transcode DVD's.

[dannyboy79]

FakeOutdoorsman, your assumptions are all correct. Although I changed from 32 bit to 64 bit Ubuntu when I switched, doing a complete OS wipe to achieve a consistent result. That said, I have now recloned both x264 and ffmpeg. After this, no problem except that it took two compiles to ffmpeg to complete. Wierd.

At any rate, does anyone have any idea where the '-vf' command option went?

... As of my previous compile (~2mo ago) this worked.

ffmpeg -i $inFile -pass 1 -an -sn -vf crop=8:0:704:480 \
-vcodec libx264 -vpre slow_firstpass -b 2048k -bt 2048k \
-threads 0 -y -flags2 +dct8x8+bpyramid -flags +loop \
-deblockalpha 2 -deblockbeta 5 -r 30000/1001 -y \
-f matroska /dev/null

FFmpeg version SVN-r19327
... bunch of output ...

[NULL @ 0x21274b0]Unknown option 'vf'
[swscaler @ 0x21284e0]Unknown option 'vf'
[NULL @ 0x2126010]Unknown option 'f'
ffmpeg: unrecognized option '-vf'

Also, any critiques regarding my settings? This is being used to transcode DVD's.

not sure, according to here it's still there: http://www.ffmpeg.org/ffmpeg-doc.html

[FakeOutdoorsman]

FakeOutdoorsman, your assumptions are all correct. Although I changed from 32 bit to 64 bit Ubuntu when I switched, doing a complete OS wipe to achieve a consistent result. That said, I have now recloned both x264 and ffmpeg. After this, no problem except that it took two compiles to ffmpeg to complete. Wierd.

At any rate, does anyone have any idea where the '-vf' command option went?

... As of my previous compile (~2mo ago) this worked.

ffmpeg -i $inFile -pass 1 -an -sn -vf crop=8:0:704:480 \
-vcodec libx264 -vpre slow_firstpass -b 2048k -bt 2048k \
-threads 0 -y -flags2 +dct8x8+bpyramid -flags +loop \
-deblockalpha 2 -deblockbeta 5 -r 30000/1001 -y \
-f matroska /dev/null

FFmpeg version SVN-r19327
... bunch of output ...

[NULL @ 0x21274b0]Unknown option 'vf'
[swscaler @ 0x21284e0]Unknown option 'vf'
[NULL @ 0x2126010]Unknown option 'f'
ffmpeg: unrecognized option '-vf'

Also, any critiques regarding my settings? This is being used to transcode DVD's.

You're using an old FFmpeg revision (r19327). As of now, FFmpeg is at r24460. Can you show the complete FFmpeg output?

I usually recommend single-pass CRF instead of two-pass VBR (as it appears what you're using) unless you are require a specific bitrate and/or output file size. I would remove:

-flags2 +dct8x8+bpyramid -flags +loop -deblockalpha 2 -deblockbeta 5

...because the presets already cover most of these options. FFmpeg will inherit the frame rate of the input so you don't need to declare -r if you want the frame rate to remain the same as the input.

[mocha]

Does anyone here know if ffmpeg has an equivalent option to mencoder's tele_src/tele_dest below:

 tele_src and tele_dest
              Enables arbitrary telecining  using  Donand  Graft's  DGPulldown
              code.   You need to specify the original and the desired framer‐
              ate; the muxer will make the video stream look like it  was  en‐
              coded at the desired framerate.  It only works with MPEG-2 video
              when the input framerate is smaller than  the  output  framerate
              and the framerate increase is <= 1.5.

              EXAMPLE:
                 tele_src=25,tele_dest=30000/1001
                      PAL to NTSC telecining

If not, what is the recommended method in ffmpeg to convert PAL to NTSC for MPEG2-DVD encoding? Thanks.

[qyot27]

Does anyone here know if ffmpeg has an equivalent option to mencoder's tele_src/tele_dest below:

 tele_src and tele_dest
              Enables arbitrary telecining  using  Donand  Graft's  DGPulldown
              code.   You need to specify the original and the desired framer‐
              ate; the muxer will make the video stream look like it  was  en‐
              coded at the desired framerate.  It only works with MPEG-2 video
              when the input framerate is smaller than  the  output  framerate
              and the framerate increase is <= 1.5.

              EXAMPLE:
                 tele_src=25,tele_dest=30000/1001
                      PAL to NTSC telecining

If not, what is the recommended method in ffmpeg to convert PAL to NTSC for MPEG2-DVD encoding? Thanks.

It's been about 2 years since I messed with ffmpeg for MPEG-2 encoding, as it had issues encoding DVD compliant streams that are free of noticeable problems (namely, I-frames being encoded at a much higher quantizer than P or B frames, which is really obvious on the first frame of a video; fixing or lessening it required breaking DVD compliancy). And I also seem to remember it didn't support *any* form of pulldown flagging, although this might have changed, or changed only to support standard 3:2 pulldown. I do know that ffmpeg does report weird status info if you give it a stream with pulldown, although I don't know the extent of how good or bad a sign this is. All of that is why I use HCenc instead.

What I do know is that DGPulldown itself, which mencoder's docs mentioned that it uses, runs perfectly in Wine (as does HCenc, for that matter). You'll just need to mux the video and audio together with mplex from the MJPEGTools project after fixing the MPEG-2 stream. I do believe that mplex is available natively, although it'll run okay in Wine as well.

[NobodyNowehreNothing]

Hey Guys,

Please if you can, help me out or just direct me to a right location, I have been stock for a whole week with this error, I have tried everything, Different versions of ubuntu (8.10,9.10 and now 10.04) , different versions of x264 , different versions of ffmpeg , everything i could, my client is now going mental!

This is now running on a Ubuntu Server 10.04 - I have gnome and everything installed, and I can VNC into the server which is setup on :1 , The server is at EC2 amazon, I did the same installation at home and I had no errors so I am going crazy! Whenever i try to use x11grab to capture screen, I get:

Segmentation fault

Here is the full version:

ffmpeg -v 2 -y -s 800x400 -r 9 -f x11grab -i :1.0   -g 300 -threads 0 -vcodec libtheora -sc_threshold -1   -cqp 22  -b 1200k  /tmp/screenCapture1.ogv
FFmpeg version SVN-r24502, Copyright (c) 2000-2010 the FFmpeg developers
  built on Jul 25 2010 20:36:29 with gcc 4.4.3
  configuration: '--enable-gpl' '--enable-version3' '--enable-nonfree' '--enable-postproc' '--enable-pthreads' '--enable-libfaac' '--enable-libmp3lame' '--enable-libopencore-amrnb' '--enable-libopencore-amrwb' '--enable-libtheora' '--enable-libvorbis' '--enable-libvpx' '--enable-libx264' '--enable-libxvid' '--enable-x11grab'
  libavutil     50.23. 0 / 50.23. 0
  libavcore      0. 0. 0 /  0. 0. 0
  libavcodec    52.84. 0 / 52.84. 0
  libavformat   52.77. 0 / 52.77. 0
  libavdevice   52. 2. 0 / 52. 2. 0
  libavfilter    1.26. 1 /  1.26. 1
  libswscale     0.11. 0 /  0.11. 0
  libpostproc   51. 2. 0 / 51. 2. 0
[x11grab @ 0x9872470] device: :1.0 -> display: :1.0 x: 0 y: 0 width: 800 height: 400
[x11grab @ 0x9872470] shared memory extension  found
Segmentation fault

Please guys, if you have any clue, help me out or point me to the right direction! Long time ubuntu user here

[FakeOutdoorsman]

-threads 0 is currently only for libx264 as far as I know. Try removing that option or use a value more suitable for your CPU.

Do you get a seg fault with a very basic command? Example:

ffmpeg -t 5 -f x11grab -r 30 -s 640x320 -i :0.0 -y x11grab.mpg

If the basic command also provides a seg fault, you should file a bug report: Reporting a Bug To The FFmpeg Project. It is important to use a minimal command in your bug report to rule out extras such as threads or external libraries.

If the basic command does work, then I suggest using libx264 to capture your screen as shown in this useful guide:
HOWTO: Proper Screencasting on Linux

[mocha]

It's been about 2 years since I messed with ffmpeg for MPEG-2 encoding, as it had issues encoding DVD compliant streams that are free of noticeable problems (namely, I-frames being encoded at a much higher quantizer than P or B frames, which is really obvious on the first frame of a video; fixing or lessening it required breaking DVD compliancy). And I also seem to remember it didn't support *any* form of pulldown flagging, although this might have changed, or changed only to support standard 3:2 pulldown. I do know that ffmpeg does report weird status info if you give it a stream with pulldown, although I don't know the extent of how good or bad a sign this is. All of that is why I use HCenc instead.

What I do know is that DGPulldown itself, which mencoder's docs mentioned that it uses, runs perfectly in Wine (as does HCenc, for that matter). You'll just need to mux the video and audio together with mplex from the MJPEGTools project after fixing the MPEG-2 stream. I do believe that mplex is available natively, although it'll run okay in Wine as well.

Thanks for the info. It's interesting considering that not much development happens on mencoder but it is superior for some MPEG2/DVD stuff compared to ffmpeg.

[ron999]

Hi
I don't seem to be able to convert audio into ogg using vorbis codec any more.
I think that previously I used a command like:

ffmpeg -i < input > -acodec libvorbis output.ogg

When I try that now it doesn't recognize libvorbis.

Unknown encoder 'libvorbis'

So I tried vorbis instead

ffmpeg -i < input > -acodec vorbis output.ogg

When I tried that it tells me to use '-strict experimental'.

encoder 'vorbis' is experimental and might produce bad results.
Add '-strict experimental' if you want to use it.

So I tried

ffmpeg -i < input > -acodec vorbis -strict experimental output.ogg

It converted OK, but when I played it back there is no sound output.

General
Complete name : output.ogg
Format : OGG
File size : 447 KiB
Duration : 3mn 18s
Overall bit rate : 18.5 Kbps

Audio
ID : 530006952 (0x1F9743A8)
Format : Vorbis
Duration : 3mn 18s
Channel(s) : 2 channels
Sampling rate : 44.1 KHz

There's information on the web that says the internal ogg encoder is no good and recommends that libvorbis is used instead.

Is this problem now caused because ffmpeg hasn't been compiled with libvorbis?
And is there some other way to use ffmpeg to convert using vorbis?
Or am I just doing something wrong?

FFmpeg version SVN-r24614, Copyright (c) 2000-2010 the FFmpeg developers
built on Jul 31 2010 15:19:46 with gcc 4.4.1
configuration: --enable-gpl --enable-version3 --enable-nonfree --enable-postproc --enable-pthreads --enable-libfaac --enable-libmp3lame --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libtheora --enable-libvpx --enable-libx264 --enable-libxvid --enable-x11grab
libavutil 50.23. 0 / 50.23. 0
libavcore 0. 2. 0 / 0. 2. 0
libavcodec 52.84. 2 / 52.84. 2
libavformat 52.77. 0 / 52.77. 0
libavdevice 52. 2. 0 / 52. 2. 0
libavfilter 1.27. 0 / 1.27. 0
libswscale 0.11. 0 / 0.11. 0
libpostproc 51. 2. 0 / 51. 2. 0

EDIT

I think I've answered my own questions now.
There's information in this blog here:-http://xiphmont.livejournal.com/51160.html

So I've re-compiled ffmpeg and added these two extras to the ./configure line
--disable-encoder=vorbis --enable-libvorbis


Now when I use the command:-

ffmpeg -i < input > -acodec vorbis output.ogg

It reports

Unknown encoder 'vorbis'

And when I use command:-

ffmpeg -i < input > -acodec libvorbis output.ogg

It converts OK and the result sounds alright.

General
Complete name : output.ogg
Format : OGG
File size : 1.33 MiB
Duration : 3mn 18s
Overall bit rate : 56.1 Kbps
Writing application : Lavc52.84.2

Audio
ID : 1396686445 (0x533FBE6D)
Format : Vorbis
Format settings, Floor : 1
Duration : 3mn 18s
Bit rate : 64.0 Kbps
Channel(s) : 2 channels
Sampling rate : 44.1 KHz
Stream size : 1.51 MiB
Writing library : libVorbis 1.2 (UTC 2007-06-22)

ron@ubuntu:~$ ffmpeg
FFmpeg version SVN-r24643, Copyright (c) 2000-2010 the FFmpeg developers
built on Aug 1 2010 10:38:14 with gcc 4.4.1
configuration: --enable-gpl --enable-version3 --enable-nonfree --enable-postproc --enable-pthreads --enable-libfaac --enable-libmp3lame --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libtheora --enable-libvpx --enable-libx264 --enable-libxvid --enable-x11grab --disable-encoder=vorbis --enable-libvorbis
libavutil 50.23. 0 / 50.23. 0
libavcore 0. 2. 0 / 0. 2. 0
libavcodec 52.84. 2 / 52.84. 2
libavformat 52.78. 0 / 52.78. 0
libavdevice 52. 2. 0 / 52. 2. 0
libavfilter 1.27. 0 / 1.27. 0
libswscale 0.11. 0 / 0.11. 0
libpostproc 51. 2. 0 / 51. 2. 0

[Mustache Villain]

Any word on when we will see ffvp8 codec, the alternative to Google's libvpx?