[Ffmpeg-devel-irc] ffmpeg.log.20150406

burek burek021 at gmail.com
Tue Apr 7 02:05:01 CEST 2015


[00:00:54 CEST] <c_14> yes
[00:02:25 CEST] <DragonsLordKERNE> I will try something like this:
[00:02:27 CEST] <DragonsLordKERNE> ffmpeg -i input.mp4 -codec:v libxvid -vtag DIVX -vf fps=fps=25 -b:v 1500k -g 250 -bf 0 -me_method full -mbd 2 -cmp chroma -trellis 1
[00:02:32 CEST] <DragonsLordKERNE> -codec:a libmp3lame -b:a 128k -ar 48000 -ac 2 -joint_stereo 0 -af 'volume=4' output.avi
[00:02:56 CEST] <DragonsLordKERNE> :|
[00:04:02 CEST] <DragonsLordKERNE> I forgot to apply LibAVDeinterlacer (Linear Blending) - Crop - Resize (Bilinear)
[00:04:16 CEST] <DragonsLordKERNE> Do you have enough patience to support me, c_14?
[00:04:37 CEST] <DragonsLordKERNE> I will pay a bear and pizza to you when you will come to Italy next time
[00:04:38 CEST] <DragonsLordKERNE> ;)
[00:07:08 CEST] <c_14> -vf pp=lb,crop=w=[]:h=[]:x=[]:y=[],scale=w=[]:h=[]
[00:07:35 CEST] <c_14> At least, that's how I would translate those terms into ffmpeg filters
[00:07:44 CEST] <c_14> everywhere there's a [] replace that with the appropriate value
[00:08:01 CEST] <c_14> You can find documentation to the individual filters here: https://ffmpeg.org/ffmpeg-filters.html
[00:08:46 CEST] <c_14> Also, not sure what the default scale algo is, you can set -sws_flags bilinear to use bilinear scaling if you want that
[00:10:27 CEST] <DragonsLordKERNE> I didn't understand how can i crop in terms of
[00:10:29 CEST] <DragonsLordKERNE> Up
[00:10:31 CEST] <DragonsLordKERNE> Down
[00:10:36 CEST] <DragonsLordKERNE> Left & right
[00:10:37 CEST] <DragonsLordKERNE> :(
[00:10:42 CEST] <c_14> Right
[00:10:55 CEST] <c_14> w and h are self-explanatory (I hope)
[00:11:29 CEST] <c_14> x is how many pixels to cut off on the left side
[00:11:30 CEST] <DragonsLordKERNE> Weight = Up
[00:11:42 CEST] <c_14> y is how many pixels to cut off on the top
[00:11:54 CEST] <c_14> weiht?
[00:11:57 CEST] <c_14> w = width, h = height
[00:12:11 CEST] <c_14> wxh is the size of the area that you want to keep
[00:12:50 CEST] <DragonsLordKERNE> OH my god
[00:13:02 CEST] <DragonsLordKERNE> so I cannot simply tell FFMPEG to cut
[00:13:10 CEST] <DragonsLordKERNE> 2 pixels from the left
[00:13:14 CEST] <DragonsLordKERNE> 2 from the right
[00:13:29 CEST] <DragonsLordKERNE> 4 from the top and so on
[00:13:42 CEST] <c_14> crop=w=iw-2:h=ih-4
[00:13:49 CEST] <c_14> wait
[00:13:55 CEST] <c_14> iw-4 and ih-8
[00:14:02 CEST] <c_14> x and y default to centered
[00:14:58 CEST] <DragonsLordKERNE> the issue is that they are not centered
[00:14:59 CEST] <DragonsLordKERNE> :(
[00:15:13 CEST] <c_14> What isn't?
[00:15:16 CEST] <c_14> The cuts?
[00:15:21 CEST] <DragonsLordKERNE> YEAH!
[00:15:37 CEST] <DragonsLordKERNE> I have a frame of 1920 x 1080
[00:15:50 CEST] <DragonsLordKERNE> i need to cut only 4 from the top
[00:15:56 CEST] <DragonsLordKERNE> 2 from the bottom
[00:16:06 CEST] <DragonsLordKERNE> 2 from the left and 4 from the right
[00:16:06 CEST] <c_14> A "simple" formula is w=iw-[total width to cut]:h=ih-[total height to cut]:x=[pixels to cut from the right]:y=[pixels to cut from the top]
[00:16:06 CEST] <DragonsLordKERNE> :|
[00:16:57 CEST] <c_14> So in that case w=iw-6:h=ih-6:x=2:y=4
[00:17:04 CEST] <c_14> (When I said right earlier, I meant left)
[00:17:39 CEST] <DragonsLordKERNE> AH! OK!
[00:17:55 CEST] <DragonsLordKERNE> That's why my confusion
[00:17:56 CEST] <DragonsLordKERNE> :)
[00:18:10 CEST] <c_14> Directions are hard.
[00:18:58 CEST] <c_14> Should I go with cut from port?
[00:19:30 CEST] <DragonsLordKERNE> ?
[00:19:45 CEST] <c_14> (Nautical term for the left side (when facing forward) of a ship)
[00:19:54 CEST] <DragonsLordKERNE> crop=w=0:h=4:x=0:y=4 it to cut 4 pixels from the top right?
[00:20:11 CEST] <c_14> (forwards to the bow)
[00:20:15 CEST] <DragonsLordKERNE> (I see ... more or less)
[00:20:22 CEST] <c_14> eh, close
[00:20:33 CEST] <c_14> that'll make a video of size 0x4
[00:20:37 CEST] <c_14> Which would be 0 pixels
[00:20:44 CEST] <c_14> You probabl want w=iw-0:h=ih-4
[00:20:48 CEST] <c_14> *probably
[00:20:53 CEST] <DragonsLordKERNE> yes, of course
[00:22:26 CEST] <DragonsLordKERNE> crop=w=1920:h=1076:x=0:y=4
[00:22:35 CEST] <DragonsLordKERNE> Did I get it? :)
[00:23:01 CEST] <DragonsLordKERNE> 4 from the top
[00:23:33 CEST] <c_14> probably, you can actually literally use "ih-0" etc though. crop supports expression logic for the arguments and ih is a variable that is replaced with the input height
[00:25:48 CEST] <DragonsLordKERNE> so, something like this?
[00:25:50 CEST] <DragonsLordKERNE> -vf pp=lb,crop=w=iw-0:h=ih-4:x=0:y=4,scale=w=[704]:h=[400] -sws_flags bilinear
[00:26:16 CEST] <DragonsLordKERNE> (I'm praying, teacher ...)
[00:26:31 CEST] <c_14> get rid of the [] in the scale thingies
[00:26:51 CEST] <c_14> And if you want to preserve aspect ration you can use -2 for either w or h
[00:28:15 CEST] <DragonsLordKERNE> I need to set the Pixel Aspect Ratio to 1:1
[00:28:40 CEST] <DragonsLordKERNE> is this option?
[00:28:42 CEST] <DragonsLordKERNE> setsar=1:1
[00:28:45 CEST] <c_14> append or prepend (not sure which is better) setsar=1
[00:28:47 CEST] <c_14> yes
[00:28:49 CEST] <c_14> that
[00:28:54 CEST] <DragonsLordKERNE> :)
[00:29:37 CEST] <DragonsLordKERNE> -vf pp=lb,crop=w=iw-0:h=ih-4:x=0:y=4,scale=w=704:h=400,setsar=1:1 -sws_flags bilinear ?
[00:29:51 CEST] <c_14> should work
[00:29:56 CEST] <DragonsLordKERNE> :))
[00:31:34 CEST] <DragonsLordKERNE> beginand end
[00:31:38 CEST] <DragonsLordKERNE> begin *and end
[00:31:49 CEST] <DragonsLordKERNE> are not explicitly set right?
[00:31:57 CEST] <c_14> of what?
[00:31:59 CEST] <DragonsLordKERNE> I need to set start and duration, right?
[00:32:05 CEST] <c_14> only if you want to cut it
[00:32:14 CEST] <DragonsLordKERNE> If I need to do a frame accurate encoding
[00:32:25 CEST] <DragonsLordKERNE> Yes, I would like to do that
[00:32:27 CEST] <c_14> If you want to cut, use -ss and -t
[00:32:37 CEST] <c_14> -t is duration, -ss is from where to start
[00:34:03 CEST] <DragonsLordKERNE> is it also possible to select something like a frame number?
[00:34:13 CEST] <c_14> yes, with the select filter
[00:34:39 CEST] <DragonsLordKERNE> But how can I detect the frame number?
[00:34:45 CEST] <DragonsLordKERNE> With VLC or some other tools?
[00:35:05 CEST] <c_14> something like select=between(n\,100\,200)
[00:35:11 CEST] <c_14> Will take all frames between 100 and 200
[00:35:21 CEST] <c_14> there's also gt (greater than) lt (less than)
[00:35:36 CEST] <DragonsLordKERNE> Yes, but I need to see this frames
[00:35:36 CEST] <c_14> https://ffmpeg.org/ffmpeg-utils.html#Expression-Evaluation
[00:35:44 CEST] <c_14> mhm
[00:35:58 CEST] <c_14> Just get the exact timestamp.
[00:35:59 CEST] <DragonsLordKERNE> in advance I mean
[00:36:03 CEST] <c_14> ffmpeg cuts are frame perfect
[00:36:23 CEST] <DragonsLordKERNE> timestamp from which software?
[00:36:32 CEST] <DragonsLordKERNE> this is what I would like to understand
[00:36:35 CEST] <c_14> mediaplayer or whatever
[00:40:13 CEST] <DragonsLordKERNE> can I also use ms?
[00:40:17 CEST] <DragonsLordKERNE> something like
[00:40:17 CEST] <DragonsLordKERNE> 00:00:16.120
[00:40:20 CEST] <c_14> yep
[00:40:35 CEST] <DragonsLordKERNE> it's -ms?
[00:40:46 CEST] <c_14> https://ffmpeg.org/ffmpeg-utils.html#Time-duration
[00:40:51 CEST] <c_14> just like you wrote
[00:42:47 CEST] <DragonsLordKERNE> Doc c_14 ... it's running
[00:42:54 CEST] <DragonsLordKERNE> Are you praying with me?
[00:42:55 CEST] <DragonsLordKERNE> XD
[01:09:54 CEST] <faLUCE> c_14: tried ffmpeg -loop 1 -t 3:02 -i 1.jpg -loop 1 -i 2.jpg -i mp3 -shortest -c:a copy out.avi , it doesn't work... only 2.jpg is added
[01:14:02 CEST] <c_14> hmm, lemme test for a bit
[01:15:59 CEST] <DragonsLordKERNE> c_14, -me_method full should be the same for -me 4 ?
[01:17:28 CEST] <c_14> hmm?
[01:17:48 CEST] <DragonsLordKERNE> I found this:
[01:17:49 CEST] <DragonsLordKERNE> http://forum.videohelp.com/threads/277807-Useful-FFmpeg-Syntax-Examples
[01:21:36 CEST] <DragonsLordKERNE> -me does not exist (no more I guess)
[01:23:06 CEST] <c_14> DragonsLordKERNE: it was probably replaced
[01:23:37 CEST] <c_14> faLUCE: I might just be an idiot
[01:23:47 CEST] <c_14> -filter_complex '[0][1]concat=n=2:v=1:a=0[v]' -map '[v]' -map 2
[01:23:49 CEST] <c_14> add that
[01:23:54 CEST] <DragonsLordKERNE> YEAH, maybe
[01:24:23 CEST] <DragonsLordKERNE> <c_14> faLUCE: I might just be an idiot <<< From my perspective ... I don't think so :)
[01:25:02 CEST] <c_14> faLUCE: after the mp3 before the output file
[01:25:22 CEST] <faLUCE> c_14: wow, this is a real complex command ;-) thanks
[01:25:59 CEST] <DragonsLordKERNE> c_14, Do you think these > http://pastebin.com/Z5HJY5Hq < are the best option in order to have a beatiful High-Quality XviD ...
[01:26:07 CEST] <faLUCE> c_14: where is indicated the time in that command?
[01:26:11 CEST] <DragonsLordKERNE> c_14, ... or did I miss something?
[01:26:16 CEST] <faLUCE> 3 minutes and 2 seconds...
[01:26:35 CEST] <c_14> faLUCE: just put that in the middle of the command you already have
[01:26:53 CEST] <faLUCE> c_14: thanks, let's try
[01:29:03 CEST] <c_14> DragonsLordKERNE: I haven't actually encoded XviD since a whiiile now. Nothing jumps out to me as wrong though.
[01:29:15 CEST] <DragonsLordKERNE> c_14, should I also set something like ildctcmp precmp subcmp and so on?
[01:30:19 CEST] <c_14> probably not necessary
[01:32:17 CEST] <faLUCE> c_14: it wants two pictures of the same size
[01:32:39 CEST] <faLUCE> c_14: anyway, you are a monster of ffmpeg :-) really  a wizard
[01:32:55 CEST] <c_14> use the scale filter on one (or both) of them
[01:33:46 CEST] <c_14> ie -filter_complex '[0]scale=180x537[tmp];[tmp][1]...'
[01:34:04 CEST] <DragonsLordKERNE> c_14, -cmp can also be numbers right?
[01:34:47 CEST] <c_14> DragonsLordKERNE: I think so, yes. Most options like that have both numbers and names
[01:35:37 CEST] <faLUCE> c_14: -filter_complex '[0][1]concat=n=2:v=1:a=0[v]' -map '[v]'  <----- How can I insert the scale filter in that expression?
[01:36:50 CEST] <c_14> Like I wrote just up there ^, '[number of input you want to scale]scale=widthxheight[tmp];[either [tmp] or 0][either [tmp] or 1]concat...
[01:36:58 CEST] <c_14> just replace the front part of the filter_complex expression
[01:38:09 CEST] <faLUCE> c_14: thanks
[01:39:18 CEST] <DragonsLordKERNE> c_14, would you be so kind to check if me_method 6 is the best?
[01:39:39 CEST] <DragonsLordKERNE> I was not able to find it out
[01:41:17 CEST] <c_14> For ffmpeg me_method 6 would be one higher than the best one. There's 5 options and it's 0-indexed
[01:41:37 CEST] <c_14> Eh
[01:41:42 CEST] <c_14> There's 6 options and it's 0-indexed.
[01:42:46 CEST] <DragonsLordKERNE> the same applies for mbd = 2 and cmp = chroma (I was not able to find the correct number)
[01:45:21 CEST] <DragonsLordKERNE> FFMPEG seems to accept also me_method = 8 :|
[01:46:11 CEST] <c_14> It'll probably accept up to INT_MAX
[01:46:16 CEST] <c_14> Not sure what it does with it though&
[01:47:21 CEST] <DragonsLordKERNE> and in order to set min quantizers and max quantizer for I-P-B-Frames
[01:47:28 CEST] <DragonsLordKERNE> is it -qmin ... -qmax ...
[01:47:35 CEST] <c_14> yes
[01:48:03 CEST] <DragonsLordKERNE> so which me_method is the best in terms of quality?
[01:48:12 CEST] <DragonsLordKERNE> This is exactly what I am not understanding
[01:48:57 CEST] <c_14> full
[01:49:04 CEST] <c_14> Well quality / filesize
[01:51:30 CEST] <DragonsLordKERNE> full is not mentioned here:
[01:51:31 CEST] <DragonsLordKERNE> https://www.ffmpeg.org/ffmpeg-codecs.html
[01:51:38 CEST] <DragonsLordKERNE> @me_method (me)
[01:52:22 CEST] <c_14> https://www.ffmpeg.org/ffmpeg-codecs.html#libxvid
[01:52:24 CEST] <c_14> It's down there.
[01:56:58 CEST] <DragonsLordKERNE> [Parsed_setsar_3 @ 0000000005780be0] num:den syntax is deprecated, please use nu
[01:56:58 CEST] <DragonsLordKERNE> m/den or named options instead <<< Do you know what does it mean?
[01:57:20 CEST] <DragonsLordKERNE> *[Parsed_setsar_3 @ 0000000005780be0] num:den syntax is deprecated, please use num/den or named options instead <<< Do you know what does it mean?
[01:59:02 CEST] <c_14> ye, you can ignore it
[01:59:11 CEST] <c_14> It just means that instead of setsar=1:1 you should use setsar=1/1
[02:00:34 CEST] <DragonsLordKERNE> Thank you very much for all your time, spent for me
[02:00:49 CEST] <DragonsLordKERNE> The conversion is running
[02:12:54 CEST] <asdf_> hi. how can i convert video to mp3?
[02:13:01 CEST] <asdf_> which i need to download?
[02:13:21 CEST] <c_14> >convert video to mp3
[02:13:25 CEST] <c_14> Please elaborate
[02:13:46 CEST] <c_14> Also, what do you mean "which i need to download"
[02:13:47 CEST] <asdf_> ONE OF this
[02:13:48 CEST] <asdf_> http://www.archive.org/details/IntroductoryIntelX86ArchitectureAssemblyApplicationsDay1Part1
[02:13:54 CEST] <asdf_> extension
[02:14:02 CEST] <c_14> You want to extract the audio from it?
[02:14:04 CEST] <asdf_> which ffmpeg i need to dnlowad
[02:14:05 CEST] <asdf_> yes
[02:14:25 CEST] <asdf_> so  ican do multiple instead in vlc
[02:14:47 CEST] <c_14> ffmpeg -i https://archive.realllonglink.mp4 out.mp3
[02:15:05 CEST] <c_14> Read https://trac.ffmpeg.org/wiki/Encode/MP3 for quality settings
[02:15:20 CEST] <c_14> https://ffmpeg.org/download.html
[02:16:31 CEST] <DragonsLordKERNE> c_14, in Xmedia Recode I have a setting that is "Audio / Video syncronization" ... Do you know if I should set something similar also with FFMPEG in order to preserve the absolute A/V sync?
[02:16:39 CEST] <asdf_> i want good quality sound
[02:16:49 CEST] <c_14> DragonsLordKERNE: most of the time you won't need to
[02:17:03 CEST] <DragonsLordKERNE> asdf_, good depends on the source
[02:17:10 CEST] <asdf_> mp3?
[02:17:11 CEST] <DragonsLordKERNE> asdf_, good depends on you
[02:17:26 CEST] <DragonsLordKERNE> good can be 320 kbps for me and 128 kbps for you
[02:17:43 CEST] <c_14> something around qscale 3/4 is a good start and you can adjust up or down to your liking
[02:17:48 CEST] <c_14> (note, lower is better)
[02:17:52 CEST] <DragonsLordKERNE> c_14, but it's better to set it or to leave it unset?
[02:17:53 CEST] <c_14> All covered in that linke I posted
[02:18:02 CEST] <c_14> DragonsLordKERNE: leave it unless you have problems
[02:18:21 CEST] <DragonsLordKERNE> Can you please explain what it does?
[02:19:47 CEST] <c_14> It probably touches the -vsync option, which basically specifies what to do with frames so that you get the right framerate/ the right frames on the right timestamp
[02:21:02 CEST] <DragonsLordKERNE> Looking at your explanation ... it would sounds better to have it always set
[02:21:18 CEST] <asdf_> nvm i guess i wil use vlc
[02:21:37 CEST] <c_14> DragonsLordKERNE: ffmpeg usually sets it to the correct one by itself
[02:22:00 CEST] <c_14> Look at the manpage entry for -vsync. https://ffmpeg.org/ffmpeg.html
[03:10:27 CEST] <electronrotoscop> ffmpeg is crashing after the line [Parsed_drawtext_4 @ 0000000005852e20] Setting 'fontcolor' to value 'white' and I think I missed a library in my build
[03:10:36 CEST] <electronrotoscop> the config was configuration: --arch=x86_64 --target-os=mingw32 --cross-prefix=x86_64-w64-mingw32- --enable-cross-compile --enable-runtime-cpudetect --pkg-config=pkg-config --enable-memalign-hack --enable-nonfree --enable-gpl --enable-version3 --disable-w32threads --enable-libfdk_aac --enable-libfribidi --enable-avisynth --enable-bzlib --enable-fontconfig --enable-iconv --enable-libfreetype --enable-libx264 --enable-zlib
[03:10:42 CEST] <electronrotoscop> anyone know what I might be missing?
[03:11:03 CEST] <electronrotoscop> full log at http://pastebin.com/keNQpXMR
[03:12:24 CEST] <electronrotoscop> the error windows gives me is http://pastebin.com/UW3AQjPe
[03:14:16 CEST] <WickedP> hi folks, does ffmpeg have a namespace when working in c++?
[03:16:52 CEST] <GWARslave> hey, quick question. I have a simple batch fie written, or want to write, where I can drag an .flv file onto it and have it output an .mp4 without re-encoding. Is there a basic command that'll keep the same output name as the input
[03:17:26 CEST] <electronrotoscop> windows?
[03:17:31 CEST] <GWARslave> yeah, windows
[03:17:48 CEST] <GWARslave> I saw somewhere that might be "${1}" for input, and 2 for output
[03:17:52 CEST] <GWARslave> but that didn't work
[03:18:34 CEST] <GWARslave> @ECHO OFF
[03:18:34 CEST] <GWARslave> ffmpeg -i "${1}".flv -c copy -copyts "${2}".mp4
[03:18:34 CEST] <GWARslave> Is all I have
[03:19:04 CEST] <electronrotoscop> “c:\path\to\ffmpeg.exe -i "%~f1" ^
[03:19:04 CEST] <electronrotoscop> -map 0:v -vcodec copy ^
[03:19:05 CEST] <electronrotoscop> -map 0:a -acodec copy ^
[03:19:05 CEST] <electronrotoscop> -f mp4 "c:\path\to\dest\%~n1-remuxed.mp4"
[03:19:05 CEST] <electronrotoscop> PAUSE
[03:19:24 CEST] <GWARslave> well then, wherever I read was a bit off
[03:19:31 CEST] <electronrotoscop> whoops that should be a " at the beginning not “
[03:19:42 CEST] <electronrotoscop> ${1} is for OSX/Linux
[03:19:47 CEST] <GWARslave> aha
[03:20:17 CEST] <electronrotoscop> probably you should take out the -map 0:v and -map 0:a, that's for multiple audio tracks
[03:20:35 CEST] <electronrotoscop> if you're going to mp4 you probably only want one video track and one (stereo or whatever) audio track
[03:20:51 CEST] <GWARslave> yeah this is just 1 video, 1 audio
[03:21:07 CEST] <electronrotoscop> so then just "c:\path\to\ffmpeg.exe -i "%~f1" -vcodec copy -acodec copy -f mp4 "c:\path\to\dest\%~n1-remuxed.mp4"
[03:21:16 CEST] <electronrotoscop> and another line with PAUSE
[03:21:27 CEST] <electronrotoscop> save as a .bat, drag file into .bat and you're off to the races
[03:21:50 CEST] <GWARslave> excellent, tanx
[03:21:53 CEST] <GWARslave> lemme give it a spin
[03:22:23 CEST] <GWARslave> wait
[03:22:35 CEST] <GWARslave> I already have ffmpeg set up for windows to recognize it wherever
[03:22:52 CEST] <GWARslave> so I leave out the directory line and just put ffmpeg right
[03:23:50 CEST] <electronrotoscop> oh yeah, sure that'll work just say ffmpeg then I guess
[03:26:11 CEST] <GWARslave> excellent, it works
[03:26:25 CEST] <GWARslave> PAUSE didn't do anything. no biggie though, I prolly missed something
[03:26:29 CEST] <electronrotoscop> if you want to go back to the same directory you dropped from, no matter where it us, use ffmpeg -i "%~f1" -vcodec copy -acodec copy -f mp4 "%~d1%~p1%~n1-remuxed.mp4"
[03:26:50 CEST] <electronrotoscop> pause just makes it hold the window up until you press any key, as long as you put it by itself on the last line
[03:27:07 CEST] <electronrotoscop> I find it useful to check to make sure ffmpeg didn't find any issues while remuxing
[03:27:14 CEST] <GWARslave> gotcha
[03:27:32 CEST] <GWARslave> d1 p1 huh
[03:27:48 CEST] <GWARslave> luckily all the work's being done in the same folder
[03:27:53 CEST] <GWARslave> but ill keep that in min
[03:27:55 CEST] <GWARslave> d
[03:28:30 CEST] <electronrotoscop> %~d1 is drive of source file %~p1 is path of source file %~n1 is filename without extension
[03:28:40 CEST] <GWARslave> gotcha
[03:28:46 CEST] <GWARslave> was tryna figure out what p was
[03:29:03 CEST] <electronrotoscop> ha ha ha
[03:30:04 CEST] <GWARslave> @ECHO OFF
[03:30:05 CEST] <GWARslave> ffmpeg -i "%~f1" -vcodec copy -acodec copy "e:\%~n1.mp4"
[03:30:05 CEST] <GWARslave> PAUSE
[03:30:11 CEST] <GWARslave> what am i missing to make the pause work
[03:30:16 CEST] <GWARslave> i feel like some quotations somewhere
[03:31:58 CEST] <electronrotoscop> not sure. maybe it doesn't work with echo off
[03:32:10 CEST] <electronrotoscop> I guess if it's not writing anything on screen anyways you don't need pause
[03:32:31 CEST] <GWARslave> hm ic
[03:32:40 CEST] <GWARslave> been quite a while since i've written simple batch files
[03:32:54 CEST] <electronrotoscop> ha ha it takes some getting used to
[03:33:01 CEST] <GWARslave> i feel like typing in LOAD ,8,1 hah
[03:33:12 CEST] <electronrotoscop> ha ha ha ha
[03:33:18 CEST] <electronrotoscop> @echo off disables all terminal output, iirc
[03:33:48 CEST] <electronrotoscop> so I'd take that out if you want to keep an eye on ffmpeg doing its thing
[03:34:30 CEST] <GWARslave> well, i took out @echo off and it works, but still doesn't pause
[03:34:42 CEST] <GWARslave> oh well
[03:35:09 CEST] <electronrotoscop> ha ha not completely essential. so it just closes by itself?
[03:35:54 CEST] <GWARslave> yeah
[03:36:05 CEST] <electronrotoscop> huh. no idea.
[03:40:18 CEST] <Hello71> @ means do not print this command like in make
[03:40:24 CEST] <Hello71> echo off means don't print future commands
[03:41:00 CEST] Action: Hello71 uses psychic debugging
[03:41:13 CEST] <Hello71> %~f1 isn't a valid expansion
[03:45:14 CEST] <GWARslave> hm
[03:59:37 CEST] <GWARslave> hmm
[03:59:58 CEST] <GWARslave> thought they were all flv codecs, apparently not
[03:59:58 CEST] <electronrotoscop> what does valid expansion mean?
[04:00:16 CEST] <electronrotoscop> getting some error messages?
[04:00:35 CEST] <GWARslave> yeah after a few of them
[04:00:52 CEST] <GWARslave> ran mediainfo, some of the .flv's are AVC video
[04:01:07 CEST] <electronrotoscop> avc should rewrap into mp4
[04:01:38 CEST] <electronrotoscop> try removing the @echo off line and put a bunch of PAUSE lines in and see if you can get it to keep the window up so you can see what it's doing
[04:02:32 CEST] <GWARslave> i'm bypaasing the batch file right now
[04:02:36 CEST] <GWARslave> and manually typing
[04:04:37 CEST] <GWARslave> http://cplusplus.bordoon.com/cmd_exe_variables.html
[04:04:45 CEST] <GWARslave> maybe what he's talking about
[04:04:48 CEST] <GWARslave> he/she
[04:05:37 CEST] <GWARslave> http://www.picpaste.com/pics/Untitled-acXky623.1428285908.png
[04:05:40 CEST] <GWARslave> is what i'm getting
[04:06:16 CEST] <GWARslave> is it trying to fit stream 0 in stream 1's spot?
[04:07:12 CEST] <electronrotoscop> so that codec isn't avc it's flv1
[04:07:23 CEST] <electronrotoscop> which I'm not super familiar with, but it's saying that flv1 isn't supported in mp4
[04:07:35 CEST] <electronrotoscop> you could give mkv a shot, but i suspect you'll just have to transcode
[04:08:24 CEST] <GWARslave> oh sorry, i meant the other way around
[04:08:28 CEST] <WickedP> hi folks, i'm having a really frustrating time trying to get ffmpeg working in my c++ project. I'm using zeranoe's dev package and have linked all the libraries etc. I enclosed the headers in extern "C". But whenever I call an ffmpeg function (e.g. av_register_all()) the program goes into an access violation error. What do I need to be doing to rectify this?
[04:08:34 CEST] <GWARslave> the avc ones were converting fine
[04:08:59 CEST] <electronrotoscop> yeah that makes sense, avc aka h264 is very common for mp4. flv1 is I think pretty old
[04:09:50 CEST] <GWARslave> Sorenson Spark
[04:09:52 CEST] <GWARslave> what a name
[04:10:09 CEST] <GWARslave> that seems to be your basic flash video codec nowadays
[04:10:54 CEST] <electronrotoscop> avc is the basic one yeah, sorenson spark was like what the first versions of youtube used
[04:11:23 CEST] <electronrotoscop> something like this might work for you
[04:11:24 CEST] <electronrotoscop> ffmpeg -i "%~f1" -pix_fmt yuv420p -vcodec libx264 -preset veryslow -tune fastdecode -tune film -crf 18 -refs 1 -qmin 4 -maxrate 1000k -bufsize 500k -acodec copy -f mp4 "%~dpn1-converted.mp4"
[04:11:31 CEST] <GWARslave> http://en.wikipedia.org/wiki/Flash_Video#Media_type_support
[04:11:35 CEST] <electronrotoscop> but if you are converting to avc then you have to make all kinds of new decisions
[04:11:48 CEST] <GWARslave> jesus son
[04:11:56 CEST] <GWARslave> i'm putting so much time and effort into porn files
[04:11:56 CEST] <GWARslave> lol
[04:12:00 CEST] <electronrotoscop> bahahahaha
[04:12:36 CEST] <electronrotoscop> sorenson spark aka flv1 seems to be the codec, where flv is the container
[04:13:28 CEST] <GWARslave> right
[04:13:48 CEST] <GWARslave> hmm, your new command is taking much longer
[04:16:23 CEST] <electronrotoscop> bahaha well it's transcoding instead of just copying frames from one file into another so
[04:16:43 CEST] <GWARslave> oh
[04:16:51 CEST] <electronrotoscop> http://torcache.net/torrent/B2905B674A8F3A1096934F13E10273761C91D573.torrent?title=[kickass.to]marvels.agents.of.s.h.i.e.l.d.s02e14.720p.hdtv.x264.immerse.glodls
[04:16:52 CEST] <GWARslave> transcoding=re-encoding?
[04:16:57 CEST] <electronrotoscop> wait ha ha wrong window
[04:16:59 CEST] <electronrotoscop> yeah
[04:17:12 CEST] <GWARslave> was tryna get away from re-encoding
[04:17:21 CEST] <GWARslave> otherwise i could just stick it in vlc player or something
[04:17:40 CEST] <electronrotoscop> try the -vcodec copy script but with other containers
[04:17:47 CEST] <electronrotoscop> maybe mkv or something can take flv1
[04:17:47 CEST] <GWARslave> got no time for anything but to learn heh
[04:18:04 CEST] <GWARslave> i'm tryna get the files into Camtasia
[04:18:11 CEST] <GWARslave> it dont read .mkv
[04:18:15 CEST] <GWARslave> nor those .flv files
[04:18:24 CEST] <electronrotoscop> mov?
[04:18:30 CEST] <GWARslave> perhaps
[04:18:35 CEST] <electronrotoscop> it might not read flv1 no matter what container it's in
[04:18:40 CEST] <GWARslave> never really have or use .mov
[04:20:22 CEST] <GWARslave> tru..
[04:21:21 CEST] <GWARslave> whew, i remembered ctrl-break
[04:21:57 CEST] <GWARslave> i was testing that on an avc file which converted fine earlier
[04:23:43 CEST] <electronrotoscop> ha ha ha ctrl-c works often too
[04:24:11 CEST] <GWARslave> ah
[04:24:31 CEST] <GWARslave> yea ur right, i used vlc to do a raw dump
[04:24:40 CEST] <GWARslave> and it still doesn't read it
[04:29:37 CEST] <electronrotoscop> fucking spark
[04:30:11 CEST] <electronrotoscop> "Macromedia later tried to find a better video codec. Starting with Flash Player 8 (released in September 2005), the preferred video codec became VP6."
[04:30:46 CEST] <electronrotoscop> it's now H.264 though. Nothing competes with H.264 on quality and compatability, except arguably vp9 or hevc
[04:31:36 CEST] <GWARslave> i was hanving a helluva time with hevc yesterday, or the day before
[04:31:57 CEST] <Hello71> electronrotoscop: H.264 just came out then
[04:31:59 CEST] <GWARslave> none of the .mkv files I had would extract properly, using several programs
[04:32:08 CEST] <GWARslave> but the files themselves worked fine
[04:32:15 CEST] <Hello71> unlike now it wasn't embedded in every smartphone
[04:32:20 CEST] <Hello71> for that matter smartphones barely existed
[04:32:54 CEST] <electronrotoscop> Hello71: ha ha yes and there have been one or two improvements to the encoders since then
[04:33:37 CEST] <electronrotoscop> with a careful eye, one should be able to see some differences in 264 since 2005
[04:33:41 CEST] <electronrotoscop> *x264
[04:34:11 CEST] <electronrotoscop> GWARslave: HEVC isn't that widely supported in editing programs yet
[04:34:31 CEST] <electronrotoscop> or, like, anywhere
[04:35:57 CEST] <GWARslave> yea i realized that, i went ahead and redownloaded all the files in 1080p instead
[04:40:42 CEST] <GWARslave> whats the recommended string for encoding the h264
[04:40:49 CEST] <GWARslave> like quality level and stuff
[04:42:01 CEST] <Hello71> https://encrypted.google.com/search?q=ffmpeg%20whats%20the%20recommended%20string%20for%20encoding%20the%20h264
[04:42:16 CEST] <GWARslave> smarty pants
[04:42:30 CEST] <electronrotoscop> ha ha you can use the string I gave you
[04:42:38 CEST] <Hello71> https://encrypted.google.com/search?q=smarty%20pants
[04:42:41 CEST] <electronrotoscop> depends on your source size
[04:43:02 CEST] <Hello71> SmartyPants Vitamins  All in One Gummy Vitamins for Kids and Adults with Omega 3 Fish Oil
[04:43:08 CEST] <electronrotoscop> hello71: you should use http://lmgtfy.com/
[04:44:07 CEST] <electronrotoscop> also: if you actually know of a good general recommendation I'd love to see it. That good search isn't giving much of value for anyone who doesn't want to learn all the basic optoins for x264 encoding
[04:44:27 CEST] <electronrotoscop> it's not an easily google-able question
[04:45:22 CEST] <electronrotoscop> GWARslave: if you're just playing back in vlc or something, you can just set -preset and -crf, the rest won't matter for you
[04:45:40 CEST] <electronrotoscop> and if you're going to be transcoding to something else, one option is to set -crf to 0 and then it's loessless
[04:45:44 CEST] <Hello71> third result on an un-customized page
[04:45:44 CEST] <electronrotoscop> *lossless
[04:46:12 CEST] <electronrotoscop> you're talking https://trac.ffmpeg.org/wiki/Encode/H.264 ?
[04:50:20 CEST] <GWARslave> sry, had a friend over, and working on a string
[04:51:14 CEST] <GWARslave> @ECHO OFF
[04:51:14 CEST] <GWARslave> ffmpeg -i %~n1.flv -vcodec libx264 -preset slow -crf 22 -acodec copy  %~n2.mp4
[04:51:14 CEST] <GWARslave> PAUSE
[04:51:17 CEST] <GWARslave> thats what i came up with
[04:51:56 CEST] <GWARslave> nah, i'm not worried about vlc. i'm just trying to get it to load into Camtasia
[04:53:31 CEST] <GWARslave> whoa
[04:53:33 CEST] <GWARslave> that was way off
[04:53:36 CEST] <GWARslave> out of sync
[04:54:11 CEST] <GWARslave> I just basically used the string that https://trac.ffmpeg.org/wiki/Encode/H.264 said
[04:54:21 CEST] <electronrotoscop> you'll want to probably use "%~1f" instead of %~n1.flv and %~dpn1.mp4 instead of %~n2.mp4
[04:54:50 CEST] <electronrotoscop> sorry that's "%~f1" not "%~1f"
[04:55:07 CEST] <GWARslave> ah, yeah
[04:55:15 CEST] <GWARslave> full path name as opposed to file name
[04:55:20 CEST] <electronrotoscop> crf 22 is pretty low quality, you could probably go higher
[04:55:21 CEST] <GWARslave> it still caem out correctly though
[04:55:26 CEST] <GWARslave> except for the syncage
[04:55:29 CEST] <GWARslave> and skippage
[04:55:40 CEST] <electronrotoscop> ha ha yes and the 1 is important, otherwise it'll try to name it blank I think?
[04:55:57 CEST] <GWARslave> ya, i was tryna go off that link I posted earlier ;P
[04:56:23 CEST] <GWARslave> i think it'll try to overwrite itself or something, i dunno
[04:56:50 CEST] <GWARslave> ima try your string you posted earlier
[04:56:51 CEST] <electronrotoscop> dunno why the sync issue
[04:57:29 CEST] <GWARslave> well, i assume transcoding the video, and leaving the audio as is would cause problems
[04:57:54 CEST] <GWARslave> wait
[04:58:01 CEST] <GWARslave> that WAS your string I used heh
[04:58:10 CEST] <GWARslave> mine is sitting in a txt file on my desktop
[04:58:11 CEST] <GWARslave> lol
[04:58:15 CEST] <electronrotoscop> ha ha
[04:58:24 CEST] <GWARslave> the drugs and sleeplessness are getting to me ;P
[04:59:49 CEST] <electronrotoscop> try ffmpeg -i "%~f1" -vcodec libx264 -preset slow -crf 18 -c:a pcm_s16le  "%~dpn1.mp4"
[04:59:53 CEST] <electronrotoscop> sorry wait
[05:00:01 CEST] <electronrotoscop> ffmpeg -i "%~f1" -vcodec libx264 -preset slow -crf 18 -c:a pcm_s16le  "%~dpn1.mov"
[05:00:41 CEST] <GWARslave> hold up
[05:00:47 CEST] <GWARslave> i'm watching the original file for the first time
[05:00:54 CEST] <GWARslave> the beginning was made out of sync on purpose
[05:00:55 CEST] <GWARslave> lol
[05:01:38 CEST] <GWARslave> only thing wrong was, the output file came out with no file name, just an extension ;P
[05:02:24 CEST] <electronrotoscop> yeah that'd be because you used 2
[05:02:41 CEST] <electronrotoscop> which refers to the name of the second file you drag onto the batch script (in this case nothing at all)
[05:03:09 CEST] <GWARslave> ho yeah
[05:03:11 CEST] <GWARslave> duh
[05:03:38 CEST] <electronrotoscop> takes some getting used to
[05:03:55 CEST] <GWARslave> yep
[05:07:43 CEST] <GWARslave> i dislike coding in the first palce
[05:08:12 CEST] <GWARslave> but since i lost my job last month and been depressed, I randomly started learning audio and video editing
[05:12:06 CEST] <electronrotoscop> it's a fun world
[05:12:22 CEST] <electronrotoscop> avid media composer has a 30 day trial option
[05:13:00 CEST] <electronrotoscop> http://en.wikipedia.org/wiki/Blender_%28software%29 is an NLE and 3d modelling program
[05:14:07 CEST] <GWARslave> how does avid compare to camtasia
[05:14:27 CEST] <electronrotoscop> I've never heard of camtasia, but actual like hollywood movies are cut on avid
[05:14:30 CEST] <GWARslave> hrm, 3d modelling, i think i did that back in high school
[05:14:50 CEST] <GWARslave> ah
[05:14:50 CEST] <GWARslave> well
[05:15:01 CEST] <GWARslave> i think camtasia's a mid-level software
[05:15:09 CEST] <GWARslave> not sure if i wanna jump into the big one yet ;P
[05:15:22 CEST] <electronrotoscop> the three things I see people cut on are Avid, Adobe Premiere and Final Cut Pro on OSX
[05:15:33 CEST] <GWARslave> altho some of the little details are annoying to me, like the animated ratio scale minus/gain
[05:15:45 CEST] <GWARslave> have you seen American Psycho?>
[05:15:56 CEST] <electronrotoscop> ha ha I actually worked on it
[05:15:58 CEST] <electronrotoscop> http://en.wikipedia.org/wiki/List_of_video_editing_software
[05:16:13 CEST] <electronrotoscop> in a very minor capacity
[05:16:16 CEST] <GWARslave> hm?
[05:16:17 CEST] <GWARslave> how so
[05:17:07 CEST] <electronrotoscop> i helped with the dailies when they were shooting
[05:17:21 CEST] <GWARslave> nice ;P. not sure what that is but still
[05:17:26 CEST] <GWARslave> what a coincidence that i asked that lol
[05:17:42 CEST] <electronrotoscop> ha ha yeah
[05:17:54 CEST] <electronrotoscop> what about it?
[05:18:14 CEST] <GWARslave> i'm working on a music video for it. Anthrax wrote a song about the book back in the 90s, so I figured I'd play around with it
[05:18:30 CEST] <electronrotoscop> sounds cool
[05:18:32 CEST] <GWARslave> been distracted the past couple weeks by other things, but i can show u hat i got. it's unlisted on youtube
[05:18:51 CEST] <electronrotoscop> ah I'm good, I have to go to bed actually
[05:18:55 CEST] <GWARslave> okie dokie
[05:19:56 CEST] <GWARslave> sleep well
[05:20:04 CEST] <GWARslave> i'll prolly pass out when i get drunk enuff
[05:45:43 CEST] <Prelude2004c> hey guys... to enable the cuda version of nvenc ?
[05:45:57 CEST] <Prelude2004c> --enable-libnvenc for that .. which was it for the cuda?
[05:49:29 CEST] <Prelude2004c> libxvid?
[05:57:14 CEST] <kepstin-laptop> i'm not sure what you mean, nvenc and cuda are completely different things.
[05:58:25 CEST] <kepstin-laptop> nvenc uses fixed-function encoding hardware on the gpu die, while cuda is a way for a software encoder to use the gpu cores.
[18:27:00 CEST] <loki_> can somebody explain me why x11grab is legacy if it is fast and light version screen grabbing?
[18:27:30 CEST] <loki_> now xcb version is used by default and when it active is see high cpu utilization
[18:45:27 CEST] <durandal_1707> Is it really that slow?
[18:47:40 CEST] <loki_> 1.5
[18:48:19 CEST] <loki_> i build ffmpeg by myself and for very long time can't get why system one works faster.
[18:48:43 CEST] <loki_> after i changed to real x11grab device i have even more better perfomance.
[18:49:44 CEST] <loki_> and bad part of that, that it is not just perf gain, there less cpu utilization for xorg too...
[19:02:45 CEST] <c_14> loki_: on my system the difference is pretty negligible
[19:02:51 CEST] <c_14> Especially once you add an encoder.
[19:03:02 CEST] <c_14> What commandline are you using to test?
[19:41:25 CEST] <seasc> loki_, might you have a build script for your ffmpeg? (assuming you included gnutls, v4l-utils and all the other stuff too?)
[19:42:43 CEST] <seasc> or c_14? i'm probably just stuck before the final line due to 'exhaustion' of writing my own 600+ lines script to build ffmpeg
[19:45:14 CEST] <seasc> build-ffmpeg.sh (using TUI) http://pastebin.com/rutufqsQ
[19:50:32 CEST] <c_14> seasc: what's the problem?
[19:52:16 CEST] <seasc> c_14, after gotting stuck with v4l, i decided to try build ffmpeg with the things i 'have', now it fails at nettle and gnutls.
[19:52:59 CEST] <c_14> config.log ?
[19:53:37 CEST] <c_14> wait
[19:53:46 CEST] <c_14> Why are you cross-compiling for your own platform?
[19:54:38 CEST] <seasc> thats been suggested once i've added --host=linux after a build failed and gave that as suggestion
[19:55:36 CEST] <seasc> dont recall from which package
[19:56:17 CEST] <c_14> you can probably just delete that entire line (line 676) it doesn't set anything helpful for most builds,
[20:01:02 CEST] <seasc> ok i'll try that, but dont i have to pull in nettle and gnutls for it too?
[20:01:36 CEST] <llogan> seasc: why --enable-ftrapv?
[20:01:54 CEST] <seasc> llogan, sounded good? :)
[20:03:20 CEST] <c_14> seasc: If I don't see the failing config.log I can't tell why nettle/gnutls isn't being detected
[20:04:26 CEST] <seasc> screenlog: http://pastebin.com/dwBq8G4f
[20:05:39 CEST] <seasc> config.log: http://pastebin.com/WBM1thFN
[20:06:09 CEST] <seasc> llogan, whats bad with --enable-ftrapv?
[20:07:51 CEST] <seasc> c_14, did you see the config.log?
[20:12:32 CEST] <c_14> seasc: it looks like you're actually having issues building libnettle, not ffmpeg.
[20:13:20 CEST] Action: c_14 has no idea how to build nettle correctly. You might want to read the official docs, look if they have a contact method for help somewhere
[20:14:09 CEST] <llogan> seasc: i've only seen it used for testing
[20:15:20 CEST] <seasc> c_14, yes, sorry for the confusion, thats why i asked if you guys had a script where it pulls th reqs for ffmpeg in as well.
[20:16:02 CEST] <seasc> llogan, would it hurt (influence performance in a bad way)?
[20:16:40 CEST] <seasc> so i could gather/adapt the missing/faulty parts
[20:21:45 CEST] <llogan> seasc: im not sure.
[20:21:55 CEST] <c_14> seasc: I have a small script that I use to build ffmpeg and a couple of deps, but I have neither nettle nor gnutls
[20:27:50 CEST] <seasc> llogan, well, i'll disable it for now, no intention for massive memory testings or alike :p thank you
[20:28:53 CEST] <seasc> i'll try a build ignore the error message and just use (--enable-XY) those built successfull. ... lets see in ~10mins
[21:17:17 CEST] <seasc> c_14, build failed, the extralibs are an attempt to include/use installed nettle/gnutls :: http://pastebin.com/tWrcSEmg
[21:36:18 CEST] <c_14> seasc: get rid of the --extra-libs
[21:40:26 CEST] <seasc> c_14, its currently building.. but why does it fail on 'exta' folders to look for libs/headers?
[21:41:19 CEST] <c_14> Because that isn't for folders.
[21:41:57 CEST] <seasc> oh, so i'd need to pass all extra libs/header each with an extra --extra-XY argument?
[21:42:17 CEST] <seasc> IF used, seems one is better of without it :)
[21:42:24 CEST] <c_14> yes, or just use extra-ldflags="-L/path -L/path2" etc
[21:43:51 CEST] <DragonsLord> Hello to everybody
[21:43:58 CEST] <DragonsLord> Hi c_14 :)
[21:44:01 CEST] <seasc> c_14, thank you
[21:44:06 CEST] <seasc> hi DragonsLord
[21:44:17 CEST] <DragonsLord> hi seasc
[21:55:23 CEST] <DragonsLord> Does someone know if it's possible to use "libavcodec deinterlacer" as a Video filter?
[21:56:13 CEST] <c_14> ffmpeg has at least 4 deinterlacers that I know of
[21:56:57 CEST] <DragonsLord> I understood that I need to 1) De-interlace 2) Crop 3) Resize exactly in this order
[21:57:29 CEST] <DragonsLord> I tried to Google for "libavcodec deinterlacer" + "FFMpeg" without finding any relevant result
[21:58:15 CEST] <RobotsOnDrugs> deinterlacers are filters and therefore would be part of libavfilter
[21:58:55 CEST] <RobotsOnDrugs> also http://ffmpeg.org/ffmpeg-all.html
[21:59:37 CEST] <RobotsOnDrugs> i prefer yadif
[22:00:53 CEST] <c_14> http://ffmpeg.org/ffmpeg-filters.html for a succincter version only with filters
[22:01:36 CEST] <DragonsLord> RobotsOnDrugs, I have found this page > https://www.ffmpeg.org/libavfilter.html <
[22:01:47 CEST] <DragonsLord> But I cannot understand what I should do
[22:01:56 CEST] <c_14> DragonsLord: are you using the libraries or the binary?
[22:02:13 CEST] <DragonsLord> c_14, good question :|
[22:02:24 CEST] <DragonsLord> The Binary on Windows
[22:02:39 CEST] <c_14> Then that page isn't what you want.
[22:03:03 CEST] <RobotsOnDrugs> DragonsLord: the libav* libraries are part of ffmpeg, and see the page i linked (or the one c_14 linked)
[22:05:55 CEST] <DragonsLord> RobotsOnDrugs, Probably I am stupid but I was unable to find the instruction in order to call the libavfilter
[22:05:56 CEST] <llogan> DragonsLord: a simple example is: -vf yadif,crop=iw-100,scale=1280:-2
[22:06:29 CEST] <DragonsLord> Thank you llogan but I am not interested in yadif
[22:06:52 CEST] <RobotsOnDrugs> it was an example
[22:07:10 CEST] <RobotsOnDrugs> -vf is used for all video filters
[22:07:43 CEST] <llogan> you mentioned you were using the binary, so i assumed you were using the binary.
[22:08:15 CEST] <DragonsLord> RobotsOnDrugs, I understood that but -vf (what?)
[22:08:27 CEST] <DragonsLord> -vf libavdeinterlacer ?
[22:08:41 CEST] <llogan> there is no such thing as libavdeinterlacer
[22:09:30 CEST] <llogan> or ""libavcodec deinterlacer"
[22:10:28 CEST] <DragonsLord> I asked this because I am trying to replace AVIdemux/Xmedia Recode with FFMpeg
[22:10:39 CEST] <DragonsLord> But I would like to use the same setting
[22:11:02 CEST] <DragonsLord> and since Xmedia Recode is a graphical interface for FFMpeg
[22:11:23 CEST] <DragonsLord> and since Xmedia Recode has libav as an interlacer
[22:11:29 CEST] <DragonsLord> I thought it has
[22:12:04 CEST] <llogan> i don't know what Xmedia Recode is. if you tell us what you want to do we can give example commands.
[22:12:34 CEST] <llogan> or you could attempt to find the exact ffmpeg commands Xmedia uses, but they will likely be "wrong"
[22:12:50 CEST] <DragonsLord> Well, I would like to:
[22:13:13 CEST] <RobotsOnDrugs> it was probably some arbitrary deinterlacer that is available in lavf
[22:13:24 CEST] <DragonsLord> 1) Apply libav as an interlacer filter (Linear blending mode)
[22:13:40 CEST] <RobotsOnDrugs> best to just pick a deinterlacer and go with it
[22:13:48 CEST] <llogan> i'm assuming it uses the binary, and just adds the old "-deinterlace" option.
[22:13:52 CEST] <DragonsLord> 2) Perform some crop (I got it yesterday with c_14)
[22:14:08 CEST] <llogan> the -deinterlace option now just maps to yadif, IIRC.
[22:14:28 CEST] <DragonsLord> 3) Perform a resize with "Scaling mode" = Bilinear
[22:19:13 CEST] <llogan> my example does most of that
[22:21:33 CEST] <DragonsLord> Does YADIF perform this operation, as far as you know?
[22:21:37 CEST] <DragonsLord> - Linear blend deinterlacing filter that deinterlaces the given block by filtering all lines with a (1 2 1) filter -
[22:21:53 CEST] <llogan> read the docs http://ffmpeg.org/ffmpeg-filters.html#yadif
[22:24:58 CEST] <DragonsLord> Sorry, I didn't understand :(
[22:25:17 CEST] <c_14> DragonsLord: lb/linblenddeint: Linear blend deinterlacing filter that deinterlaces the given block by filtering all lines with a (1 2 1) filter.
[22:29:20 CEST] <DragonsLord> so -vf pp=lb should be fine?
[22:29:42 CEST] <c_14> should
[22:30:31 CEST] <DragonsLord> and -sws_flags bilinear is for scaling, right?
[22:31:29 CEST] <llogan> probably better to use flags option when you use scale filter
[22:32:52 CEST] <DragonsLord> llogan, I did something like this > http://pastebin.com/tSu0EksY <
[22:33:05 CEST] <DragonsLord> Would you be so kind to review the option I set?
[22:33:25 CEST] <llogan> perform all filtering in one filtergraph
[22:34:10 CEST] <llogan> scale=w=704:h=400:flags=bilinear
[22:34:17 CEST] <llogan> then remove -sws_flags
[22:34:29 CEST] <llogan> bicubic is the default, right? i forget
[22:35:40 CEST] <DragonsLord> In AVIdemux Bilinear is the Default
[22:35:52 CEST] <DragonsLord> I do not really know how it is set in FFMpeg
[22:37:31 CEST] <DragonsLord> llogan, I reviewed the resulting file
[22:37:37 CEST] <llogan> yes, appears to be bicubic. at least for my lazy cli test.
[22:37:51 CEST] <DragonsLord> and I found a not perfect A/V sync
[22:38:01 CEST] <DragonsLord> Do you have any idea on how I can fix it?
[22:42:05 CEST] <seasc> I'm confused... libx264+libx265 are built fine (with EC:0), but i have to remove "--enable-libx264 --enable-libx265"  from the configure line or ffmpeg or it will NOT build.. however it has libx264 enabled when building without --enable-libx264...(at least i have binaries for that one) --> http://pastebin.com/09ZjvS7z
[22:44:01 CEST] <seasc> outout of builit ffmpeg binary: http://pastebin.com/VJ9pjzb2
[22:44:40 CEST] <DragonsLord> :)
[22:44:42 CEST] <c_14> seasc: >WARNING: library configuration mismatch
[22:44:49 CEST] <DragonsLord> OK fflogger is a Bot
[22:45:01 CEST] <c_14> the binary is loading information from system libraries
[22:45:17 CEST] <c_14> *ffmpeg libraries installed through a different method
[22:45:51 CEST] <seasc> c_14, yes, must do so or i'd need to pull in (and compile) nettle, gnutls and alot of others
[22:45:59 CEST] <DragonsLord> llogan, how I can save all the console output on Windows?
[22:46:37 CEST] <seasc> c_14, so i would need to remove the libav* from my system before i build ffmpeg?
[22:46:38 CEST] <c_14> seasc: export LD_LIBRARY_PATH=$PREFIX/lib (before testing the binary)
[22:47:19 CEST] <seasc> PREFIX beeing here.. as example my ~/ffmpeg_chroot ?
[22:47:22 CEST] <c_14> also, either set PKG_CONFIG_PATH so that ffmpeg can find the x264/x265.pc files or so that the include path is in the cflags
[22:47:25 CEST] <c_14> yes
[22:48:46 CEST] <llogan> DragonsLord: i don't know anything about Windows
[22:49:12 CEST] <seasc> uh oh.... PKG_CONFIG_PATH is to the include dirs? i had set it to "$CHROOT"/usr/lib, was that wrong?
[22:50:21 CEST] <seasc> DragonsLord, not sure, sould work like on linux... ffmpeg.exe args opts  ... > mylog.txt
[22:50:21 CEST] <c_14> seasc: it's set to include/pkgconfig
[22:50:56 CEST] <seasc> c_14, 	PKG_CONFIG_PATH="$DIR_LIB/pkgconfig" ?
[22:51:25 CEST] <DragonsLord> Thank you, llogan & seasc
[22:51:52 CEST] <c_14> yes, though you might want PKG_CONFIG_PATH="$DIR/LIB/pkgconfig:$PKG_CONFIG_PATH" in case there's already something in there you might want
[22:52:28 CEST] <seasc> c_14, good hint, thanks alot!
[23:01:02 CEST] <ssss>  I've a question on stack overflow regarding ffmpeg, Can Someone please help me with it ? http://stackoverflow.com/questions/29479100/ffmpeg-how-to-set-histogram-size
[23:01:02 CEST] <DragonsLord> Hello glebihan and ssss
[23:01:20 CEST] <ssss> Thank you.
[23:02:43 CEST] <c_14> ssss: put a scale in front of the split ?
[23:05:37 CEST] <ssss> @c_14 I'm fairly new to ffmpeg can you please point on the right direction in order to put a scale in front of the split ?
[23:08:07 CEST] <c_14> [a]format=gray,histogram=mode=waveform:waveform_mode=column,vflip,scale=fooxbar,split[c]
[23:09:45 CEST] <ssss> that sounds promissing ! if you've an account at SO please post your answer ther so I can accept it. thank you very much c_14 :)
[23:10:38 CEST] <seasc> c_14, still having the same error (configuration mismatch) and i have no idea why its refering to NON-$CHROOt directory/ies ?? Same configure line (the dots indicating --enable-libXY, etc) : http://pastebin.com/09ZjvS7z
[23:11:07 CEST] <seasc> Could it be that *some* built codec/library is refering to outside of $CHROOT?
[23:11:17 CEST] <c_14> seasc: did you export LD_LIBRARY_PATH before running ffmpeg?
[23:11:39 CEST] <DragonsLord> llogan, I put the option -report in order to collect a log
[23:11:49 CEST] <seasc> c_14, yes with no change
[23:12:05 CEST] <seasc> neither for run the compiled, nor for after compiling another one
[23:12:44 CEST] <c_14> can you pastebin the output of `ls $LD_LIBRARY_PATH'
[23:14:47 CEST] <seasc> c_14, when i try to use --enable-x264 : http://pastebin.com/NwhTiV89  ..... only libnettle files and pkgconfig folder
[23:15:12 CEST] <c_14> Can you pastebin config.log ?
[23:15:23 CEST] <seasc> so changing that dir should help? of ffmpeg i assume?
[23:15:49 CEST] <c_14> The problem is usually that ffmpeg can't find the pkgconfig files
[23:16:02 CEST] <c_14> You can check with pkgconfig --exists --print-errors libx264
[23:16:20 CEST] <seasc> http://pastebin.com/YypfwneV
[23:16:33 CEST] <seasc> i'll try setting another path to LD_LIBRARY_PATH
[23:16:44 CEST] <seasc> coz libnettle is not used anymore anyway
[23:17:15 CEST] <c_14> And LD_LIBRARY_PATH needs to point to the location where the ffmpeg libraries are installed
[23:17:55 CEST] <c_14> oh, and it's x264 not libx264 (for the pkgconfig check)
[23:18:50 CEST] <seasc> hmm pkgconfig 127: command not found
[23:19:01 CEST] <c_14> It's pkg-config
[23:20:06 CEST] <c_14> In short, you either need pkg-config to find the .pc files (by setting PKG_CONFIG_PATH accordingly), or you need to setup the include and link paths correctly (by setting the CFLAGS AND LDFLAGS)
[23:22:48 CEST] <DragonsLord> Does anyone know what [mov,mp4,m4a,3gp,3g2,mj2 @ 0000000000321c40] Invalid SampleDelta -448 in STTS, a
[23:22:48 CEST] <DragonsLord> t 2 st:0 mean?
[23:22:53 CEST] <seasc> c_14, i dont have any .pc files (of 264/265)...  but i dont really care about that as long ffmpeg shows a library config error.. or are these 2 connected?
[23:24:35 CEST] <c_14> They are related. ffmpeg uses pkg-config (which uses the .pc files) for much of it's dependency detection
[23:26:42 CEST] <c_14> x264 detection still has a fallback if your CFLAGS and LDFLAGS are set correctly though
[23:30:53 CEST] <DragonsLord> c_14, Do you know what the error I have just pasted means?
[23:31:00 CEST] <c_14> No clue, sorry
[23:37:31 CEST] <ssss> @seasc I'm not sure it it's related  but I had the exact same error while trying to run the ffmpeg inside the src_dir and not the one in /usr/local/bin/ffmpeg
[23:40:53 CEST] <seasc> c_14, you asked for the LD_LIBRARY_PATH, which looks alot better than just the the nettle.so's from the previous attmept.. http://pastebin.com/hhNMVCGv
[23:41:22 CEST] <seasc> ssss, thanks for the hint, not sure it applies, since i execute the comiled ffmpeg from its $CHROOT/bin
[23:41:34 CEST] <c_14> ye, that looks better
[23:43:53 CEST] <smo_> hi, do you think the patchs in ffmpeg git "avcodec: add unpack packed B-frames bitstream filter..." can fix slow decoding of some avi files with the message "Video uses a non-standard and wasteful way to store B-frames"
[23:43:54 CEST] <smo_> ...
[23:43:55 CEST] <smo_> ?
[23:45:03 CEST] <kepstin-laptop> no, the bitstream filter is only useful in cases where you're not decoding the video, e.g. if you're using -c:v copy to turn it into an mkv file
[23:46:42 CEST] <smo_> arf go this problem when build ffmpeg in chromium (libffmpegsumo), don t know how to fix this small artefact problem
[23:59:29 CEST] <DragonsLord> c_14, ffmpeg -i "concat:video01.avi|video02.avi|video03.avi|video04.avi|video05.avi" -c copy video.avi ... Does it make sense for you?
[23:59:43 CEST] <c_14> Don't use the concat protocol.
[23:59:46 CEST] <c_14> Use the demuxer
[00:00:00 CEST] <c_14> https://trac.ffmpeg.org/wiki/Concatenate
[00:00:00 CEST] --- Tue Apr  7 2015


More information about the Ffmpeg-devel-irc mailing list