[Ffmpeg-devel-irc] ffmpeg.log.20150610

burek burek021 at gmail.com
Thu Jun 11 02:05:02 CEST 2015


[00:55:04 CEST] <rig-bone> hi
[00:55:27 CEST] <rig-bone> I need to get the iframes of a mp4 as yuv420p
[00:55:40 CEST] <rig-bone> what is the command for this
[00:58:19 CEST] <Compn> anoop_r : what can i help you with ?
[01:00:02 CEST] <llogan> rig-bone: maybe this "select='eq(pict_type,I)',format=yuv420p"
[01:02:12 CEST] <rig-bone> from the cli?
[01:02:14 CEST] <anoop_r> ok
[01:03:01 CEST] <llogan> rig-bone: yes, i assumed you want to use the ffmpeg cli tool
[01:03:32 CEST] <anoop_r> i am using ffmpeg cli with a c# gui
[01:03:48 CEST] <llogan> that's just the filtergraph. you'll need to add whatever other options you need or want
[01:04:34 CEST] <anoop_r> ffmpeg -i Air.mkv -c:v libx265 -preset ultrafast -x265-params crf=30 -c:a aac -strict experimental -q:a 10 output.mp4
[01:05:00 CEST] <anoop_r> but media info shows audio is still constant bit rate
[01:06:45 CEST] <Compn> anoop_r : try -b 128k instead of -q:a 10
[01:07:06 CEST] <anoop_r> that works
[01:07:08 CEST] <ac_slater_> hey all. I'm muxing 2 streams into MPEGTS via simple av_write_frame and a custom AVIO context. This is working great. Anyone know how I can get packet durations in microseconds?
[01:07:12 CEST] <anoop_r> but vbr mode not working
[01:08:15 CEST] <durandal_1707> i do not think native aac support vbr
[01:08:33 CEST] <llogan> how do you know mediainfo isn't wrong?
[01:09:30 CEST] <anoop_r> well can tell me anyother tool to determine
[01:09:54 CEST] <anoop_r> vbr support as per official ffmpeg docs
[01:10:16 CEST] <llogan> if you use another -aac_coder option it may show up as "variable" in medianifo
[01:10:23 CEST] <llogan> ffmpeg -h encoder=aac
[01:10:30 CEST] <llogan> results may vary.
[01:11:36 CEST] <anoop_r> i had used this command line ffmpeg -i input.wav -strict experimental -c:a aac -q:a 6 output.m4a
[01:12:48 CEST] <durandal_1707> ac_slater: from packet duration and stream dps iirc
[01:13:21 CEST] <rig-bone> llogan: what would the entire command be?
[01:13:36 CEST] <llogan> i don't know. you didn't mention your desired output format
[01:14:05 CEST] <anoop_r> ffmpeg -i Air.mkv -c:v libx265 -preset ultrafast -x265-params crf=30 -c:a aac -strict experimental -q:a 10 output.mp4
[01:14:30 CEST] <llogan> yes, you showed that or a variant of it at least 3 times i think
[01:14:42 CEST] <anoop_r> ya
[01:15:00 CEST] <llogan> you ignored the suggestions i gave you
[01:15:01 CEST] <anoop_r> i did it as instructed on the docs
[01:15:17 CEST] <llogan> wiki != docs
[01:15:32 CEST] <anoop_r> ok
[01:15:52 CEST] <anoop_r> then can you give me corrected command line
[01:15:58 CEST] <anoop_r> for variable bit rate
[01:16:12 CEST] <anoop_r> i don't have libfdk
[01:16:21 CEST] <anoop_r> i only use natvie aac
[01:16:29 CEST] <llogan> see -aac_coder option in "ffmepg -h encoder=aac"
[01:17:19 CEST] <llogan> see if any of those options will make mediainfo give you what you think you want
[01:19:18 CEST] <anoop_r> anywhere can i get detailed discrption on the modes specified
[01:19:43 CEST] <anoop_r> AAC encoder AVOptions:
[01:19:45 CEST] <anoop_r>   -stereo_mode       <int>        E...A... Stereo coding method (from -1 to 1) (default 0)
[01:19:45 CEST] <anoop_r>      auto                         E...A... Selected by the Encoder
[01:19:45 CEST] <anoop_r>      ms_off                       E...A... Disable Mid/Side coding
[01:19:45 CEST] <anoop_r>      ms_force                     E...A... Force Mid/Side for the whole frame if possible
[01:19:45 CEST] <anoop_r>   -aac_coder         <int>        E...A...  (from 0 to 3) (default 2)
[01:19:45 CEST] <anoop_r>      faac                         E...A... FAAC-inspired method
[01:19:45 CEST] <anoop_r>      anmr                         E...A... ANMR method
[01:19:46 CEST] <anoop_r>      twoloop                      E...A... Two loop searching method
[01:19:46 CEST] <anoop_r>      fast                         E...A... Constant quantizer
[01:19:47 CEST] <anoop_r>   -aac_pns           <int>        E...A... Perceptual Noise Substitution (from 0 to 1) (default 0)
[01:19:47 CEST] <anoop_r>      disable                      E...A... Disable PNS
[01:19:48 CEST] <anoop_r>      enable                       E...A... Enable PNS (Proof of concept)
[01:20:06 CEST] <llogan> use a pastebin service if you ever need to paste multiple lines on IRC
[01:20:20 CEST] <llogan> http://ffmpeg.org/ffmpeg-codecs.html#aac
[01:20:31 CEST] <anoop_r> sorry
[01:27:40 CEST] <anoop_r> thanks
[01:27:56 CEST] <anoop_r> i had used -q:a 10 -aac_coder faac and now it shows variable
[01:31:29 CEST] <anoop_r> anyway to show progress as % in ffmpeg encoding
[01:36:50 CEST] <anoop_r> http://pastebin.com/ijsWbG62
[01:37:17 CEST] <anoop_r> anway to get rid of Past duration 0.833321 too large messages
[01:46:47 CEST] <bbloom_> is it possible for keyframes to be split across multiple packets?
[01:49:43 CEST] <Agrajag-> g'day, i'm trying to use ffmpeg to grab audio from an IP webcam and then use sox to identify if there was noise during that time (baby monitor kind of thing). i can do this with "ffmpeg -i 'rtsp://192.168.1.10:88/videoMain' -t 5 -vn -acodec copy -f wav - 2>/dev/null | sox -t .wav - -n stat | grep 'Maximum amplitude'". i could keep doing this in a loop for every 5 seconds, however it seems to take some time (~5secs) to setup reading from the
[01:54:21 CEST] <rig-bone> llogan: I need to be able to extract an iframe from a mpeg4 h.264 video
[01:54:38 CEST] <rig-bone> and I want that iframe to be in its internal y'uv format
[01:54:48 CEST] <rig-bone> so I can see the bits, binary wise
[01:59:45 CEST] <rig-bone> $ ffmpeg -i example.mp4 test.yuv
[02:00:04 CEST] <rig-bone> how can I do this, the command above fails
[02:06:55 CEST] <rig-bone> llogan: another fast idea, can ffmpeg tell me the planar packing that an mp4 video is using in info?
[02:07:48 CEST] <rig-bone> is that a part of the video info that ffmpeg can find
[02:13:40 CEST] <yumbox> hello. How can I see the encoding options of a VP9 encoded webm?
[02:17:10 CEST] <rig-bone> ffmpeg -i file.webm
[02:19:15 CEST] <yumbox> rig-bone: that doesnt show all encoding options.
[02:19:25 CEST] <yumbox> that hardly shows any.
[02:21:00 CEST] <yumbox> rig-bone: this is what I'm looking for: http://sprunge.us/fHeR
[02:21:33 CEST] <yumbox> however, on a webm, this is the sparse output: http://sprunge.us/UOiL
[02:21:43 CEST] <yumbox> it doesnt show any encoding options
[02:23:10 CEST] <llogan> i believe that comes from the external encoding library itself and libvpx probably just doesn't include that info, AFAIK
[02:23:38 CEST] <rig-bone> oh there you are llogan
[02:23:52 CEST] <yumbox> when I encode a file to webm I can use ffmpeg -loglevel verbose
[02:23:55 CEST] <rig-bone> hopefully I believed giving faulty information would bring you back
[02:23:58 CEST] <yumbox> it then shows me the encoding options
[02:24:07 CEST] <rig-bone> so, can you help me after yumbox
[02:24:22 CEST] <llogan> rig-bone: sorry, i didn't know an answer for your question
[02:25:02 CEST] <yumbox> llogan: so its impossible to determite the orignial encoding options of a VP9 encoded file?
[02:25:44 CEST] <llogan> i'm not sure. i just don't think libvpx provides that info like libx264 does
[02:26:01 CEST] <yumbox> llogan: it does show it before encoding
[02:26:07 CEST] <yumbox> before/during
[02:26:40 CEST] <llogan> that's just the verbose log of ffmpeg parsing the options you give it
[02:27:03 CEST] <yumbox> llogan: but it shows this: [libvpx-vp9 @ 0x7f83178edc00]
[02:27:10 CEST] <llogan> hell, i don't know
[02:29:26 CEST] <yumbox> llogan: http://sprunge.us/IaaY
[02:30:15 CEST] <llogan> ah, i see. i thought you were referring to the "Reading option" stuff.
[02:30:45 CEST] <yumbox> when I use "-crf 33" instead of "-crf default" I get this message:
[02:30:46 CEST] <yumbox> [libvpx-vp9 @ 0x7ff83c07ac00]   VP8E_SET_CQ_LEVEL:            33
[02:31:04 CEST] <yumbox> so I can't seem to figure out what the default crf is
[02:31:44 CEST] <yumbox> the wiki of vp9 just lists crf can be set 0-63 with lower being higher quality
[02:31:52 CEST] <yumbox> it doesnt list the default value
[02:37:46 CEST] <yumbox> llogan: is there a different channel where I might find answers to this question?
[02:38:59 CEST] <llogan> ffmpeg-user mailing list
[02:40:12 CEST] <rig-bone> how do I convert mp4 to y4m
[02:40:29 CEST] <rig-bone> since that would be an upcoding process anyway
[02:40:31 CEST] <rig-bone> how
[02:41:33 CEST] <yumbox> I dont think ive ever heard of upcoding
[02:43:27 CEST] <rig-bone> what I mean by upcoding is that y4m is a raw format
[02:43:31 CEST] <rig-bone> and mp4 is a compressed format
[02:43:57 CEST] <rig-bone> I'm asking to "upcode" which is that I want to get the mp4 video in a y4m format
[02:44:08 CEST] <rig-bone> I know it isn't directly possible
[02:44:21 CEST] <yumbox> rig-bone: did you try just doing ffmpeg -i file.mp4 file.y4m ?
[02:44:44 CEST] <rig-bone> so I'm instead asking how I could get the iframes of the mp4 video and convert those as yuv frames
[02:45:26 CEST] <rig-bone> :encoder did not produce proper points, making some up:
[02:45:57 CEST] <rig-bone> then it surprisingly encodes, but I'm not sure if this is what I want
[02:46:16 CEST] <rig-bone> maybe there's a more sane way than trying to create a raw video from a compressed one
[02:47:09 CEST] <voltagex> llogan: after compiling an ffmpeg with vbi support I still can't extract subtitles
[02:49:09 CEST] <rig-bone> infact, each time I run this command, it literally chooses "RANDOM" frames from the video to make the mp4 from
[02:49:18 CEST] <rig-bone> I get different output each time
[02:51:08 CEST] <llogan> voltagex: try ffmpeg-user mailing list
[02:55:27 CEST] <llogan> rig-bone: maybe 'ffmpeg -i input -vf "select='eq(pict_type,I)',format=yuv420p" output.yuv' will give you want. add -vsync (0 or 2 I think) if you don't like how it repeats the frame to fill empty PTS
[02:58:22 CEST] <rig-bone> yeah its choosing random pts for me because the encoder isn't producing them
[03:16:09 CEST] <rig-bone> llogan: I'd kiss you on the cheek right now if I could
[03:16:14 CEST] <rig-bone> thanks man
[03:18:34 CEST] <rig-bone> its odd how my original video was 750 frames and this has become 794
[03:18:40 CEST] <rig-bone> i'll find out why
[03:37:19 CEST] <rig-bone> wee, so switching -vsync vfr gives me JUST the iframes
[03:37:25 CEST] <rig-bone> this is exactly what I needed!
[03:37:34 CEST] <rig-bone> oh if only I could donate money
[03:37:37 CEST] <rig-bone> I would
[03:37:39 CEST] <rig-bone> hehe
[03:45:02 CEST] <rig-bone> llogan can I force it to select certain pts
[03:52:38 CEST] <llogan> rig-bone: http://ffmpeg.org/ffmpeg-filters.html#select_002c-aselect
[05:53:34 CEST] <_vaibhavingale_> I have converted the flac file to m4a. and tried to play using html5 audio tag. The song played back within a second and when I check in the network stats of browser, the downloaded stream was of just 100kb when stream started playback. One thing I observe this is not consistent with every raw file. z.B I have converted another flac file in m4a but it behaves like mp3 i.e taking around 6-10 seconds before actual playback and gets downloaded atlast 3
[05:54:17 CEST] <_vaibhavingale_> can some one tell me or show me the right direction to git it done.
[05:55:06 CEST] <_vaibhavingale_> *get . I want to stream the static audio files in the browser but there should not be delay for prebuffering to start the playback
[05:58:05 CEST] <relaxed> _vaibhavingale_: are you using -movflags faststart?
[06:42:07 CEST] <_vaibhavingale_1> No.
[06:42:51 CEST] <relaxed> do so
[06:43:16 CEST] <_vaibhavingale_1> I used -vn flag to get rid of meta and album art
[06:43:25 CEST] <_vaibhavingale_1> cool let me try
[07:17:38 CEST] <techtopia> hello
[07:18:15 CEST] <techtopia> lets say i have a 3 min long video, and i want to encode the first 60 seconds and the last 60 seconds
[07:18:25 CEST] <techtopia> but want to omit the middle 60 seconds
[07:18:44 CEST] <techtopia> is their anyway to do this from an ffmpeg encoding line?
[07:18:56 CEST] <techtopia> with first cutting the video with another application
[07:29:48 CEST] <chungy> techtopia: you'd probably have to use intermediate files and then concat it
[07:31:13 CEST] <techtopia> yeah it's fine for it to be encoded to 2 files
[07:31:16 CEST] <techtopia> and concat later
[07:31:36 CEST] <techtopia> im just looking for a way to not have to process videos before feeding them to ffmpeg
[07:32:38 CEST] <techtopia> at the moment that 3 minute example video would get cut before hand into input01.ts and input02.ts and would get encoded seperatly then get concatted
[07:32:51 CEST] <techtopia> if i can cut the first step out it will save me a lot of hasstle
[07:44:49 CEST] <_vaibhavingale_> @relaxed: Thanks a lot, it reduced the delay significantly.
[07:45:07 CEST] <relaxed> you're welcome
[08:37:11 CEST] <harisk92> "expansion=strftime is deprecated"
[08:37:16 CEST] <harisk92> ?
[08:39:30 CEST] <harisk92> what do I use now
[08:39:36 CEST] <harisk92> for drawtext
[08:40:42 CEST] <harisk92> to set time for text value in this format %Y-%m-%d %H:%M:%S ?
[08:51:28 CEST] <harisk92> guys
[09:10:18 CEST] <durandal_1707> isnt it in dica?
[09:10:31 CEST] <durandal_1707> docs
[09:10:41 CEST] <c_14> harisk92: :text=%{localtime\:%a %b %d %Y} <- like that
[09:27:42 CEST] <harisk92> c_14: I found it
[09:27:47 CEST] <harisk92> I need gmtime
[09:27:50 CEST] <harisk92> ty
[09:31:12 CEST] <harisk92> hey c_14 I have this text='%{gmtime\:%Y-%m-%d %H\:%M\:%S}'
[09:31:15 CEST] <harisk92> and I get
[09:31:24 CEST] <harisk92> gmtime requires 1 argument at most
[09:32:34 CEST] <c_14> harisk92: probably escaping breaking somewhere
[09:33:10 CEST] <harisk92> hmm I don't need to escape anything in text file ?
[09:33:12 CEST] <harisk92> right
[10:06:47 CEST] <harisk92> c_14 when positioning drawtext
[10:06:55 CEST] <harisk92> can I use ih and iw
[10:08:06 CEST] <c_14> The values you can use are listed in the documentation, you can use h and w
[10:08:16 CEST] <harisk92> I tried to do following :x=(iw-20):y=(ih-20)
[10:08:48 CEST] <harisk92> owh c_14
[10:08:50 CEST] <harisk92> i see
[10:08:57 CEST] <harisk92> i can use W,w or
[10:09:12 CEST] <harisk92> main_w
[10:24:35 CEST] <RobertNagy> Is it possible to get webp working with piping/streams?
[10:25:02 CEST] <RobertNagy> Right now ffmpeg will output webp to a pipe however the output will be corrupted
[10:25:28 CEST] <RobertNagy> probably because the encoder tries to go back to the beginning of the file and write information in the header after finishing encoding a frame.
[10:25:44 CEST] <RobertNagy> which is not supported when streaming
[10:26:15 CEST] <RobertNagy> not sure if it is possible to fix, but ffmpeg should atleast throw an error?
[10:31:15 CEST] <c_14> If webp requires seeking backwards in output, then yes ffmpeg should throw an error. Can you make a bugreport?
[10:42:03 CEST] <relaxed> RobertNagy: is your input webp?
[10:42:07 CEST] <RobertNagy> no
[10:42:12 CEST] <RobertNagy> the output is
[10:42:38 CEST] <relaxed> what are you piping the output to?
[10:42:50 CEST] <RobertNagy> a http response
[10:43:45 CEST] <RobertNagy> I think the encoder should probably encode the frame in memory, write the header and then write to output
[10:43:58 CEST] <RobertNagy> that way it would work for piping as well
[10:45:02 CEST] <RobertNagy> https://github.com/FFmpeg/FFmpeg/blob/6b56fcbd10a56d2a7a116c0de4ee032b895456c2/libavformat/webpenc.c
[10:45:43 CEST] <RobertNagy> *muxing the frame in memory
[11:08:35 CEST] <RobertNagy> done, https://trac.ffmpeg.org/ticket/4609
[11:59:44 CEST] <t4nk626> Hello All,
[12:00:02 CEST] <t4nk626> I am trying to execute the following command
[12:00:03 CEST] <t4nk626> ahttp://pastebin.com/fYAwJeAn
[12:00:13 CEST] <t4nk626> http://pastebin.com/fYAwJeAn
[12:00:41 CEST] <t4nk626> and it keep saying : Unrecognized option 'filter_complex "[0:0] [0:1] [1:0] [1:1] concat=n=2:v=1:a=1 [v] [a]" -map "[v]" -map "[a]" -vcodec libx264'.
[12:00:55 CEST] <t4nk626> Any idea?
[13:32:20 CEST] <smo_80> hi
[13:32:52 CEST] <smo_80> is it possible to cross compil ffmpeg for android with pthreads enabled ?
[13:35:51 CEST] <beaver> smo_80: :)
[13:36:47 CEST] <beaver> smo_80: oui avec --enable-pthreads
[13:37:16 CEST] <beaver> http://docs.collectiveaccess.org/wiki/Compiling_ffmpeg
[14:29:58 CEST] <harisk92> hey c_14 is it possible to use epoch timestamp in output filename ?
[14:31:53 CEST] <c_14> harisk92: just use %s ?
[14:32:17 CEST] <harisk92> thx
[14:33:09 CEST] <zachary> ffmpeg publishes camera to rtmp server, but get error: av_interleaved_write_frame(): Operation not permitted
[14:33:35 CEST] <zachary> ffmpeg -rtbufsize 1500M -r 25 -f dshow -i video="Integrated Camera":audio="Internal Microphone (Conexant 20672 SmartAudio HD)" -threads 4 -vcodec libx264 -b:v 150k -preset ultrafast -pix_fmt yuv420p -tune zerolatency -acodec libmp3lame -b: a 32k -bufsize 3000k -f flv rtmp://ip/app
[15:05:48 CEST] <hendry> hi, has anyone here created a video editor with ffmpeg? wondering how brain dead http://dabase.com/blog/A_suckless_video_editor_proposal/ is...
[15:13:19 CEST] <Compn> hendry : well, hes basically talking about a nle editor except done on the command line. which is just a simple edit decision list format
[15:14:12 CEST] <Compn> so to answer your question, adding EDL support to ffmpeg would be fairly easy
[15:14:32 CEST] <Compn> mplayer has a very very simple edl system already
[15:14:39 CEST] <Compn> i think vlc was talking about getting one too
[15:14:55 CEST] <Compn> just seems like most people dont care about it and just run it manually
[15:20:34 CEST] <Compn> are there already video editors using ffmpeg? of course, http://www.openshotvideo.com/ https://github.com/jliljebl/flowblade http://www.blender.org
[15:21:06 CEST] <JEEBsv> hendry: you can also look into vapoursynth and vapoursynth editor
[15:21:21 CEST] <JEEBsv> there's two source filters in that based on lavf/lavc
[15:21:37 CEST] <JEEBsv> not sure if there's a lavfi based filtering thing around but I would guess it wouldn't be too hard to implement
[15:23:33 CEST] <zachary> ffmpeg -rtbufsize 1500M -r 25 -f dshow -i video="Integrated Camera":audio="Internal Microphone (Conexant 20672 SmartAudio HD)" -threads 4 -vcodec libx264 -b:v 150k -preset ultrafast -pix_fmt yuv420p -tune zerolatency -acodec libmp3lame -b: a 32k -bufsize 3000k -f flv rtmp://ip/app
[15:23:35 CEST] <Compn> any screenshots of vaporsynth on the vaporsynth website? :P
[15:23:48 CEST] <zachary> could some guy help check my ffmpeg command error?
[15:24:38 CEST] <Compn> zachary : paste error to pastebin
[15:26:20 CEST] <zachary> Compn: http://pastebin.com/WNKbAE3M
[15:26:55 CEST] <JEEBsv> Compn: not much to screenshot with vapoursynth itself since its main interface is Python scripting. Then there's the Vapoursynth Editor, which lets you do simple preview/code completion etc
[15:27:22 CEST] <zachary> each time, the command can only run around 1 minute
[15:30:24 CEST] <zachary> Compn: http://pastebin.com/J5j7yUH3, this is a more detailed pastebin
[15:35:18 CEST] <Compn> zachary : looks like problem on your rtmp side
[15:35:39 CEST] <Compn> did you provide the correct rtmp url and authorization/login to the rtmp ?
[15:35:55 CEST] <zachary> Compn: i am pretty sure, url correct
[15:36:02 CEST] <Compn> what ver ffmpeg ?
[15:36:15 CEST] <Compn> nightly build or so ?
[15:36:30 CEST] <zachary> Compn: also, the ffmpeg runs around 30 minutes, and then failed with the error above
[15:36:43 CEST] <zachary> i get it days ago
[15:37:27 CEST] <zachary> Compn: ffmpeg version N-72734-g5cf84e5
[15:37:44 CEST] <Compn> is your network fast enough ?
[15:37:47 CEST] <Compn> or wifi problems ?
[15:41:30 CEST] <Compn> zachary : i see this thread on the forum looks exactly the smae problem as you , https://ffmpeg.zeranoe.com/forum/viewtopic.php?t=657&start=40
[15:43:36 CEST] <Compn> someone said that ffmpeg tcp is crap and they use tcp relay and it works fine
[15:43:39 CEST] <zachary> Compn: it seems this post does not give solution
[15:43:42 CEST] <Compn> a few pages into that thread
[15:43:46 CEST] <zachary> Compn: i will look into it
[15:44:11 CEST] <Compn> you have to read whole thread :D
[15:46:33 CEST] <Compn> kind of sucks we need a librtmp maintainer
[15:49:06 CEST] <zachary> Compn: i will install tcp relay as a try
[15:49:12 CEST] <zachary> and come back later on
[15:49:23 CEST] <zachary> Compn: thanks very much for your help.
[15:51:18 CEST] <Compn> no pr oblem
[15:58:21 CEST] <nyuszika7h> how can I make a "windowboxed" (black bars on all sides) video fill the whole screen?
[16:07:50 CEST] <k_sze> hmm, is the new NVENC or QSV encoding stuff redistributable?
[16:09:03 CEST] <c_14> nvenc isn't
[16:09:20 CEST] <k_sze> but qsv is?
[16:11:10 CEST] <k_sze> Would be nice if I can take advantage of QSV.
[16:11:32 CEST] <k_sze> Does it have any limitation in terms of pixel format or resolution?
[16:13:01 CEST] <JEEBsv> 8bit 4:2:0 YCbCr
[16:13:31 CEST] <JEEBsv> resolution wise I have no idea, with older stuff you probably have to limit yourself to 1080p, newer stuff might support 2160p
[16:18:24 CEST] <k_sze> 420...
[16:18:32 CEST] <k_sze> useless for me then.
[16:18:36 CEST] <k_sze> I need at least 422. :(
[16:22:28 CEST] <JEEBsv> if nvenc supported that I'd be surprised :P
[16:25:06 CEST] <Compn> nyuszika7h : vf expand
[16:26:08 CEST] <k_sze> hmm, there's also mmal
[16:26:14 CEST] <k_sze> what kind of hardware provides mmal?
[16:28:11 CEST] <k_sze> ARM processors?
[16:30:04 CEST] <k_sze> JEEBsv: nvenc supports YUV444 lossless since the second generation (Maxwell GM107/GM108)
[16:30:08 CEST] <k_sze> http://en.wikipedia.org/wiki/Nvidia_NVENC
[16:30:52 CEST] <JEEBsv> yeah, but that's just the profile's name
[16:31:12 CEST] <JEEBsv> unless it specifically says it can encode 4:4:4 YCbCr
[16:31:41 CEST] <k_sze> "it" as in "ffmpeg"?
[16:31:44 CEST] <JEEBsv> the wikipedo article seems to hint at that, but I'd honestly be quite surprised. I would expect 4:2:0 YCbCr lossless
[16:32:01 CEST] <k_sze> I'm not surprised, actually.
[16:32:20 CEST] <k_sze> Maxwell-based GPUs are really pretty freaking powerful.
[16:32:34 CEST] <JEEBsv> that doesn't matter, the encoding is done with ASICs
[16:32:38 CEST] <JEEBsv> not the GPGPU stuff
[16:32:41 CEST] <k_sze> We do machine vision stuff with it and it's really mind blowing.
[16:32:52 CEST] <k_sze> true
[16:33:14 CEST] <JEEBsv> anyways, does the 960 enter that sphere?
[16:33:20 CEST] <JEEBsv> I could try whatever it has :P
[16:33:20 CEST] <k_sze> but I mean, I would expect the ASIC to be powerful as well.
[16:33:34 CEST] <JEEBsv> yes, but end-user ASICs really don't do !4:2:0
[16:33:44 CEST] <JEEBsv> encoders are a bit more of a special case, but still
[16:33:51 CEST] <k_sze> gtx 960 should be maxwell
[16:37:21 CEST] <JEEBsv> ok, I'll be damned
[16:37:46 CEST] <JEEBsv> it seems to be able to do 4:4:4 according to documentation and what the input pixel format stuff notes
[16:38:01 CEST] <k_sze> see? :)
[16:38:21 CEST] <JEEBsv> which kind of means that you could even do lossless RGB with it if you just fed it RGB instead of YCbCr (and set the flags right for it to be flagged correctly instead of YCbCr)
[16:38:45 CEST] <JEEBsv> might be useful for screen capture as long as your HDD can handle it
[16:38:58 CEST] <k_sze> Well, lossless is lossless, you could even pretend the RGB is YCbCr. :P
[16:39:14 CEST] <JEEBsv> you don't need to pretend
[16:39:28 CEST] <JEEBsv> unless nvidia forcibly writes into parameter sets that it's YCbCr
[16:39:41 CEST] <JEEBsv> 'cause the amount of data and its structure would be the same
[16:40:26 CEST] <k_sze> Yes
[16:40:42 CEST] <k_sze> but that's all beside the point.
[16:40:56 CEST] <k_sze> I'm working with some HP ProDesk mini PCs.
[16:41:02 CEST] <k_sze> can't stick an nvidia card in. :P
[16:41:17 CEST] <k_sze> :(
[16:44:17 CEST] <k_sze> It's hard to believe that this is not a solved problem with newish CPUs and their integrated graphics.
[16:45:35 CEST] <k_sze> I mean, jee, I'm only doing 1080p at 30 fps. What happens when people need to do like 4k video at 200+ fps with those crazy CameraLink cameras.
[16:52:00 CEST] <Gonbe> Hi, I've read this: http://wiki.hydrogenaud.io/index.php?title=Fraunhofer_FDK_AAC#FDK_License this: https://ffmpeg.zeranoe.com/forum/viewtopic.php?t=877 and this: https://trac.ffmpeg.org/wiki/Encode/AAC and skimmed through this: https://lists.ffmpeg.org/pipermail/ffmpeg-devel/2013-February/139156.html
[16:52:02 CEST] <Gonbe> So basically, is there no way to legally distribute ffmpeg with libfdk support (for windows)? Is building ffmpeg with everything dynamically linked and including the dynamic link library for libsdk also not an option for distribution?
[16:53:38 CEST] <c_14> What you're looking for, is a lawyer.
[16:54:52 CEST] <c_14> afaik dynamic linking + distributing is now allowed unless the library is a "system library"
[16:55:05 CEST] <c_14> s/now/not/
[16:56:45 CEST] <Gonbe> c_14: thanks.
[17:01:03 CEST] <Gonbe> by system libraries, within the context of windows, I'm assuming: http://en.wikipedia.org/wiki/Microsoft_Windows_library_files is that correct? if not that can be quite ambiguous&
[17:01:27 CEST] <c_14> It's about as ambiguous as your nearest lawyer decides it to be.
[17:01:33 CEST] <Gonbe> hahaha
[17:02:17 CEST] <c_14> Technically the nearest Judge.
[18:41:41 CEST] <xreal> Is this correct to select streams? -c:v:3 libx264 -preset faster -crf 30 -c:a:5 copy
[18:41:56 CEST] <xreal> argh, I didn't use -map.
[20:04:23 CEST] <nyuszika7h> is it possible to import subtitle timings from one srt into another?
[20:11:59 CEST] <gouessej> Hi
[20:12:18 CEST] <gouessej> I'm trying to convert a MP3 video into OGV
[20:12:45 CEST] <gouessej> I use the following command line: ffmpeg -i Video20150609_185820.mp4 Video20150609_185820.ogv -qscale 0
[20:12:56 CEST] <gouessej> I use FFMPEG 2.7, the latest static vuild
[20:12:58 CEST] <gouessej> build
[20:13:52 CEST] <gouessej> the output framerate is ... miserable
[20:13:56 CEST] <gouessej> about 4 fps
[20:14:22 CEST] <gouessej> whereas the input has a framerate of about 30 fps
[20:16:09 CEST] <gouessej> Done: http://pastebin.com/KhmGpmW4
[20:16:47 CEST] <JEEBsv> do you mean the frame rate as the encoding rate?
[20:16:50 CEST] <JEEBsv> or when you watch the clip?
[20:17:11 CEST] <JEEBsv> oh
[20:17:13 CEST] <gouessej> when I watch the clip
[20:17:13 CEST] <JEEBsv> I see
[20:17:18 CEST] <gouessej> even with ffplay
[20:17:21 CEST] <JEEBsv> yeah
[20:17:25 CEST] <JEEBsv> I just noticed your issue
[20:17:37 CEST] <JEEBsv> you see, options for encoding go between the input and the output
[20:17:49 CEST] <JEEBsv> see how it even tells you that trailing options were found
[20:17:49 CEST] <gouessej> I tried with -r and -framerate
[20:18:42 CEST] <JEEBsv> 1) you set qscale 0 to no output 2) default bit rate is 200kbps 3) theora will drop pictures if the rate is too low
[20:18:58 CEST] <llogan> i guess i should ask: why do you even want to use theora?
[20:19:18 CEST] <gouessej> I wanted to use -sameq
[20:19:23 CEST] <gouessej> but it has been dropped
[20:19:28 CEST] <JEEBsv> which is not what you think it is
[20:19:37 CEST] <JEEBsv> anyway, did you read what I wrote?
[20:19:43 CEST] <JEEBsv> it should be rather obvious what your issue is
[20:19:46 CEST] <JEEBsv> if you did
[20:19:46 CEST] <gouessej> I tried with -qscale:v 1
[20:19:58 CEST] <JEEBsv> gouessej: look at your command line
[20:20:14 CEST] <llogan> the location of the option matters
[20:20:15 CEST] <JEEBsv> options for encoding go BETWEEN input and output
[20:20:23 CEST] <gouessej> oh sorry
[20:20:38 CEST] <JEEBsv> as I said, it even tells you that there are trailing options
[20:20:50 CEST] <JEEBsv> it expected another output
[20:20:55 CEST] Action: llogan still wonders why theora is desired
[20:21:21 CEST] <JEEBsv> like `ffmpeg -i hurr.mp4 -c:v libx264 out.mkv -c:v libvpx out.webm`
[20:21:35 CEST] <JEEBsv> libx264 for the first output file, and libvpx for the seccond
[20:22:20 CEST] <JEEBsv> I wonder if ffmpeg could actually just error out at that trailing options message
[20:22:35 CEST] <gouessej> llogan: I had similar problems with webm, it was a bit better when I used mpeg1video
[20:22:44 CEST] <llogan> for theora, using qscale:v/q:v, 0 is the crappiest, and 10 is the least crappiest/most bloated
[20:22:45 CEST] <JEEBsv> lol
[20:23:04 CEST] <JEEBsv> gouessej: let me guess that you had trailing options in the same way :P
[20:23:31 CEST] <gouessej> To have a constant quality (but a variable bitrate), use the option -qscale n when n is between 1 (excellent quality) and 31 (worst quality).
[20:23:32 CEST] <gouessej> :s
[20:23:44 CEST] <gouessej> I thought that 1 would give me something very good
[20:23:45 CEST] <llogan> that's for the mpeg* family
[20:23:53 CEST] <llogan> that range.
[20:24:13 CEST] <JEEBsv> also constant quality is just not true
[20:24:17 CEST] <JEEBsv> it's constant quantizer
[20:24:34 CEST] <llogan> where did you find that suggestion?
[20:24:46 CEST] <JEEBsv> we don't have a single rate control mechanism that would have thousands of chinese children looking at various alternatives
[20:24:49 CEST] <gouessej> in the documentation: http://ffmpeg.org/ffmpeg.html
[20:25:03 CEST] <gouessej> http://ffmpeg.org/ffmpeg.html#Tips
[20:25:25 CEST] <JEEBsv> that's probably something written ages ago and nobody cares to edit it
[20:25:29 CEST] <llogan> i should burn those
[20:25:36 CEST] <llogan> "realvideo"
[20:25:48 CEST] <JEEBsv> it's mostly written in early 2000s or so
[20:26:06 CEST] <JEEBsv> the closest we have atm is libx264's and libx265's CRF, but even that is most definitely not "constant quality"
[20:26:53 CEST] <JEEBsv> libvpx also has the -crf option linked to some rate control of its own (which differs between VP8 and VP9, too), no idea about those but the last I remember from reading the VP8 docs it's different to what the libx264/265 algo does
[20:27:22 CEST] <JEEBsv> otherwise the non-bitrate based alternative is QP so at some point in time constant quantizer indeed was the closest we were :P
[20:27:32 CEST] <gouessej> QP?
[20:27:51 CEST] <JEEBsv> quantizer
[20:28:04 CEST] <smo_80> hi, just tring ffmpeg on my set top box under android with soc broadcom BCM7252 can i get hw accel with this ?
[20:28:09 CEST] <JEEBsv> but yeah, lemme see when CRF got implemented...
[20:28:23 CEST] <smo_80> got many many errors when doing transcoding :/ think it s too slow
[20:28:30 CEST] <JEEBsv> http://git.videolan.org/?p=x264.git;a=commit;h=0ddc9d5524a48882ac804948775fd7a35b3a07da
[20:28:34 CEST] <JEEBsv> almost ten years
[20:28:59 CEST] <JEEBsv> calls it "quality" but the same person scolded me years later when I happened to call it "constant quality"
[20:29:36 CEST] <gouessej> higher is better with crf, isn't it?
[20:29:41 CEST] <gouessej> I can give it a try
[20:30:24 CEST] <JEEBsv> it only works with libx264 and libx265
[20:30:57 CEST] <iive> higher crf means lower bitrate and more loss
[20:31:18 CEST] <JEEBsv> with x264 the default is 23, with x265 26. if the result of a few thousand frames (encode a few parts of your input at 500 frames or so) is bad, you lower. if it looks good you go higher.
[20:31:31 CEST] <JEEBsv> then at some point you find equilibrium at the highest value that still looks good for you
[20:32:34 CEST] <gouessej> Sorry for my dumb question but how can I force the use of libx264?
[20:32:42 CEST] <JEEBsv> -c:v libx264
[20:33:46 CEST] <gouessej> Could not write header for output file #0 (incorrect codec parameters ?): Invalid argument
[20:33:53 CEST] <JEEBsv> the error is generally before that
[20:34:03 CEST] <gouessej> [ogg @ 0x4b15be0] Unsupported codec id in stream 0
[20:34:06 CEST] <JEEBsv> yup
[20:34:14 CEST] <JEEBsv> no AVC/H.264 mappings for OGG
[20:34:18 CEST] <JEEBsv> use another container :P
[20:35:55 CEST] <gouessej> I don't want to use the format covered by software patents
[20:36:04 CEST] <JEEBsv> too bad
[20:36:05 CEST] <JEEBsv> everything is
[20:36:08 CEST] <gouessej> and I don't want to rely on Google
[20:36:56 CEST] <c_14> Use matroska?
[20:37:09 CEST] <JEEBsv> he seems to have an issue with AVC :P
[20:37:29 CEST] <c_14> Then don't use h264?
[20:37:29 CEST] <gouessej> c_14: Thanks, I'll give it a look
[20:37:49 CEST] <gouessej> My smartphone uses h264 by default :(
[20:37:50 CEST] <c_14> Though you'll be hard pressed to find a decent lossy video codec that isn't mpeg or maybe vp9.
[20:40:28 CEST] <gouessej> Is there anything that can give me a better result with ffmpeg?
[20:40:52 CEST] <c_14> Better result than what? Better in what way?
[20:41:31 CEST] <gouessej> Better framerate when I play the video
[20:42:21 CEST] <JEEBsv> either use an encoder that doesn't drop frames, or push the rate higher so whatever piece of shit you're using no longer drops frames
[20:43:09 CEST] <gouessej> How do you push the rate higher? -r or -framerate just before the output?
[20:43:36 CEST] <JEEBsv> when I talk of rate it's the BIT rate
[20:43:45 CEST] <gouessej> ok sorry
[20:43:45 CEST] <JEEBsv> by whatever rate control method you're using
[20:43:49 CEST] <llogan> JEEBsv: already told you what the issue was
[20:44:01 CEST] <llogan> type failure
[20:45:19 CEST] <llogan> just do this: ffmpeg -i input -q:v 5 -q:a 6 output.ogv
[20:45:26 CEST] <llogan> it will probably give you what you think you want
[20:46:56 CEST] <gouessej> I have just tried with -b:v 64k -bufsize 64k and it isn't better. I'm going to try llogan's suggestion
[20:47:58 CEST] <llogan> 64k for a 1920x1080 video using libtheora...........
[20:48:14 CEST] <JEEBsv> nîn
[20:48:19 CEST] <gouessej> llogan thanks, it's really better :)
[20:49:44 CEST] <gouessej> llogan: Please can you explain to me why it seems to drop no frame with your options? I'd like to understand
[20:50:05 CEST] <JEEBsv> because you set the rate so goddamn low
[20:50:08 CEST] <JEEBsv> 64 kilobits
[20:50:50 CEST] <JEEBsv> which is 64 000 bits (8000 bytes) per second
[20:51:43 CEST] <gouessej> Why does -q:v 5 -q:a 6 just work as expected? I'm sure I used something similar yesterday. I probably put the options in the wrong order
[20:51:49 CEST] <JEEBsv> yes
[20:52:04 CEST] <JEEBsv> if you had your stuff after the output they were ignored because the options would have been applied to the next output
[20:52:17 CEST] <JEEBsv> I really guess ffmpeg should error out in that case
[20:52:27 CEST] <JEEBsv> of course then you guys still wouldn't read
[20:52:37 CEST] <JEEBsv> you would just complain Y THIS NO WORK!!!!!
[20:52:56 CEST] <gouessej> No, I just want to understand and I really appreciate your explanations
[20:53:02 CEST] <gouessej> Really
[20:53:14 CEST] <gouessej> It's my fault, I should have read a bit more the man
[20:53:57 CEST] <gouessej> Should I guess that -q:v 1 -q:a 1 might preserve a bit more the video?
[20:54:24 CEST] <JEEBsv> you should have noted the comment before that theora's quantizer doesn't go like MPEG's :P
[20:54:27 CEST] <gouessej> oops llogan said the opposite earlier
[20:54:37 CEST] <llogan> "<logan> for theora, using qscale:v/q:v, 0 is the crappiest, and 10 is the least crappiest/most bloated"
[20:54:49 CEST] <llogan> same for Vorbis, AFAIK
[20:56:20 CEST] <llogan> some info here. https://trac.ffmpeg.org/wiki/TheoraVorbisEncodingGuide
[20:56:41 CEST] <gouessej> thanks llogan
[20:57:07 CEST] <llogan> http://ffmpeg.org/ffmpeg-codecs.html#libvorbis
[20:57:14 CEST] <llogan> http://ffmpeg.org/ffmpeg-codecs.html#libtheora
[21:04:57 CEST] <gouessej> Is it possible to edit the metadata without converting the video again?
[21:05:24 CEST] <llogan> yes. ffmpeg -i input -c copy -metadata title="Theora sucks" output
[21:08:04 CEST] <gouessej> What is wrong with threora in your humble opinion?
[21:08:45 CEST] <Compn> it sucks and there is no hw support for it
[21:13:18 CEST] <gouessej> ffplay plays the video correctly but not VLC, it shows "0:00" as a duration :s
[21:13:44 CEST] <gouessej> Stream #1: not enough frames to estimate rate; consider increasing probesize
[21:14:09 CEST] <llogan> it's old, and there are better modern alternatives
[21:16:04 CEST] <gouessej> How can I increase the probesize?
[21:18:41 CEST] <gouessej> The duration is 0, other players don't play the video
[21:18:47 CEST] <gouessej> Is it a bug?
[21:22:22 CEST] <llogan> i don't know. ask VLC
[21:23:20 CEST] <gouessej> I have increased the probesize
[21:24:13 CEST] <gouessej> Other players refuse starting the video, I hope that the increase of the probesize will help in fixing the duration
[21:26:47 CEST] <smo_80> hello
[21:27:10 CEST] <durandal_1707> hello
[21:27:10 CEST] <smo_80> what can do this type of errors?
[21:27:12 CEST] <smo_80> http://paste.ubuntu.com/11691892/
[21:27:25 CEST] <smo_80> slow hardware, slow connexion .... wrong config ...?
[21:28:29 CEST] <podman> afternoon!
[21:28:42 CEST] <gouessej> Increasing the probesize doesn't solve my problem
[21:29:46 CEST] <smo_80> (just builded ffmpeg ffmpeg version n2.4.2 for android
[21:30:48 CEST] <gouessej> ffprobe indicates the correct duration
[21:31:20 CEST] <podman> I'm trying to figure out what the following compile flags do but i'm having a hard time finding out: --enable-hardcoded-tables --enable-avresample --enable-vda
[21:34:14 CEST] <llogan> smo_80: that output means nothing without any context
[21:34:21 CEST] <llogan> such as your actual command
[21:35:14 CEST] <smo_80> ffmpeg -i 'rtsp://......' -c:v libx264 -c:a copy -sn test.mkv
[21:35:20 CEST] <smo_80> nothing special llogan
[21:36:08 CEST] <smo_80> (ffmpeg running on a freebox 4k, set top box in france)
[21:36:22 CEST] <smo_80> android
[21:36:33 CEST] <c_14> podman: enable-avresample enables the avresampler in addition to swresample; enable-vda enables a hardware acceleration subsystem (for OSX afaik)
[21:36:58 CEST] <podman> c_14: gotcha. any idea what enable-hardcoded-tables does?
[21:37:44 CEST] <gouessej> "Past duration 0.647652 too large" :(
[21:38:07 CEST] <c_14> precomputes certain tables and writes them into the binary instead of computing them on the fly (iirc)
[21:38:50 CEST] <podman> c_14: i figured as much. I was just trying to figure out what that actually does with regard to regular usage of FFMPEG
[21:39:30 CEST] <smo_80> llogan, ffmpeg config http://paste.ubuntu.com/11691958/
[21:39:45 CEST] <kyleogrg> hello
[21:40:30 CEST] <kyleogrg> I'd like to stream-copy a VOB into an MKV, adding a cropping parameter (10,0,0,0).
[21:41:08 CEST] <kyleogrg> How do I set an mkv's cropping parameter in ffmpeg?
[21:41:14 CEST] <gouessej> I'll try to solve my last problem alone. Thank you for your help. Goodbye
[21:41:47 CEST] <kyleogrg> I am not encoding.  This is a cropping parameter for the video player.
[21:45:09 CEST] <smo_80> hi c_14 any idears on this type of errors ? http://paste.ubuntu.com/11691992/
[21:47:10 CEST] <Perkol> Hello. Where can i get libavcodec56 for ubuntu 14.04?
[21:47:11 CEST] <c_14> Looks like a mix between corrupt build and corrupted packets.
[21:48:39 CEST] <smo_80> ok for the build but it comes from slow hardware or slow connexion ?
[21:49:02 CEST] <smo_80> hard to find good android buils :/
[21:49:10 CEST] <smo_80> i m compiling it now
[21:49:48 CEST] <Perkol> Hello. Where can i get libavcodec56 for ubuntu 14.04? It's required in deps for dev builds
[22:16:58 CEST] <kyleogrg> how can i encode variable frame rate with libx264?
[22:19:29 CEST] <kyleogrg> with -vsync vfr?
[22:22:16 CEST] <JEEBsv> a lot of lavf/lavc should support VFR input by default
[22:22:34 CEST] <JEEBsv> it doesn't try to fit the output to any frame rate unless you specifically set one
[22:22:48 CEST] <JEEBsv> it takes the input timestamps, and fits them to the timescale of the output format
[22:23:53 CEST] <kyleogrg> ok
[22:28:34 CEST] <kyleogrg> if i stream-copy VOB to MKV in ffmpeg, will the MKV have the chapters info?
[22:29:34 CEST] <kyleogrg> or if i convert it
[22:30:10 CEST] <Compn> dont stream copy vob to mkv
[22:30:17 CEST] <Compn> just ... no
[22:30:25 CEST] <Compn> probably you want to use makemkv to do chapter encodings
[22:31:25 CEST] <kyleogrg> yeah.  but isn't makemkv copying the vob/mpg stream?
[22:32:09 CEST] <Compn> the number of video players that handle vob in mkv will be slim, and the number of them supporting the dvd chapters inside the mkv are none.
[22:32:41 CEST] <Compn> ignore my advice (10+ years of multimedia formats) at your own peril.
[22:32:54 CEST] <kyleogrg> okay
[22:33:10 CEST] <kyleogrg> now, vlc seems to show the chapters
[22:33:29 CEST] <kyleogrg> and i just want it to work for vlc
[22:34:05 CEST] <kyleogrg> anyhow, i have a different question
[22:34:35 CEST] <kyleogrg> supposing i want to always play this video with a crop setting
[22:34:50 CEST] <kyleogrg> can i tell vlc to always crop that video
[22:35:59 CEST] <Compn> crop setting in mkv maybe
[22:36:02 CEST] <Compn> i dont know how to set that :)
[22:36:06 CEST] <JEEBsv> lol
[22:36:13 CEST] <kyleogrg> haha
[22:36:13 CEST] <JEEBsv> I kind of waited for someone to mention that
[22:36:23 CEST] <JEEBsv> because that's one of the things you MUST NOT USE
[22:36:23 CEST] <kyleogrg> Just went over that with JEEBsv
[22:36:31 CEST] <JEEBsv> because it's not specified
[22:36:33 CEST] <kyleogrg> he says it's not often supported
[22:36:49 CEST] <JEEBsv> not supported is not even the main pain point
[22:36:52 CEST] <Compn> h264 has crop settings...
[22:36:55 CEST] <JEEBsv> yes
[22:37:08 CEST] <JEEBsv> and the matroska field is made to override those
[22:37:13 CEST] <Compn> of course
[22:37:17 CEST] <JEEBsv> literally for the demuxer to override crop values
[22:37:27 CEST] <JEEBsv> literally to overwrite parameter set values
[22:37:41 CEST] <kyleogrg> hmm, but if I'm *encoding* to x264, couldn't I actually crop the video stream?
[22:37:44 CEST] <JEEBsv> that was the original implementation :D
[22:37:49 CEST] <JEEBsv> yes
[22:37:50 CEST] <JEEBsv> of course
[22:37:59 CEST] <Compn> kyleogrg : see, in mplayer, you just create a filename config , like mymovie.avi.conf and then put vf=crop=cropsetitings
[22:38:06 CEST] <Compn> but this is vlc we are talking about
[22:38:12 CEST] <Compn> and who knows about that dumb program :D
[22:38:16 CEST] <Compn> ask in #videolan
[22:38:20 CEST] <kyleogrg> hmm, never really used mplayer
[22:38:29 CEST] <Compn> dont use mplayer, its eeeevil
[22:38:39 CEST] <kyleogrg> you make it sound awesome
[22:38:49 CEST] <JEEBsv> I wonder if mpv kept that feature
[22:38:56 CEST] <Compn> mpv hates everything
[22:39:01 CEST] <JEEBsv> no, no it doesn't
[22:39:04 CEST] <kyleogrg> I bet you can put other settings in the config file as well
[22:39:11 CEST] <Compn> other settings, bah!
[22:39:21 CEST] <Compn> you can even have per-directory settings :P
[22:39:24 CEST] <Compn> i think
[22:39:32 CEST] <kyleogrg> wow
[22:39:33 CEST] <JEEBsv> mpv does remove a helluva lot of fringe features though :P
[22:40:00 CEST] <JEEBsv> and I consider with first look that random config file reading is kind of such a thing
[22:40:51 CEST] <kyleogrg> http://www.mplayerhq.hu/DOCS/man/en/mplayer.1.html#CONFIGURATION%20FILES
[22:40:53 CEST] <kyleogrg> :D
[22:41:34 CEST] <JEEBsv> http://mpv.io/manual/master/#configuration-files
[22:41:42 CEST] <JEEBsv> yeah, file specific stuff was kept
[22:41:46 CEST] <JEEBsv> I'm kind of surprised
[22:43:33 CEST] <Compn> i'm not, anime people like diff configs
[22:45:19 CEST] <podman> did ffmpeg start automatically rotating videos?
[22:45:23 CEST] <JEEBsv> kyleogrg: anyways if you're really perverted feel free to go through 14496-12 (iso base media file format aka "mp4"), I'm pretty sure it can do random croppin
[22:45:26 CEST] <JEEBsv> *cropping
[22:46:08 CEST] <c_14> podman: judging by the changelog, yes
[22:46:09 CEST] <Compn> podman : yes
[22:46:21 CEST] <podman> c_14: just saw that
[22:46:34 CEST] <podman> That's a "problem" but also kind of neat
[22:46:52 CEST] <c_14> You should be able to turn it off (hopefully).
[22:47:03 CEST] <Compn> why? are your rotation tags incorrect in your mp4 files? :P
[22:47:04 CEST] <c_14> -noautoratet
[22:47:13 CEST] <c_14> *rotate
[22:47:24 CEST] <podman> Compn: i was already rotating, so it's getting rotated twice
[22:47:33 CEST] <podman> c_14: awesome, thanks
[22:47:36 CEST] <Compn> ehe
[22:47:47 CEST] <podman> i imagine a lot of people have code for that
[22:47:57 CEST] <Compn> probablyt
[22:48:14 CEST] <c_14> hmm, would it be possible for the autorotate to detect if a rotate filter is currently in the filterchain and then not rotate?
[22:48:23 CEST] <kyleogrg> interesting
[22:48:41 CEST] <kyleogrg> anyway, i'm going to look at SMplayer
[22:48:54 CEST] <podman> I think I'll just rely on ffmpeg to rotate stuff correctly and pull out my rotation code
[22:49:00 CEST] <kyleogrg> it supposedly remembers all settings for every file you open
[22:49:28 CEST] <Compn> smplayer does do that
[22:49:40 CEST] <Compn> also remembers where in the film you are when you quit, so it can resume later
[22:50:27 CEST] <kyleogrg> cool
[22:50:37 CEST] <kyleogrg> i'm trying to figure out how to crop
[22:53:57 CEST] <kyleogrg> do you happen to know?
[23:07:48 CEST] <kyleogrg> got it, never mind
[00:00:00 CEST] --- Thu Jun 11 2015



More information about the Ffmpeg-devel-irc mailing list