[Ffmpeg-devel-irc] ffmpeg.log.20160502

[burek]

[01:55:01 CEST] <Guest62857> word
[02:02:23 CEST] <datam0sh> im on ffmpeg 3.0 I'm trying to find a setting to encode a mp4 video that will work on windows and mac/ios with a HTML5 emebd
[02:03:55 CEST] <andrew_46> I have a problem accessing ffmpeg from git today: fatal: read error: Connection reset by peer
[02:04:15 CEST] <furq> datam0sh: what settings do you have which aren't working
[02:04:35 CEST] <andrew_46> Using
[02:04:37 CEST] <andrew_46> git://source.ffmpeg.org/ffmpeg.git
[02:04:47 CEST] <andrew_46> Just me?
[02:05:25 CEST] <furq> not working here either
[02:05:50 CEST] <datam0sh> furq should i post the ffmpeg string here?
[02:05:55 CEST] <furq> pastebin it
[02:06:02 CEST] <furq> and the output as well if possible
[02:06:24 CEST] <datam0sh> http://pastebin.com/EQaKXkQ8
[02:06:44 CEST] <furq> add -pix_fmt yuv420p
[02:07:06 CEST] <datam0sh> it encodes successfully just fine , but this string only works on Windows/Android - if i swap -vcodec libx264 for mpeg4 - then it works on ios/mac but not windows
[02:07:11 CEST] <furq> ^
[02:08:38 CEST] <datam0sh> k trying it now
[02:13:44 CEST] <furq> you wouldn't normally need that but you're converting from rgb24 so i assume it's defaulting to yuv444p
[02:13:54 CEST] <furq> which won't play in browsers (except maybe chrome)
[02:14:55 CEST] <andrew_46> furq: Thanks for confirming
[02:15:14 CEST] <furq> andrew_46: i normally just use the snapshot anyway
[02:16:46 CEST] <datam0sh> it did play in chrome well - encoding it now with  -pix_fmt yuv420p
[02:17:33 CEST] <furq> also don't use mpeg4, it's 2016
[02:19:51 CEST] <c_14> But what about the good old xvid days
[02:20:14 CEST] <furq> ;_;
[02:20:30 CEST] <furq> why can't we just go back to selectavision
[02:21:45 CEST] <datam0sh2> thanks furq - it worked! im in heaven
[02:21:57 CEST] <furq> i have that effect on people
[02:24:19 CEST] <datam0sh2> made me a believer
[03:21:41 CEST] <klaxa> the "correct" way to do html5 video is to use webm (with vp9/opus or vp8/vorbis)
[03:21:53 CEST] <klaxa> the correctest way to do html5 video is to not do html5 video
[03:22:21 CEST] <furq> is that really more correct than mp4
[03:23:18 CEST] <klaxa> i tried to make a html5 video viewer yesterday but it only supports those (on linux with firefox and chromium)
[03:42:44 CEST] <Prelude2004c> hey guys
[03:42:59 CEST] <Prelude2004c> anyone know what this means.. i am getting my screen scrolling with this " new max_dts_to_pts:7106600 "
[03:43:05 CEST] <Prelude2004c> its  a satelite source.
[04:34:07 CEST] <grublet> furq: lets go back to mechanical televisions and celluloid
[04:35:03 CEST] <grublet> but yeah when in doubt -pix_fmt yuv420p ends up being something i throw on the command line out of habit
[04:35:30 CEST] <grublet> shame higher bit depth got uptook faster than beter chroma
[04:36:06 CEST] <furq> chrome supports 4:4:4 now
[04:36:15 CEST] <furq> hopefully firefox will follow suit
[04:36:33 CEST] <furq> then that's all the browsers which aren't massive deliberate bastards covered
[05:39:41 CEST] <grublet> im still on palemoon 25.6
[10:52:01 CEST] <Fyr> guys, is XAVC a format invented only for Samsung cameras and it has nothing to with PC?
[10:52:29 CEST] <Fyr> or XAVS is used when encoding/decoding video and it's somekind of a codec/container?
[10:53:04 CEST] <Fyr> pdeiwikia says nothing.
[10:58:21 CEST] <relaxed> from wikipedia it looks like the name of their h264 encoder
[11:01:31 CEST] <relaxed> well, more of a format like AVCHD
[11:01:42 CEST] <Fyr> >>XAVC is a recording format that was introduced by Sony on October 30, 2012. XAVC is a format that will be licensed to companies that want to make XAVC products
[11:01:53 CEST] <Fyr> so, it's their H264 encoder?
[11:02:07 CEST] <relaxed> but I would guess ffmpeg can decode it
[14:04:49 CEST] <Carlrobertoh> Hi ! I'm tring to code a screen casting program in ffmpeg using c++. The raw h264 byte stream can be played by VLC, but if i'm trying to save the file as .mp4 then it wont be played. Any suggestions ?
[14:05:13 CEST] <Carlrobertoh> "trying to encode the file as .mp4"
[14:05:28 CEST] <Fyr> and not remux?
[14:05:44 CEST] <Fyr> encoding is used for codecs, muxing is used for containers.
[14:06:43 CEST] <Carlrobertoh> const char* filename = "test.h264";     output_format_ctx->oformat = av_guess_format( NULL, filename, NULL );
[14:06:45 CEST] <Carlrobertoh> works fine
[14:06:57 CEST] <Carlrobertoh> but if i change filename to "test.mp4"
[14:06:59 CEST] <Carlrobertoh> then it wont work
[14:07:29 CEST] <Carlrobertoh> i've found something about annex b
[14:07:57 CEST] <Carlrobertoh> that i need to change some NALU header
[14:59:12 CEST] <_prcreeper_> Hello
[14:59:40 CEST] <_prcreeper_> I don`t understood - how to stream via RTSP using ffserver
[14:59:43 CEST] <_prcreeper_> I don`t understood - how to stream via RTSP using ffserver?
[15:01:21 CEST] <_prcreeper_> In FAQ and man this is very unclear
[15:07:43 CEST] <_prcreeper_> Any body there?
[15:47:16 CEST] <Prelude2004c> hello
[15:47:29 CEST] <Prelude2004c> anyone know what this means.. i am getting my screen scrolling with this " new max_dts_to_pts:7106600 "  how do i get rid of the max_dts_pts errors ?
[15:57:35 CEST] Action: penguin42 is trying to convert an rtsp stream (that seems to be h264) from a cctv cam to something a web browser will play but trying not to use too much CPU - currently I'm going to .mp4  and using -f segment   to spit out few minute long chunks;  my only problem is that it wont play the video until the end fo the segment
[15:57:45 CEST] <penguin42> I was looking for some help as to if -movflags can fix that
[15:58:20 CEST] <c_14> I'm pretty sure most web browsers won't play fragmented mov, so probably nothing you can do there (besides making the segments shorter...)
[15:58:29 CEST] <penguin42> ah
[15:58:32 CEST] <c_14> you can try -movflags +frament
[15:58:38 CEST] <c_14> *fragment
[15:59:01 CEST] <c_14> eh
[15:59:16 CEST] <c_14> -movflags frag_keyfram or -movflags frag_size or something
[16:00:15 CEST] <penguin42> yeh I'd played with that but without success
[16:00:31 CEST] <penguin42> any other alternatives to mp4 that would be better, or is that the best?
[16:01:10 CEST] <c_14> Well, if you want a web browser to play it your only choice is mp4 or webm and if you chose webm you'll have to reencode to vp9 or vp8 which takes a decent amount of cpu
[16:01:14 CEST] <c_14> and is rather slow
[16:01:22 CEST] <penguin42> yeh, that was my previous attempt
[16:02:16 CEST] <penguin42> hth designed mp4 without allowing it for streaming....
[16:04:48 CEST] <roxlu> Hey! Can I use ffmpeg as a tool to accept live rtmp data and stream it to a client? (or maybe one of it's utils)
[16:06:24 CEST] <penguin42> c_14: Thanks
[16:06:54 CEST] <_prcreeper_> Hello
[16:06:59 CEST] <_prcreeper_> I don`t understood - how to stream via RTSP using ffserver?
[16:07:09 CEST] <_prcreeper_> In FAQ and man this question is very unclear
[16:07:26 CEST] <JEEB> ffserver is black magic known by few
[16:07:36 CEST] <JEEB> and unless someone steps up to maintain ffserver, it will get removed
[16:07:49 CEST] <JEEB> look into other streaming servers if you actually require such
[16:08:12 CEST] <_prcreeper_> What?
[16:08:54 CEST] <roxlu> JEEB: yeah, I wanted to quickly test a rtmp stream I create; would be nice to have a simple server that doesn't need lots of setup.
[16:10:21 CEST] <navlelo_> _prcreeper_: look at the dummy config file for ffserver
[16:10:41 CEST] <Prelude2004c> C_14 .. > any idea on  i am getting my screen scrolling with this " new max_dts_to_pts:7106600 "  how do i get rid of the max_dts_pts errors ? ... not sure what to do
[16:11:02 CEST] <navlelo_> _prcreeper_: it shows how to connect and what to set up... it is a little bit of configuration that is needed to make stuff work
[16:17:16 CEST] <Prelude2004c> or anyone else?
[16:20:28 CEST] <c_14> Prelude2004c: I can't even find that string anywhere in the source (or any subpart thereof)
[16:20:44 CEST] <_prcreeper_> Ok, i tries to set the rtsp port in ffsever`s config file, but how to prevent ffmpeg to use the rtsp?
[16:21:43 CEST] <navlelo_> _prcreeper_: what do you mean by that?
[16:22:28 CEST] <_prcreeper_> How to stream to an rtsp feed via ffmpeg
[16:22:43 CEST] <_prcreeper_> Sorry for my English
[16:22:54 CEST] <Prelude2004c> really?
[16:23:24 CEST] <Prelude2004c> that is odd...  i am def. getting that error.. just not sure hwo to deal with it
[16:24:42 CEST] <c_14> Can you pastebin the complete output?
[16:25:25 CEST] <navlelo_> _prcreeper_: the ffserver docs + the dummy config will explain how it works. Basically ffmpeg does the streaming and outputs the stream into ffserver which serves the stream on a configured port
[16:25:35 CEST] <_prcreeper_> Output of what?
[16:25:56 CEST] <navlelo_> the stream itself
[16:27:06 CEST] <_prcreeper_> Of ffmpeg or ffserver?
[16:29:01 CEST] <navlelo_> you call ffmpeg like so: ./ffmpeg -i rtsp://cameraip/stream.sdp http://ffserverip:port/feed1.ffm
[16:29:15 CEST] Action: penguin42 wonders how short a segment size on the segment mux makes sense
[16:29:35 CEST] <_prcreeper_> ffmpeg -i /dev/video0 rtsp://localhost:5039/feed.ffm
[16:30:07 CEST] <navlelo_> that seems reasonable
[16:30:43 CEST] <navlelo_> im off now... good luck :)
[16:32:05 CEST] <_prcreeper_> Pastebined output here http://pastebin.com/RT3JXQ0y
[16:34:02 CEST] <_prcreeper_> rtsp://localhost:5039/feed.ffm: Protocol not found
[16:45:38 CEST] <_prcreeper_> And how to stream screen via ffmpeg?
[16:59:52 CEST] <Prelude2004c> c_14 sure sorry for the delay
[17:01:08 CEST] <_prcreeper_> Who is now how to capture and stream screen via ffmpeg?
[17:18:42 CEST] <andrey_utkin> is there an option to drop SAR/DAR info from input file? i have some mp4 files with it, and some without. Of course setting same SAR everywhere is ok, but i wonder can i go discarding SAR info.
[17:20:48 CEST] <kepstin> andrey_utkin: if you're re-encoding the video, you can use the 'setsar' video filter to change it to whatever you want
[17:21:11 CEST] <kepstin> you can't really "discard it", but setting it to 1:1 (square pixels) is what you probably want to do?
[17:21:20 CEST] <andrey_utkin> kepstin, yes i know
[17:24:36 CEST] <kepstin> if you want to modify an mp4 file without re-encoding, maybe one of the MP4Box tools can do it?
[17:25:38 CEST] <andrey_utkin> kepstin, what i am talking about as "discarding" it, is that file (h264-in-mp4) MAY carry explicit SAR info which is shown e.g. in ffprobe at elementary stream info
[17:26:04 CEST] <andrey_utkin> no, i'm actually reencoding
[17:26:42 CEST] <kepstin> I suspect that ffmpeg always explicitly sets the sar/par metadata fields when encoding.
[17:26:58 CEST] <kepstin> I checked the MP4Box docs, it does have support for changing or removing that
[17:32:59 CEST] <andrey_utkin> kepstin, in fact no, unless explicitly asked, ffmpeg does not populate sar, at last in case of mp4 container
[17:33:39 CEST] <andrey_utkin> or unless this info is set in input file which is undergoing transcoding
[17:39:54 CEST] <tokam> I would like to create a video by a photo and a mp3 file.
[17:40:09 CEST] <tokam> which command must I use to have the photo displayed as part of the video for the whole length?
[17:40:35 CEST] <Prelude2004c> hey guys.. i found more info " cur_dts is invalid " .. that is what i am getting on start
[17:40:40 CEST] <tokam> is it -vframes 1?
[17:41:32 CEST] <furq> tokam: -loop 1 -i in.png -i in.mp3 -shortest out.mkv
[17:42:21 CEST] <tokam> what does -threads auto do
[17:42:26 CEST] <tokam> and -vframes 1
[17:43:03 CEST] <furq> -vframes 1 writes one frame to the output file
[17:43:15 CEST] <tokam> -shortest?
[17:43:27 CEST] <furq> that command will loop forever unless you pass -shortest
[17:43:32 CEST] <tokam> ok thanks
[17:43:33 CEST] <furq> in that instance it'll stop processing when the mp3 ends
[17:44:08 CEST] <tokam> is avi or mpeg a better format?
[17:44:13 CEST] <tokam> I like to upload it on facebook.
[17:44:18 CEST] <furq> use mp4
[17:44:30 CEST] <tokam> what is -loop 1 than? I assumed that this is the looping, but you said it is shortest?
[17:44:46 CEST] <tokam> [aac @ 0x20ee0c0] The encoder 'aac' is experimental but experimental codecs are not enabled, add '-strict -2' if you want to use it.
[17:44:49 CEST] <furq> -loop loops the next input stream forever
[17:44:49 CEST] <tokam> that's what I get for mp4
[17:45:05 CEST] <furq> if you didn't pass -shortest then that command would run forever because one of your inputs is looping forever
[17:45:11 CEST] <Mista_D> Any idea on EAC3 decoding please? http://pastebin.ca/3588795
[17:45:15 CEST] <furq> and yeah just add -strict -2
[17:45:20 CEST] <furq> or upgrade to a newer ffmpeg
[17:45:35 CEST] <tokam> ok thanks
[17:45:45 CEST] <tokam> mp4 seems to encode quicker than avi
[17:46:13 CEST] <tokam> but it takes a while also. maybe it is stupid to use an 1,2 MB high resolution image
[17:46:42 CEST] <tokam> the file size is only 600kb yet but it only compiled it up to 5 seconds in that entire time
[17:47:44 CEST] <tokam> only 9 seconds done yet, is there a way to speed that up?
[17:50:52 CEST] <tokam> furq: thanks
[18:50:25 CEST] <thebombzen> w00t the bug report I submitted got fixed :D
[18:50:58 CEST] <thebombzen> makes you love ffmpeg devs. submit crash bug report, fixed two days later. pretty good
[18:56:47 CEST] <thebombzen> hey, I'm trying to figure out how to use -vf scale to force yuv420p. it appears I have to use flags=dst_format=yuv420p but I don't know the integer id for yuv420p. does anyone know (and/or have a better way of doing this)?
[18:57:33 CEST] <vade> do all packets that go to a muxer have to have increasing dts timestamps, or only packets destined to go to a particular STREAM associated with a muxer need monotonically increases DTS relative to that stream?
[18:57:34 CEST] <vade> I ask because im having an issue where the first frame of video out to an MP4 encoded in h.264 is complaiing that its DTS is less than the LAST audio samples I wrote to the output audio stream
[18:57:36 CEST] <pfelt> -pix_fmt ?
[18:57:43 CEST] <pfelt> thebombzen: ^
[18:57:44 CEST] <vade> and I am using av_interleaved_write_frame too :X
[19:14:51 CEST] <cybrNaut> ffmpeg -vtag xvid -i input.avi -vf 'scale=720:-1' output.avi  <= why does that produce an "FMP4" video, and not an "XVID"?
[19:16:11 CEST] <cybrNaut> FaQ says "-vtag xvid" is all that's needed, if I understand correctly
[19:19:59 CEST] <vade> gah
[19:21:42 CEST] <DHE> cybrNaut: order of parameters matters. put it before the output file, not the input file
[19:22:01 CEST] <cybrNaut> DHE: thanks!  will try that..
[19:22:07 CEST] <DHE> ffmpeg [input1-opts] -i input1 [...]   [output1-opts] output1 [...]
[19:23:14 CEST] <cybrNaut> ah
[19:24:18 CEST] <cybrNaut> i do think it's bizarre that it doesn't automatically give me the same format as the input file (the input file is already Xvid)
[19:25:43 CEST] <cybrNaut> ffprobe says it's correct now
[19:26:32 CEST] <DHE> I haven't tried it that way, but does it work right? it says the codec is xvid when playing it back and not just that it's labeled xvid?
[19:26:44 CEST] <DHE> I normally use the full "-c:v xvid" process
[19:27:43 CEST] <timbucii> Hi, What is the command to change the input of my tv card from antenna to composite?
[19:27:44 CEST] <cybrNaut> DHE: how can I do a real test?  vlc "media info" also says it's XviD, but presumably that's still just the metadata
[19:28:14 CEST] <kepstin> cybrNaut: ffmpeg doesn't make any attempt to use the same codecs as the input file. If unspecified, the output codecs are automatically picked only by looking at the output format.
[19:28:28 CEST] <DHE> would that include the vtag option?
[19:29:01 CEST] <DHE> cybrNaut: ffprobe will do it, but you need to read the full output
[19:29:14 CEST] <DHE> also note that you didn't set a video bitrate etc, so you get defaults.
[19:32:50 CEST] <cybrNaut> Stream #0:0: Video: mpeg4 (Simple Profile) (xvid / 0x64697678), yuv420p, 720x405 [SAR 1:1 DAR 16:9], 30 tbr, 30 tbn, 30 tbc <= i suppose that's the relevant part?
[19:33:02 CEST] <cybrNaut> the metadata section only says "encoder         : Lavf54.29.104"
[19:33:39 CEST] <cybrNaut> anyway, the result is blocky.
[19:34:06 CEST] <cybrNaut> i shrunk the dimensions to 720x-1 (to preserve aspect ratio)
[19:34:26 CEST] <cybrNaut> when I play the shrunk version in its native size, it should not be blocky
[19:34:46 CEST] <furq> isn't xvid always advanced simple profile
[19:34:53 CEST] <furq> or is that just ffprobe not distinguishing
[19:36:06 CEST] <cybrNaut> furq: no idea
[19:36:34 CEST] <furq> it's not ffprobe, i just checked
[19:36:50 CEST] <furq> it's been too long since i used xvid to remember if it encodes SP
[19:37:03 CEST] <furq> but it could be that someone's just written the wrong fourcc to your source
[19:37:39 CEST] <cybrNaut> ah, i wonder if that's it.. so the source file might have untrustworthy metadata?
[19:37:50 CEST] <furq> maybe
[19:38:51 CEST] <cybrNaut> The file came from a school teacher, and her classrooms video play complains that the dimensions of her video file are too big, with a max of 720x548
[19:39:28 CEST] <cybrNaut> so I'm merely trying to shrink it to work w/out doing too much transcoding, and without knowing the model of her player
[19:39:50 CEST] <furq> oh never mind
[19:42:39 CEST] <cybrNaut> i guess the blockyness won't do much harm.. it's just a short video she wants to show and sharpness probably doesn't matter
[19:42:58 CEST] <furq> http://sprunge.us/EWZj
[19:43:01 CEST] <furq> looks like i'm wrong anyway
[19:43:14 CEST] <cybrNaut> i'd just like to know what went wrong, in case I need to do this again
[19:44:32 CEST] <furq> oh hang you're reencoding
[19:44:54 CEST] <furq> the input codec doesn't affect the output codec selection, it'll use the default for the container if you don't specify
[19:45:02 CEST] <furq> which for avi is the internal mpeg4 encoder
[19:45:10 CEST] <furq> you'll need to specify -c:v libxvid
[19:45:19 CEST] <cybrNaut> furq: right, that's what kepstin mentioned
[19:45:29 CEST] <cybrNaut> ah, ok, will try that then
[19:45:29 CEST] <furq> oh so he did
[19:46:01 CEST] <cybrNaut> furq: DHE said he uses "-c:v xvid"
[19:46:09 CEST] <furq> either should work i think
[19:46:15 CEST] <cybrNaut> k
[19:46:24 CEST] <furq> but yeah if you were remuxing to the same container then afaik it would keep the same fourcc
[19:47:26 CEST] <DHE> cybrNaut: okay, I don't actually use xvid specifically anymore, so I may have gotten the exact codec name wrong
[19:47:32 CEST] <kepstin> by "remuxing", you mean -c:v copy? I'd expect it to pick a new fourcc based on the codec if it's reencoding
[19:47:43 CEST] <furq> yeah exactly
[19:49:07 CEST] <cybrNaut> i'm getting errors with the -c:v option.  "[libxvid @ 0xbc2a60] Xvid: Could not create encoder reference"
[19:49:21 CEST] <cybrNaut> and "Error while opening encoder for output stream #0:0 - maybe incorrect parameters such as bit_rate, rate, width or height"
[19:50:04 CEST] <furq> use -vf scale=720:-2
[19:50:19 CEST] <furq> -2 is the same as -1 but it'll ensure the height is mod2
[19:50:31 CEST] <thebombzen> pfelt: I'm trying to do it with -vf scale because of a filtergraph
[19:50:44 CEST] <thebombzen> furq: do you know the swscale ID for yuv420p?
[19:50:52 CEST] <furq> no idea
[19:51:28 CEST] <cybrNaut> furq: "Size values less than -1 are not acceptable." (maybe the debian stable version is too old)
[19:51:35 CEST] <furq> oh
[19:51:36 CEST] <furq> maybe
[19:51:43 CEST] <furq> i think jessie-backports has 2.8.6
[19:51:56 CEST] <furq> regular jessie is still using libav
[19:51:57 CEST] <cybrNaut> ffmpeg version 1.0.10
[19:52:01 CEST] <furq> yeah that's libav
[19:52:17 CEST] <furq> er, no it isn't
[19:52:20 CEST] <furq> it's super old either way
[19:52:54 CEST] <furq> ffmpeg from backports should be fine
[19:53:17 CEST] <cybrNaut> ok, i'll try upgrading to a backport
[19:55:13 CEST] <cybrNaut> i hate to get dirty like this.. this will be the first backport i use in jessie
[19:55:49 CEST] <furq> mixing stable and backports is officially sanctioned
[19:56:10 CEST] <furq> i'd love to tell you it's never broken anything for me but i only ever run testing
[19:59:01 CEST] <vade> Im a big confused about writing audio and video to the same muxer - to be clear of some assumptions: I use the same AVOutputFormat and set the audio and video codec, and init an AVCodec* from that - and make an AVFormatContext from my outputFormat, and then attach new AVStreams* and configure them as appropriate via avformat_new_stream(outputFormatContext, codecForStream). I configure the stream and the streams codec, and the avcodec_open2 them
[19:59:02 CEST] <vade> and then avio_open my outputFormatXontext, and finally write the header ?
[19:59:41 CEST] <vade> once I do that - I should be able to tell my muxer to send a write a packet - and based on the packet stream index, it goes to the write stream?
[19:59:46 CEST] <cybrNaut> aptitude -t jessie-backports upgrade ffmpeg <= is that the command?  or should I /remove/ followed by /install/?
[20:00:22 CEST] <furq> probably better to purge it first so you don't have any old bullshit lying around
[20:05:47 CEST] <cybrNaut> so it removed about 2mb of files, and installing the backports version adds 25mb in files
[20:05:59 CEST] <cybrNaut> my have you grown, ffmpeg
[20:07:47 CEST] <timbucii> Booya! 2016-03-20 12:10:24 :P
[20:07:54 CEST] <cybrNaut> now when I do the conversion, it goes to work instead of failing with errors immediately
[20:08:19 CEST] <cybrNaut> but i get lots of warnings "Invalid pixel aspect ratio 406/405, limit is 255/255 reducing"
[20:09:58 CEST] <cybrNaut> btw, why is "-vtag xvid" inferior to the "-c:v libxvid" option?
[20:10:05 CEST] <cybrNaut> should I use both, perhaps?
[20:11:09 CEST] <furq> one is the fourcc, one is the encoder
[20:11:16 CEST] <furq> the latter ought to imply the former
[20:12:32 CEST] <cybrNaut> it's still blocky.  I used: ffmpeg -i input.avi -c:v libxvid -vf 'scale=720:-2' output.avi
[20:14:38 CEST] <kepstin> it's probably blocky because you didn't set any quality settings...
[20:14:51 CEST] <kepstin> you'll probably want to set a bitrate (and with xvid, maybe even do a 2-pass encode)
[20:19:12 CEST] <MrSassyPants> how do I enable scene cut keyframes when using libvpx / vp9 ?
[20:24:54 CEST] <cybrNaut> kepstin: thanks.. that made a big difference.  I used these options => -c:v libxvid -q:v 10 -vf 'scale=720:-2'
[20:25:09 CEST] <cybrNaut> i'll try a 2 pass approach next, using => -c:v libxvid -q:v 10 -vf 'scale=720:-2' -pass 2
[20:25:47 CEST] <cybrNaut> Error reading log file 'ffmpeg2pass-0.log' for pass-2 encoding
[20:25:58 CEST] <kepstin> using -q:v with 2-pass mode doesn't make sense
[20:26:11 CEST] <kepstin> only use it if you're using bitrate (-b:v)
[20:26:25 CEST] <cybrNaut> ah, does the qscale incorporate number of passes?
[20:26:37 CEST] <cybrNaut> that's what I get for trying to mix different examples
[20:26:44 CEST] <furq> you'd need to run -pass 1 first
[20:26:52 CEST] <furq> -pass 2 is second pass, not two passes
[20:27:15 CEST] <furq> and yeah it only really makes sense if you're targeting a bitrate
[20:27:54 CEST] <vade> ah my bad. the AVPacket filled by the encoder doesnt keep the stream_index populated
[20:27:57 CEST] <cybrNaut> well, i'll keep with the simpler approach now, which I think qscale gives
[20:28:03 CEST] <vade> it was getting set to zero behind my back
[20:28:10 CEST] <furq> are you sure that the target player doesn't support h.264
[20:28:14 CEST] <furq> that'd make your life a bit easier
[20:28:42 CEST] <cybrNaut> furq: no idea.  I just know it's old enough to be limited to 720x resolution
[20:29:26 CEST] <cybrNaut> that has to be old.. i mean, even if the output is limited, the device should be able to scale as needed if the input is big
[20:40:29 CEST] <cybrNaut> furq: my phone supports h.264.  If I want to transcode a video for my phone, what codec option should I used to get h.264?  The ffmpeg default?
[20:40:45 CEST] <furq> the default settings for .mp4 are pretty ok
[20:41:00 CEST] <cybrNaut> cool
[20:42:03 CEST] <cybrNaut> i used the default settings and it came out blocky on my phone, but i probably needed to add a qscale, which I'm trying now
[20:42:21 CEST] <furq> don't use -q:v for x264
[20:42:23 CEST] <furq> use -crf
[20:42:36 CEST] <furq> the default is 23, lower values are higher quality
[20:43:26 CEST] <cybrNaut> ok, will try -crf
[20:44:31 CEST] <cybrNaut> -crf -vf 'scale=480:-2' <= those options give => Unable to find a suitable output format for 'scale=480:-2'
[20:44:41 CEST] <furq> you forgot the argument after -crf
[20:44:49 CEST] <cybrNaut> ah
[20:45:04 CEST] <cybrNaut> better
[20:50:23 CEST] <MrSassyPants> Question: Does 2-Passing do anything for constant-quality encodes?
[20:51:50 CEST] <DHE> with -qp ? no
[20:51:52 CEST] <DHE> or -crf
[20:52:22 CEST] <MrSassyPants> -crf
[20:53:01 CEST] <DHE> yeah, no use
[20:53:08 CEST] <DHE> 2-pass is for bitrate based outputs
[20:55:51 CEST] <MrSassyPants> ok. How to set scene cut keyframes in libvpx / vp9 ? I can't seem to find the right doc
[20:56:43 CEST] <BlackBishop> ffmpeg -i file.mkv -codec:s:0:24 srt file.srt ( trying to extract subtitle #24 ( Stream #0:17(kor): Subtitle: hdmv_pgs_subtitle, 1920x1080)) to a separate file ), anyone can tell me what I'm doing wrong ?
[20:59:17 CEST] <BlackBishop> or a way to hardcode it in the stream ?
[21:32:05 CEST] <DryEagle> hi, does ffmpeg work with wildcards? if i have a directory of 200 files all .aac and i want to wrap them to .m4a is there an easy way to do that?
[21:32:54 CEST] <furq> for f in *.aac; do ffmpeg -i "$f" -c copy "${f%.*}.m4a"; done
[21:34:53 CEST] <DryEagle> ok, that worked for some of the files, a lot of them returned this: http://pastebin.com/8qQmBgXK
[21:35:46 CEST] <hyponic> i have been trying to burn subtitles on a video following the wiki guide but when doing so the subtitles does not show. ffmpeg -i video.mkv -vf subtitles=video.mkv out.avi
[21:36:08 CEST] <hyponic> no errors indicating that it's not getting burned but still nothing on the video.
[21:36:25 CEST] <pfelt> hyponic: is the option on the subtitles filter supposed to be the file containing the subtitles?
[21:37:06 CEST] <Fyr> >>This filter requires ffmpeg to be compiled with --enable-libass.
[21:37:13 CEST] <pfelt> it'll have to be the right format too
[21:37:52 CEST] <hyponic> pfelt yes. i even tried ffmpeg -i video.avi -vf subtitles=subtitle.srt out.avi after extracting the .srt from the video.
[21:38:35 CEST] Action: pfelt would think that would work provided the library is correctly installed
[21:38:56 CEST] <hyponic> ffmpeg is compiled with: --enable-libass
[21:39:00 CEST] <vade> Hi all - im using libavformat and libavcodec to re-encode a video to AAC + h.264 mp4  - im having an issue where my video is timed perfectly but audio is WILDLY off - I imagine its my PTS handling - are there any examples of conversion between demuxer -> decoder -> encoder -> muxer correctly for both video and audio packets?
[21:39:19 CEST] <hyponic> pfelt should i install anything extra?
[21:39:20 CEST] <Prelude2004c> hey guys.. any way to take a bunch of input files.. shuffle them and play them one by one .. without using concat ? i don't want the stream to end and sence i am piping the output to the ffmpeg and segmenting .. the segments always start up at 0 after each file.... i want to have one continuous stream that never ends.. but at the end of the file list i would like it to sort again and keep going..
[21:39:49 CEST] <Prelude2004c> do i have to add 2 processes .. like 1 process output to UDP and the other take from UDP ? or can it somehow be done in the script where it can just stdout or somethign
[21:40:16 CEST] Action: pfelt is working up changes to the movie filter that would work if you have an external process to decide what to play when
[21:40:40 CEST] <pfelt> i couln't find any way to do something similar without coding it
[21:46:26 CEST] <Prelude2004c> pfelt , you talking to me about tht ?
[21:47:04 CEST] <pfelt> Prelude2004c: yep
[21:47:17 CEST] <pfelt> i've got a weird segfault atm
[21:47:45 CEST] <pfelt> but it switches files for me fairly well
[21:47:48 CEST] <Prelude2004c> oh ic... ya i was trying to find a way to code it to do that.. i mean.. we can take the list of files.. play them one after the other.. i am sure i twill work if i write it out to udp://locallhost:xxxx but i wanted to keep it all in the same file
[21:47:55 CEST] <hyponic> pfelt libass-dev is installed and the ffmpeg is compiled with --enable-libass and still nothing. here is the outout. http://pastebin.com/NUnVZaau
[21:48:00 CEST] <Prelude2004c> and it doesn't mess up the segmentation ?
[21:48:53 CEST] <pfelt> Prelude2004c: so my use case was to switch between live streams and output to a local decklink sdi card.  that all works pretty well.  i modified movie to have a command and i use zmq with it to tell it what file to read from
[21:49:17 CEST] <Prelude2004c> in realtime ?
[21:49:42 CEST] <pfelt> Prelude2004c: yep.  i'm reading in ffmpeg output through a named pipe on the filesystem and i can realtime swap between them
[21:49:59 CEST] <pfelt> i'm hoping to be able to clean up the code and submit back to main
[21:50:10 CEST] <Prelude2004c> ya i think if i run a udp listen on localhost and then just push things to that udp port i could kind of do the same thing no ?
[21:50:31 CEST] <Prelude2004c> i was trying to avoid doing that because i have to basically run 2 sperate processes.. and i wanted to do it all with stdin and stdout
[21:50:33 CEST] <pfelt> it didn't work for me when i tried that
[21:50:46 CEST] <pfelt> pts timestamps get all screwed up
[21:51:06 CEST] <pfelt> basically if i let the first stream play for 3 minutes, then i swapped, i had to wait 3 minutes before stuff would start showing up
[21:57:29 CEST] <pfelt> i can post the patch if you want it
[21:57:49 CEST] <pfelt> (i think my segfault is an av_freep() that i'm doing wrong
[22:59:53 CEST] <hyponic> Trying to burn subtitles and get this error: No usable fontconfig configuration file found, using fallback. i am on ubuntu and unable to find out where the fonts.conf should be etc..
[23:06:04 CEST] <c_14> /etc/fonts/fonts.conf
[23:06:20 CEST] <c_14> ubuntu should have one though...
[23:06:25 CEST] <c_14> Also, that's not an error
[23:06:27 CEST] <c_14> It's a warning
[23:09:40 CEST] <hyponic> c_14 warning.. but could it be the reason why the subs are not showing on the video?
[23:10:02 CEST] <c_14> possible, probable even
[23:10:11 CEST] <c_14> What does fc-list output?
[23:10:19 CEST] <c_14> does it output anything?
[23:10:45 CEST] <hyponic> c_14 The program 'fc-list' is currently not installed.
[23:11:42 CEST] <hyponic>  apt-get install fontconfig now it's showing some fonts
[23:12:58 CEST] <c_14> now try burning subtitles again
[23:14:15 CEST] <hyponic> c_14 [Parsed_subtitles_0 @ 0x40b7d60] No usable fontconfig configuration file found, using fallback. Fontconfig error: Cannot load default config file
[23:19:33 CEST] <c_14> hyponic: install fontconfig-config
[23:20:05 CEST] <hyponic> c_14 fontconfig-config is already the newest version.
[23:20:22 CEST] <c_14> Then you should have a fonts.conf
[23:20:36 CEST] <c_14> At the very least fontconfig-config is listed as the owner of the fonts.conf on the computer I just checked on
[23:22:35 CEST] <hyponic> c_14 i do have the fonts.conf but for some reason ffmpeg can't detect it
[23:32:56 CEST] <c_14> How did you get your build of ffmpeg?
[23:35:05 CEST] <hyponic> c_14 got it off some guy. static build or something
[23:39:39 CEST] Action: c_14 has no idea. It should work
[23:39:55 CEST] <hyponic> c_14 where should i get it from in ubuntu? i am unable to compile my own though.
[23:40:35 CEST] <c_14> Was it the static build from here http://johnvansickle.com/ffmpeg/ ?
[23:41:36 CEST] <hyponic> yes
[23:41:50 CEST] <c_14> then it should work
[00:00:00 CEST] --- Tue May  3 2016