[Ffmpeg-devel-irc] ffmpeg.log.20170226

[burek]

[00:00:27 CET] <klaxa> because my allocted outputs all have AVStream->codec->codec_id == AV_CODEC_ID_NONE
[00:00:34 CET] <klaxa> but i guess that doesn't matter then?
[00:00:56 CET] <klaxa> (AVStream->codec is AVCodecContext)
[00:01:22 CET] <klaxa> however when reading the AVStream->codec AVCodecContext has these fields populated with the correct values
[00:02:19 CET] <klaxa> and avformat_new_stream() wants const AVCodec *c
[00:02:33 CET] <klaxa> but how do i get the proper AVCodec then if it is nowhere set?
[00:02:46 CET] <klaxa> AVStream->codec->codec is NULL also for reading streams
[00:03:49 CET] <klaxa> i want to copy the streams, the remuxing.c example uses deprecated fields
[00:03:53 CET] <klaxa> i started working from there
[00:12:51 CET] <klaxa> hmm it seems all the codec parameters in the output file are set correctly, but no frames are written, well, seems like the mistake is somewhere else
[00:48:27 CET] <DHE> klaxa: we're in a grace period where the fields are merely deprecated. you should avoid using them and eliminate the compiler warnings in preparation for the version where they're gone entirely
[00:49:55 CET] <DHE> AVStream->codecpar still has most (if not all) of the basic identification fields you need including codec_id
[01:42:20 CET] <memeka> hi, are there any plans on adding v4l2 m2m to ffmpeg?
[01:43:16 CET] <JEEB> depends on if anyone cares about it. check the issue tracker (trac.ffmpeg.org) and add a feature request if there's none yet
[01:43:39 CET] <JEEB> basically people don't necessarily know or care, and the lack of knowledge can be fixed with a feature request
[02:47:50 CET] <TerNer> JEEB: Are you there?
[02:47:58 CET] <JEEB> never
[02:48:03 CET] <TerNer> :o
[02:48:45 CET] <TerNer> I have checked out fragmented-mp4, but there's also needed duration information in mdhd header
[02:50:14 CET] <JEEB> there's that kind of stuff for a *fragment* (will take a look at the exact box names in addition to moof), but since in a live situation you'd just be putting one fragment after another there really isn't a "duration" until you've written everything down, if you ever have
[02:51:00 CET] <JEEB> and fragments can be as long or short as you want, that really doesn't matter. they don't even technically have to start on a RAP
[02:51:56 CET] <TerNer> :o
[02:52:36 CET] <TerNer> can you tell me easier ;-; I'm not good at english...
[02:57:01 CET] <Aison> hello
[02:57:19 CET] <Aison> is it possible that ffmpeg has problems with decoding truehd 7.1?
[02:57:32 CET] <Aison> I just got this message: Dependent substream decoding is not implemented.
[03:00:27 CET] <JEEB> Aison: check with latest FFmpeg HEAD
[03:00:41 CET] <JEEB> if it still happens, report it with a sample if possible @ trac.ffmpeg.org
[03:02:21 CET] <Aison> hmm, another mkv file is not generating this warning, but the sound is still completely wrong
[03:02:38 CET] <Aison> I guess it is a problem with truehd decoder
[03:02:51 CET] <Aison> eac3 7.1 works
[03:02:58 CET] <Aison> k
[03:04:25 CET] <JEEB> mpeg-ts and matroska can generally be cut with something simple like dd
[03:04:33 CET] <JEEB> so you can try to cut the first 20MiB or so
[03:04:40 CET] <JEEB> and see if you can replicate
[03:04:44 CET] <JEEB> thus you can make a small sample
[03:04:48 CET] <Aison> ok
[03:08:15 CET] <TerNer> JEEB, I tested with increasing/decreasing duration but It's not affect at result video. thank you alot for help!
[03:33:01 CET] <Aison> ok, mplayer plays truehd correctly, vlc doesn't and so ffmpeg
[03:33:05 CET] <Aison> ffmpeg (git)
[03:33:14 CET] <Aison> so I create a sample tomorrow
[03:33:57 CET] <Aison> looks like vlc and ffmpeg are using the same decoder which is not working
[03:35:27 CET] <Aison> on the other hand, it looks like mplayer is also using avcodec
[03:35:28 CET] <Aison> hmm
[04:03:46 CET] <solrize> this is probably the wrong place to ask, but i'm trying to watch a video lecture and there's a lot of electrical buzzing (amplifier feedback or so) in the audio.  is there a decent way to get rid of that?  it might be 60 hz hum with a lot of harmonics i.e. i don't think i can take it out with a simple notch filter.  thanks
[04:04:25 CET] <solrize> actually i should probably ask on audacity instead of here.  just thought of that
[05:23:46 CET] <adeeln_> I'm starting out with libavcodec. Could anyone please provide a good starting point? Currently going through the codebase[https://github.com/FFmpeg/FFmpeg/search?utf8=%E2%9C%93&q=libavcodec]
[05:24:17 CET] <DHE> there's a number of sample applications in doc/examples/*.c have you checked them out?
[05:28:06 CET] <adeeln_> Yes, I did. Actually I was having some trouble using the data structure. I asked this question on SO[http://stackoverflow.com/questions/42455677/ffmpeg-swsfilter-structure-usage]. Could you please check it out?
[05:31:56 CET] <adeeln_> I also want to get started on a qualification task, any help would be appreciated!
[05:43:52 CET] <durandal_1707> adeeln_: that is very less used sws feature
[05:45:22 CET] <adeeln_> I was just tinkering with it. Thought it was fun.
[05:46:45 CET] <durandal_1707> adeeln_: you should better ask on mailing list
[05:47:00 CET] <adeeln_> Okay. I will, from now on.
[05:55:23 CET] <durandal_1707> adeeln_: on what you are working?
[05:56:46 CET] <adeeln_> On XMP decoding, atomnuker is guiding me.
[05:57:17 CET] <atomnuker> I got as far as take libavcodec/dpx.c, remove everything dpx related
[05:57:30 CET] <atomnuker> and add an entry to allcodecs.h and allcodecs.c
[05:57:41 CET] <atomnuker> (for AV_CODEC_ID_XMP)
[05:58:09 CET] <durandal_1707> ah, that. feel free to ask on ffmpeg-devel channel
[05:58:47 CET] <atomnuker> durandal_1707: can you continue (explain about libavcodec/Makefile, get data on the XMP, format, etc)?
[05:59:05 CET] <durandal_1707> yes, will try
[06:00:21 CET] <durandal_1707> adeeln_: you are free to ask questions any time
[06:00:35 CET] <adeeln_> Okay. Thank you.
[06:01:21 CET] <durandal_1707> right now i was sleeping though, what timezone you are?
[06:01:53 CET] <adeeln_> UTC +5
[06:03:01 CET] <durandal_1707> did you added AV_CODEC_ID_XPM to libavcodec/allcodecs.h?
[06:03:51 CET] <adeeln_> No, I can't find 'libavcodec/xpmdec.c'
[06:04:17 CET] <durandal_1707> adeeln_: you are supposed to create such file
[06:04:25 CET] <adeeln_> Oh, got it.
[06:04:49 CET] <durandal_1707> adeeln_: i think i sent basic xpm decoder i while ago
[06:05:13 CET] <durandal_1707> just search mailing list for xpm patches
[06:05:32 CET] <adeeln_> I see. I will look into it.
[06:05:50 CET] <durandal_1707> you of course need to extend it to be full featured
[06:06:38 CET] <durandal_1707> that means support for more than 256 colors
[06:07:02 CET] <adeeln_> Okay.
[06:07:07 CET] <durandal_1707> even that is straightforward
[06:10:25 CET] <durandal_1707> https://ffmpeg.org/pipermail/ffmpeg-devel/2012-June/126647.html
[06:11:23 CET] <durandal_1707> you need just to clean this patch and remove colors name stuff for now
[06:11:59 CET] <durandal_1707> and extend it for >256 colors
[06:12:41 CET] <adeeln_> Understood. Where do I post the result when I'm done? (considering I can do it).
[06:13:06 CET] <durandal_1707> to ffmpeg-devel mailing list
[06:13:25 CET] <adeeln_> Okay.
[06:21:51 CET] <adeeln_> Can't really find the old message you sent me. Searched the archive[http://www.ffmpeg-archive.org/].  Was this[https://ffmpeg.org/pipermail/ffmpeg-devel/2012-June/126647.html] just it?
[06:34:34 CET] <adeeln_> durandal_1707: I've done this so far: 1) copy libavcodec/dpx.c into libavcodec/xpmdec.c 2) renamed ff_dpx_decoder into ff_xmp_decoder 3) add AV_CODEC_ID_XMP to libavcodec/allcodecs.c and libavcodec/avcodec.h
[06:34:49 CET] <adeeln_> Should I move the discussion to ffmpeg-devel?
[06:35:13 CET] <durandal_1707> adeeln_: yes
[08:29:38 CET] <Aakash> I have made a rails app that can stream live videos to facebook rtmp server and deployed it on Digital Ocean. I have used nginx as web server and using a linux OS. The problem is that after streaming for some time(that keeps on varying) FFmpeg process gives the error  av_interleaved_write_frame(): Broken pipe  Pastebin link of command and logs: http://pastebin.com/vC1vLDEg
[08:29:55 CET] <Aakash> Can anyone tell me the root cause of this errir?
[08:30:02 CET] <Aakash> error*?
[11:25:27 CET] <Ekho> when using map_metadata to add metadata from a file can I map the metadata to a specific stream and not the entire file? the [STREAM] tag seems to still add the metadata to the entire file, not sure I understand what it does exactly
[13:25:22 CET] <BluesInTheNet> What video format would use least amount of cpu during playback if the video is just a single unchanging image?
[13:29:24 CET] <adeeln_> Could it be H.264? Don't know for sure.
[13:35:37 CET] <BtbN> CPU usage during playback isn't really a concern, unless you go for hevc or something like it.
[13:36:25 CET] <JEEB> I think the AVC decoder is so optimized now that you can just use that (with -tune fastdecode if you really want to)
[13:36:50 CET] <JEEB> but the best part for you if it's a single picture, is that if you just encode it once and set its duration to the duration of the whole video :P
[13:37:09 CET] <JEEB> suddenly you've got that background image you needed really cheaply
[13:38:54 CET] <BtbN> and then some crappy player barfs on it
[13:39:03 CET] <BtbN> or YouTube just re-encodes to 2 second gops
[13:40:07 CET] <JEEB> well youtube will do that anyways
[13:40:14 CET] <JEEB> and in that case you've just saved bandwidth
[13:40:24 CET] <JEEB> because you've transferred the image only once :P
[13:40:39 CET] <BtbN> I'd just upload lossless stuff to YouTube if bandwidth and size allows
[13:40:45 CET] <JEEB> yes
[13:41:03 CET] <JEEB> even in that, having a single image vs the same image over 9000 times does save bandwidth, though :)
[13:41:22 CET] <bencoh> :]
[13:41:45 CET] <BluesInTheNet> thx
[13:51:16 CET] <DHE> lower the framerate
[14:13:41 CET] <JEEB> yes, that's what adding a single picture with all the duration you need means ;)
[14:13:57 CET] <JEEB> and then utilize a container format that contains an actual duration, such as ISOBMFF
[14:14:15 CET] <DHE> ah, yes that works
[14:14:48 CET] <JEEB> stuff like matroska actually lacks a duration field, so the duration of the last picture is often left for guessing
[14:15:19 CET] <JEEB> (there's a metadata field later added by mkvtoolnix, but, uhh... > trusting metadata fields where you have to parse a string to get a number)
[14:15:52 CET] <JEEB> also that duration field IIRC is global, not track specific
[14:16:26 CET] <JEEB> oh well, as long as it gets improved in matroska v5 which they will be making after they get the current crapshoot standardized :)
[17:37:14 CET] <ftg> hello everyone, here is what ffmpeg -i outputs for my file: Video: h264 (High) (avc1 / 0x31637661), yuv420p, 1920x1080, 12733 kb/s, 30 fps, 30 tbr, 60k tbn, 120k tbc (default)
[17:37:46 CET] <ftg> I'd like to convert it into an ogg file to provide with video HTML5 player
[17:38:17 CET] <ftg> what are the best options for ffmpeg to do it without too much quality loss
[17:38:19 CET] <ftg> ?
[17:39:18 CET] <BtbN> every browser supports h264 anyway. So why convert?
[17:41:44 CET] <JEEB> ftg: theora is the worst alternative out of the freetard ones if you really want such. both vp8 and vp9 are available in all of the newer browser versions even if you strip out AVC/AAC support
[17:42:24 CET] <JEEB> build FFmpeg with libvpx support (preferably the latest tag that chromium builds) and look at examples at the webm wiki
[17:43:25 CET] <furq> or you know
[17:43:29 CET] <furq> just leave it as h264
[17:43:36 CET] <ftg> furq: ok
[17:43:52 CET] <JEEB> realistically, that's a good alternative as well as long as the bit rate is matching your requirements
[17:43:59 CET] <JEEB> or I'd even say, the best alternative in that case
[17:44:18 CET] <ftg> JEEB: may I degrade the video quality a little bit then?
[17:44:37 CET] <JEEB> yes, anything but lossless compression has some sort of loss
[17:44:54 CET] <ftg> because it is from a go Pro and file is heavy
[17:44:57 CET] <JEEB> and lossy (which your source most likely is) to lossless is lolhueg
[17:45:13 CET] <furq> if you're going to convert it then you might as well keep it as h264
[17:45:15 CET] <JEEB> (plus nothing supports lossless AVC other than lavc and a few other decoders)
[17:45:24 CET] <furq> the only reason not to do that is if you're concerned about licensing
[17:45:32 CET] <JEEB> yes, libx264 with CRF+maxrate|bufsize is your go-to choice
[17:45:55 CET] <JEEB> encode like 2500 frames starting with -crf 23, and going lower if it looks bad, or going higher if it looks good
[17:45:56 CET] <furq> i guess vp9 will potentially yield a smaller file at the same quality but it'll also take ten times longer
[17:46:10 CET] <JEEB> (-vframes 2500 for the 2500 frames)
[17:46:27 CET] <JEEB> then you have found the highest CRF that still looks good for you after a few tries
[17:46:31 CET] <furq> consider downscaling to 720p as well
[17:46:37 CET] <JEEB> and you can use that value for the rest of the similar clips
[17:47:02 CET] <JEEB> and then set maxrate+bufsize according to the minimum requirements you have regarding someone watching without additional buffering
[17:47:26 CET] <JEEB> (it will control the bit rate so that it limits bit rate to maximum of MAXRATE over BUFSIZE)
[17:53:33 CET] <ftg> JEEB: thx for the info! apart of this, I am also looking for a python lib that could extract a thumbnail image from this video, do you have a tool in mind?
[17:53:56 CET] <nadermx> hey all, I'm trying to convert a mp4 to mp3 in one command run with a thumbnail, I posted on SO http://superuser.com/questions/1182824/convert-mp4-to-mp3-with-thumbnail
[17:54:13 CET] <nadermx> the command I tried that works for a flv to mp3 doesn't work with mp4 to mp3
[18:22:20 CET] <forgon> I'd like to convert '.mkv' made of lossless libx264 and lossless flac to '.webm' so that I could upload to YouTube.
[18:22:40 CET] <JEEB> uhh, lossless AVC should work just fine o_O
[18:22:43 CET] <JEEB> other than for the file size
[18:22:47 CET] <forgon> I know both formats are fairly similar, so how much quality do I have to lose?
[18:23:02 CET] <JEEB> and you definitely don't want to use libvpx if you want good price performance lossy encoding
[18:23:59 CET] <dl2s4> doesnt encode youtube your uploads anyway?
[18:24:45 CET] <BtbN> which is why you upload lossless there
[18:24:58 CET] <BtbN> and mkv with h264 and flac should be fine to upload to YT
[18:26:46 CET] <dl2s4> yeah
[18:27:31 CET] <forgon> I'll try when my video is finished.
[18:27:33 CET] <JEEB> yup, which is why the only issue with it is the file size - in which case you either go to a place that has better interwebs or have to recompress and make your upload less nice in order to upload it better
[18:27:51 CET] <forgon> But I read that this is not possible to have flac and webm.
[18:28:04 CET] <forgon> And that YouTube will never support mkv.
[18:28:23 CET] <realies> av1 is the way to go
[18:29:12 CET] <JEEB> forgon: it takes that shit in just fine, it will re-encode whatever you push into it
[18:29:19 CET] <JEEB> it has a FFmpeg from circa 2011 or so
[18:29:20 CET] <JEEB> inside
[18:29:41 CET] <JEEB> so whatever that old FFmpeg can read it most likely can upload just fine
[18:29:46 CET] <JEEB> realies: not finished yet
[18:30:10 CET] <realies> JEEB, should be by the end of next month, no?
[18:30:20 CET] <JEEB> I would... be surprised :D
[18:30:37 CET] <JEEB> I remember at FOSDEM atomnuker was saying after the end of 2017 or so for bit stream freeze
[18:30:51 CET] <realies> wait, what
[18:30:52 CET] <JEEB> or at least his slides had something to that effect
[18:31:03 CET] <JEEB> https://fosdem.org/2017/schedule/event/om_av1/attachments/slides/1795/export/events/attachments/om_av1/slides/1795/av1_update.pdf
[18:31:32 CET] <JEEB> see the "After codec bitstream gets frozen experiments that didn't make it get removed"
[18:31:40 CET] <JEEB> ** - Won't happen until the end of the year
[18:32:02 CET] <forgon> What is the fastest way to cut my file cleanly into several parts at specific points?
[18:32:02 CET] <JEEB> which I read as the bit stream not being frozen until the end of the year
[18:32:25 CET] <JEEB> forgon: -ss and -t with -c copy
[18:33:10 CET] <realies> JEEB, thanks for the info, I was relying on information from this https://youtu.be/lzPaldsmJbk that stated a bitstream freeze by the end of nov 2016
[18:33:39 CET] <forgon> JEEB: So, that would allow me to get one interval and write it to a file, correct?
[18:33:45 CET] <JEEB> yes
[18:33:57 CET] <forgon> And I do that several times to get several parts, correct?
[18:34:40 CET] <JEEB> yes
[18:35:48 CET] <forgon> Another question: Can I extract a single shot as a png image out of a video with ffmpeg?
[18:35:58 CET] <JEEB> yes
[18:36:20 CET] <forgon> How?
[18:37:07 CET] <JEEB> you set output file to something dot png and do vframes 1 for that output?
[18:38:18 CET] <forgon> K.
[18:41:15 CET] <bencoh> JEEB: do they actually concat that properly?
[18:41:25 CET] <bencoh> (results from -ss/-t and -c copy)
[18:41:34 CET] <JEEB> that was not mentioned
[18:41:37 CET] <JEEB> he just wanted it to be cut
[18:41:39 CET] <bencoh> huhu :)
[18:41:43 CET] <JEEB> no idea how well the concat shit works
[18:41:43 CET] <bencoh> true
[18:41:55 CET] <forgon> No need to concat.
[18:42:15 CET] <bencoh> forgon: don't you intend to push resulting files to youtube?
[18:42:28 CET] <forgon> I have a 66 Warzone2100 video and neeed to cut it into 5 parts for YouTube which has a 15 minute limit.
[18:42:36 CET] <bencoh> oh, okay
[18:42:36 CET] <forgon> ^66^66m
[18:42:39 CET] <bencoh> nevermind then
[19:14:11 CET] <forgon> Can I stop ffmpeg in som eway and continue after a reboot?
[19:14:46 CET] <BtbN> no
[19:15:27 CET] <jkqxz> Not built in.  Run inside a VM and make a snapshot.
[19:24:40 CET] <realies> can u do it without a vm snapshot?
[19:25:00 CET] <realies> a generic way of snapshotting a process' ram usage
[19:25:10 CET] <realies> is that not a thing
[19:27:18 CET] <bencoh> actually it is, somehow
[19:28:30 CET] <bencoh> thè question is how would you handle io (reopen fds?)
[19:30:21 CET] <realies> that should be a function of that tool i guess
[20:16:13 CET] <Sweepyoface> Hey, I'm using the palettegen and filter_complex method to convert an mp4 into a gif, but it's turning 3.3MB into 22.7MB
[20:16:29 CET] <Sweepyoface> Is there a way to get this down without losing quality? It has to be <10MB
[20:16:55 CET] <JEEB> you thought gif would be well compressed? too bad, E_NOT_GOING_TO_HAPPEN (although there are some options in the encoder to optimize the palette etc)
[20:17:52 CET] <Sweepyoface> It seems to be able to convert >300MB avi/mov to a ~1MB gif
[20:18:04 CET] <Sweepyoface> but I can't go straight from AE unless I find a way to crop without also losing quality
[20:18:28 CET] <Sweepyoface> because the effect I used breaks when changing composition size
[20:21:12 CET] <Sweepyoface> I just converted a 1.2GB avi to a 3MB gif using the same method
[20:21:23 CET] <Sweepyoface> but if the input is only 3MB it gets inflated to 22MB
[20:21:49 CET] <Sweepyoface> because I had to run it through camtasia to crop it
[20:22:51 CET] <Sweepyoface> Maybe I can just convert first and then crop the gif
[20:24:22 CET] <durandal_1707> whatever you are doing, you are doing it wrong
[20:25:25 CET] <Sweepyoface> I need to convert a 1920x1080 AE animation (because changing the size in AE breaks it) to a 506x811 <10MB gif
[20:26:36 CET] <Sweepyoface> When I convert to a 1920x1080 gif it works fine
[20:26:39 CET] <Sweepyoface> but it has to be cropped
[20:26:51 CET] <voxadam> Is it possible to convert an encrypted/DRMed ISMA file to a non-encrypted/non-DRMed file of some type using ffmpeg (or any other tool)?
[20:27:34 CET] <voxadam> According to media info the file is an encrypted MPEG-4 PIFF with AAC audio.
[20:28:57 CET] <BtbN> ffmpeg does not support DRM
[20:32:31 CET] <voxadam> It supports encrypted files, right?
[20:32:39 CET] <voxadam> http://ffmpeg.org/ffmpeg-all.html#Audible-AAX
[20:33:01 CET] <DHE> sort of. I know HLS supports encryption but it lacks any means of doing a "secure handshake" or whatever is required to get the protected key.
[20:33:16 CET] <BtbN> hls supports decryption with a static AES key iirc
[20:33:19 CET] <BtbN> which seems super pointless
[20:33:36 CET] <voxadam> I have the "activation bytes
[20:33:45 CET] <voxadam> I have the "activation bytes" which I assume is a key.
[20:33:58 CET] <DHE> then you should be able to specify it as per the example
[20:34:47 CET] <voxadam> My problem seems to be related to the fact that the file I'm trying to convert isn't AAX but rather IMSA.
[20:34:57 CET] <DHE> BtbN: HLS keys are specified by a URL which has a 16 byte (128 bit AES) payload. but any good implementation will require something like cookies, HTTP authentication, or something else in order to get the key
[20:35:17 CET] <BtbN> But what's the point of that? Can just put the stream itself behind auth.
[20:35:23 CET] <DHE> voxadam: they're part of a family of containers. give it a try.
[20:35:43 CET] <DHE> BtbN: CDNs can carry the video payload, and a single secure point can hand out the keys
[20:35:50 CET] <BtbN> hm, yeah
[20:36:11 CET] <BtbN> does not really protect against re-distributions in terms of DRM
[20:36:35 CET] <DHE> well that's DRM for you
[20:36:45 CET] <__jack__> HLS security is pretty .. secure
[20:38:51 CET] <__jack__> it's also vague on how to secure your stream
[20:39:22 CET] <__jack__> let's take verimatrix, for instance : to fetch the AES key, you need to GET an https url (which is written on the playlist)
[20:39:34 CET] <__jack__> the https server does client-side authentification via TLS
[20:39:58 CET] <__jack__> which means only a trusted client (which have a signed certificate) can get the key and play the movie
[20:40:10 CET] <voxadam> This is what I get when I try to convert the IMSA file: https://paste.fedoraproject.org/paste/Rvrxv3NLm8zNa7MeC5I-ql5M1UNdIGYhyRLivL9gydE=
[20:40:11 CET] <DHE> so TLS client certs as authentication
[20:40:29 CET] <__jack__> yes
[20:43:14 CET] <voxadam> If I specify output.aac I only receive a single error "[aac @ 0x4cdd740] channel element 2.15 is not allocated" but the result doesn't play in vlc or ffplay.
[21:06:14 CET] <Darby_> hi guys i have solved many issue for compile a ffmpeg static build
[21:06:34 CET] <Darby_> i have a issue with libssh
[21:06:35 CET] <Sweepyoface> Anyone able to help me with this?
[21:08:54 CET] <Darby_> https://clbin.com/4BOjM
[21:32:31 CET] <BtbN> Darby_, static libs have no idea of dependencies. And it seems like libssh does not correctly specify all of its deps in the pkg-config file.
[21:32:41 CET] <BtbN> this looks like it needs to link against openssl
[21:37:25 CET] <Darby_> hi BtbN
[21:37:42 CET] <Darby_> have you a solution?
[21:37:51 CET] <BtbN> link against openssl yourself
[21:38:08 CET] <BtbN> Or just enable it, ffmpeg can use it itself, and that should workaround the issue with libssh
[21:38:25 CET] <Darby_> can you make me a example, please?
[21:38:46 CET] <BtbN> --enable-openssl
[21:39:34 CET] <Darby_> i have --enable-openssl
[21:39:52 CET] <Darby_> my issue is with libssh
[21:40:05 CET] <BtbN> no, your issue is with libssh not finding openssl functions.
[21:41:58 CET] <Darby_> BtbN, take a look please
[21:42:00 CET] <Darby_> https://clbin.com/vN3lX
[21:42:29 CET] <BtbN> I'm not going to fix your script for you.
[21:42:36 CET] <Darby_> no man
[21:42:57 CET] <Darby_> i don't want it
[21:43:01 CET] <BtbN> make sure pkg-config is working, and that should sort most things
[21:43:44 CET] <Darby_> i'm tryng to compile a ffmpeg full static binary
[21:44:08 CET] <Darby_> debian libssh-dev don't have static .a files
[21:45:04 CET] <Darby_> i'm compiling libssh in my custom dirs for incluede static library to ffmpeg binary
[21:48:01 CET] <Darby_> BtbN, how can i link openssl function to openssl?
[21:48:11 CET] <Darby_> *libssh
[21:51:43 CET] <BtbN> you can't. Static libs don't support dependencies.
[21:51:49 CET] <BtbN> That's why pkg-config helps
[21:52:28 CET] <Darby_> i can't compile with static libssh?
[21:53:56 CET] <BtbN> you have to pass all dependencies to the linker as well
[21:54:03 CET] <BtbN> Or use pkg-config...
[21:54:24 CET] <Darby_> how can i do this?
[21:54:33 CET] <Darby_> i'm not a expert
[22:00:04 CET] <Darby_> BtbN, please give me a little help
[22:04:22 CET] <DHE> do you have a static lib for openssl? you need libcrypto.a and libssl.a
[22:04:44 CET] <DHE> or possibly just libcrypto, I'm not sure of the specifics.
[22:06:39 CET] <Darby_> thanks DHE i can compile it
[22:14:47 CET] <Allaen5> Hi, I am trying to write chapter markers to an audio stream via -map_metadata, ouput file is a .m4a file. I can get it to work when I am using an existing input file but not when I am reading a stream from the pipeline e.g. BinaryWriter | -f s16le -ar 22050 -ac 2 -i pipe:0 -i "<path>\metadata.txt" -map_metadata 1 -id3v2_version 3 -write_id3v1 1 -y -b:a 64004 -metadata album="<album>" "<Output-File-Name-With-Full-Path.m4a"
[22:46:46 CET] <Darby_> DHE thanks i have compiled libssl and now i have libssl.a and libcrypto.a in my custom build directory
[00:00:00 CET] --- Mon Feb 27 2017