[Ffmpeg-devel-irc] ffmpeg.log.20161125

[burek]

[00:14:37 CET] <superguest> Hi
[00:17:56 CET] <Kadigan> Here's something I've always been meaning to ask: whenever there's an outdated API, an error in one of the modules or whatever that makes a dev comment pour out, I keep thinking -- "Why would I, the end-user, want to hear about AVStream.codecpar?!"
[00:18:12 CET] <furq> it's to keep you on your toes
[00:18:36 CET] <Kadigan> Seriously, though. Is there a flag that I can use to disable these messages?
[00:18:51 CET] <furq> nothing other than -v error
[00:18:52 CET] <Kadigan> (or better yet, could there be a flag I could use to ENABLE them, so default would be hiding them?)
[00:18:54 CET] <furq> which is probably a bit much
[00:19:25 CET] <furq> there should probably be something like -v deprecated just above -v warning
[00:19:42 CET] <Kadigan> It's annoying because there's nothing informative in there for ME, you know?
[00:19:56 CET] <furq> yeah that warning does nothing other than generate confusion on irc
[00:20:02 CET] <Kadigan> I'm like "Ooops, did I fuxxor something up? Oh, no, it's just another one of THOSE messages..."
[00:21:00 CET] <Kadigan> I mean, no, I get it - it's a message spat out by some part of code interacting with another part of code. It's just useless to me, so I find it odd to have it included in every output.
[00:22:11 CET] <Kadigan> Also, this kind of thing is fairly common in the open-source world, isn't it -- deprecating interfaces, APIs and all that. I mean, Microsoft took over a decade to finally throw some wonky backsupport away - as a good example of competetive shrink-wrap.
[00:30:04 CET] <c_14> Well nobody's paying us to support ancient interfaces.
[00:30:11 CET] <c_14> And there's always the effort to make things "better".
[01:14:40 CET] <Kadigan> "Better", I get. The issue, however, is with how "better", contrary to popular belief, is actually a -subjective- term.
[01:15:29 CET] <Kadigan> Though in this case I withold further comments, as I have no idea how the API looked before and thus whether there are reasons backing the split into .codec and .codecpar.
[01:15:38 CET] <furq> the issue isn't the split
[01:15:57 CET] <furq> the issue is either that ffmpeg still uses the old api, or that it tells users about that fact at the default loglevel
[01:16:20 CET] <Kadigan> I don't think I'm running latest, though...
[01:16:28 CET] <furq> it still does it on 3.2 iirc
[01:16:53 CET] <Kadigan> Well, it printing at default loglevel is definitely the issue, there I agree.
[01:17:40 CET] <Kadigan> There's a number of other things I'd see gone at default, too. Like how libfdk_aac spits out an error about VBR 590 being outside the range of 1-5... that's just silly.
[01:18:04 CET] <Kadigan> I'm not quite decided on which part of that mismatch is sillier.
[01:18:41 CET] <furq> it'd be nice if ffmpeg just didn't show deprecation warnings from the libs at all
[01:18:51 CET] <furq> it's a completely useless thing for a user to see
[01:18:59 CET] <Kadigan> I don't know, man... When I build a program, I leave stuff like that in, maybe with a -v switch or something. When I build a lib though... I tend to kill that stuff for -release.
[01:19:12 CET] <furq> i mean ffmpeg specifically, noto the libs
[01:19:14 CET] <furq> -o
[01:19:48 CET] <furq> i can imagine you'd want to avoid another set of ifdefs though
[01:19:54 CET] <furq> i'm sure there are plenty already
[01:19:58 CET] <Kadigan> That's assuming that the libs pump their errors/deps into a queue managed by ffmpeg,
[01:20:05 CET] <Kadigan> and not just direct to screen.
[01:20:11 CET] <Kadigan> Which I've seen happen.
[01:25:13 CET] <furq> i'm pretty sure they don't
[01:25:29 CET] <furq> well that was unclear
[01:25:41 CET] <furq> i'm pretty sure the libs just print error messages directly
[01:26:12 CET] <furq> so you'd need to fuck with the libs to fix this
[01:59:50 CET] <Kadigan> I guess.
[04:56:55 CET] <Sashmo> does anyone have a suggestion how I can get a sine generated to beep every second on the second, the same second that my clock is ticking ?  this looks and sounds good, but its not timed correctly.... http://pastebin.com/xt6xkUNR
[05:49:11 CET] <tomed> it takes way too long to use x264
[05:49:16 CET] <tomed> how do i make it faster
[09:42:37 CET] <superguest> hi
[09:45:53 CET] <kerio> waddup
[10:05:20 CET] <superguest> Hey Kerio
[10:06:05 CET] <superguest>  https://ffmpeg.org/ffmpeg-filters.html#Text-expansion
[10:11:31 CET] <superguest> it is said that 2ndArg in text=${pts\:localtime\:2ndArg} is "an offset added to the timestamp"
[10:11:52 CET] <superguest> Also, " By default, YYYY-MM-DD HH:MM:SS format will be used."
[10:12:31 CET] <superguest> How, and in what format, is the 2ndArg specified?
[10:13:30 CET] <superguest> Say I want to offset the burn-in timecode by +1 day, how do I do so?
[10:19:14 CET] <superguest> Note that sequences of the form %{...}  are expanded and that 'pts' is expanded as a function that prints the timestamp of the current frame.
[10:20:51 CET] <superguest> I've asked this here before, but no one responded.  Considering there's 360+ ppl in here I thought I could get a quick response.
[10:21:19 CET] <superguest> I guess I'll take this to video.stackexchange next.
[10:21:37 CET] <furq> i assume it's either seconds or a time spec
[10:23:08 CET] <superguest> furq, time spec in what format? that's the crux of my question really.
[10:23:22 CET] <furq> https://ffmpeg.org/ffmpeg-utils.html#time-duration-syntax
[10:26:13 CET] <superguest> furq, thanks!  I'll try that ^^
[10:49:23 CET] <techtopia> when encoding to wav to ac3 with ffmpeg
[10:49:32 CET] <techtopia> im loosing some of the audio channels
[10:50:18 CET] <techtopia> ffmpeg -i input.wav -vn -sn -acodec ac3 -q:a 640 output.ac3
[10:50:36 CET] <techtopia> ^ doing this and lose the center channel, anyone know why?
[10:54:49 CET] <techtopia> also just
[10:55:11 CET] <techtopia> ffmpeg -i input.wav -acodec ac3 output.ac3
[10:55:26 CET] <techtopia> the channel with dialogue just disappears
[11:03:14 CET] <ikevin> techtopia, take a look at https://trac.ffmpeg.org/wiki/AudioChannelManipulation
[11:09:28 CET] <techtopia> i don't want to remove the channels
[11:09:57 CET] <techtopia> i want to keep them all and just encode from 24bit pcm in a wav to ac3 at a lower bitrate
[11:20:14 CET] <ikevin> techtopia, you maybe need to map channels
[11:21:43 CET] <techtopia> hmm
[11:25:19 CET] <kerio> but why ac3 and not opus :^)
[11:26:21 CET] <techtopia> i don't mind kerio, opus would be fine if it works
[11:26:46 CET] <techtopia> atm i have a 7000kbps pcm wav thats 2.9gib
[11:27:00 CET] <techtopia> i just want to reduce it in size while retaining all the audio channels
[11:27:33 CET] <kerio> is ffmpeg -i whatever.wav whatever.ac3 actually dropping channels?
[11:27:41 CET] <kerio> like, is it saying that it's outputting in stereo
[11:29:27 CET] <techtopia> no it shows as 6channels still
[11:29:36 CET] <techtopia> but there is silence where there was dialogue before
[11:29:45 CET] <kerio> how are you playing the file?
[11:29:58 CET] <techtopia> mediaplayerclassic and vlc
[11:30:10 CET] <kerio> i'm not saying both are wrong
[11:30:14 CET] <kerio> but could you try mpv?
[11:30:29 CET] <techtopia> okay
[11:30:30 CET] <kerio> try other containers as well, like .mka
[11:31:38 CET] <furq> try another encoder as well
[11:31:43 CET] <furq> it might be an issue with the ac3 encoder
[11:32:09 CET] <kerio> if it's 48kHz, opus! :D
[11:32:51 CET] <techtopia> it's the same Keridos larger periods of silcene where dialogue used to be
[11:33:06 CET] <kerio> :(
[11:33:15 CET] <kerio> ffmpeg -i whatever.wav whatever.opus
[11:33:25 CET] <techtopia> ok trying now
[11:35:20 CET] <kerio> furq i has a question about opus
[11:35:40 CET] <kerio> would a s24 audio file sound better than a s16 one
[11:35:46 CET] <kerio> compressed at the same average bitrate
[11:36:42 CET] <furq> assuming you think s24 wav sounds better than s16, then sure
[11:37:12 CET] <kerio> i mean, do i have any reason to drop the bit depth before compressing?
[11:37:17 CET] <techtopia> http://pastebin.com/baRPgxin
[11:37:17 CET] <kerio> it should end up being pretty much the same, right
[11:37:25 CET] <techtopia> sam4e issue in opus Kerio :(
[11:37:33 CET] <kerio> techtopia: but the channels are there :(
[11:37:47 CET] <techtopia> yes but the audio no longer is
[11:38:33 CET] <furq> does it happen with opusenc
[11:38:42 CET] <furq> or $standalone_encoder
[11:39:14 CET] <ikevin> don't you have multiple languages for your dialogue?
[11:40:40 CET] <techtopia> it's just english
[11:48:16 CET] <techtopia> furq yes
[11:48:27 CET] <techtopia> it happens with eac3to also
[12:03:05 CET] <techtopia> hmm using ac3_fixed
[12:03:26 CET] <techtopia> i get silence still in media player classic, but extreamly muffled voices in vlc
[12:04:12 CET] <techtopia> i think it might have somthing to do with going from 7000kb/s to 660 kb/s
[12:07:47 CET] <techtopia> http://i.imgur.com/DK6IdDN.jpg
[12:07:53 CET] <techtopia> plays great in sourc e:(
[12:19:16 CET] <techtopia> Stream #0:0: Audio: pcm_s24le ([1][0][0][0] / 0x0001), 48000 Hz, 5.1, s32 (24 bit)
[12:19:29 CET] <techtopia> shouldnt it be s24 not s32?
[12:38:25 CET] <kerio> techtopia: i wonder if that's the issue
[13:18:08 CET] <gabrieliv> Is there any example of video decoding with the new _send/receive_packet/frame() API?
[13:23:35 CET] <BtbN> ffmpeg.c, however good of an example that is
[13:29:28 CET] <kerio> techtopia: have you figured out the issue?
[13:29:41 CET] <kerio> maybe your samples are just all effectively divided by 256
[13:34:45 CET] <techtopia> im still working on it kerio
[13:34:48 CET] <techtopia> no luck so far
[14:06:07 CET] <gabrieliv> What is the desired codec if I want to decode.mpeg video into .pgm frames?
[14:08:06 CET] <gabrieliv> Also, what is the right way to pass exactly one frame to the decoder?
[16:45:42 CET] <Sashmo> does anyone have a suggestion how I can get a sine generated to beep every second on the second, the same second that my clock is ticking ?  this looks and sounds good, but its not timed correctly.... http://pastebin.com/xt6xkUNR
[17:27:49 CET] <DHE> Sashmo: you're sending this out to multicast by UDP. players will perform at least minimal buffering so getting it time-synchronized is an exercise in futility
[17:30:46 CET] <kerio> DHE: gopless h264 with periodic intra refresh
[17:30:48 CET] <kerio> :3
[17:32:02 CET] <DHE> nevertheless bitrate variations or just variable frame sizes will require some form of buffering on the client side, even if it's something small like 200 milliseconds.
[17:32:20 CET] <DHE> intra refresh helps, but I don't think it's good enough for true CBR
[17:32:51 CET] <kerio> that one blog post that's only accessible through archive.org claimed 1 frame of latency
[17:33:10 CET] <Sashmo> dhe: I would like to agree with you but I dont thing thats correct.  The output portion has nothing to do with the sync of the beep and the video of the time.
[17:33:50 CET] <DHE> Sashmo: my point is there's buffering everywhere. even your sound card has some buffering going on
[17:34:09 CET] <DHE> err, wait...
[17:34:52 CET] <DHE> oh, derp... you mean the on-screen clock of the generated video...
[17:35:27 CET] <DHE> I'm gonna go stand over here now
[18:27:36 CET] <c_14> Sashmo: the beep_factor doesn't do what you think it does
[18:27:40 CET] <c_14> It doesn't beep every second
[18:27:47 CET] <c_14> It beeps a second after it last stopped beeping
[18:27:58 CET] <c_14> so it'll go out of sync with the timestamp
[18:28:06 CET] <Sashmo> c_14: yeah thats what I am seeing
[18:28:15 CET] <Sashmo> c_14: any ideas how to sync it up?
[18:28:21 CET] <Sashmo> Trying to make my own test card
[18:31:10 CET] <c_14> Besides changing the code?
[18:31:46 CET] <c_14> Creating the beep shouldn't be that hard, but getting it synced to the system clock is much harder.
[18:32:25 CET] <c_14> maybe sine filter with frequency you want, setpts set the timestamps to RTC-RTCSTART and then aselect to select only those parts around the second changes
[19:09:55 CET] <persina> Is there a difference between using the -loop argument versus the -concat?
[19:15:07 CET] <markvandenborre> ah, persina , you are interested in similar things as what I am interested in, at the same time
[19:16:15 CET] Action: markvandenborre is reading up on how to cleanly and efficiently prepend a preroll (title screen) and append a postroll (closing screen) to an existing video
[19:21:44 CET] <durandal_1707> concat demuxer or filter
[19:24:49 CET] <markvandenborre> durandal_1707: I'm reading up on that
[19:27:14 CET] <markvandenborre> to generate an image (problem solved) and convert that to video for a few seconds (still looking at that)
[19:27:53 CET] <markvandenborre> I can realise the separate steps, but I'm spending a bit more time than I like banging my head into the wall trying to solve it elegantly
[19:27:57 CET] <markvandenborre> :-)
[19:34:26 CET] <markvandenborre> durandal_1707: thx for the hint
[19:35:25 CET] <markvandenborre> the tricky thing is not as much to get this to work even
[19:35:45 CET] <markvandenborre> getting it to work efficiently without reencoding anything in the video part
[19:35:50 CET] <markvandenborre> that is the tricky bit...
[19:36:36 CET] <Billys> hello everyone!
[19:36:43 CET] <Mysoft> markvandenborre yeah... when aiming for doing that
[19:36:49 CET] <Mysoft> i usually save as .ts format
[19:36:53 CET] <Mysoft> so concat is just binary copy
[19:37:51 CET] <Mysoft> but if both contents are the same format... you can just concat them easily with the "codec copy"
[19:38:17 CET] <Billys> im preim extremely noob on ffmpeg
[19:38:19 CET] <Billys> and not the luxury of time to search on forum for my problem
[19:38:20 CET] <Billys> is here the place to ask for help?
[19:38:27 CET] <markvandenborre> Mysoft: I have an image that I want to display for a few seconds
[19:38:53 CET] <markvandenborre> easy with -framerate 1/5 -i image.jpg
[19:38:56 CET] <Mysoft> i think it would be better to first encode it into the same format as the video
[19:39:04 CET] <markvandenborre> then use that in a filter complex
[19:39:22 CET] <Mysoft> generating the N seconds and then just do the normal concat
[19:39:23 CET] <markvandenborre> so that I could use -codec copy you mean?
[19:39:29 CET] <Mysoft> yeah
[19:39:43 CET] <markvandenborre> that is something I'm trying to avoid if at all possible
[19:39:54 CET] <markvandenborre> cleaning up temporary files can get messy
[19:40:09 CET] <Mysoft> well i guess the filters will work fine
[19:40:29 CET] <markvandenborre> I wonder if I could have my cake and eat it
[19:40:44 CET] <Mysoft> since they are in a predicted order... and if i remember right there's options for timings with that
[19:41:01 CET] <Mysoft> but they would "override" the begin of the video
[19:41:03 CET] <Mysoft> instead of "concat"
[19:41:19 CET] <Mysoft> and that obviously requires encoding not just copy...
[19:41:33 CET] <Mysoft> so
[19:41:47 CET] <markvandenborre> I was hoping that I could accomplish something with filter_complex magic
[19:42:11 CET] <Mysoft> then what you can do is pipe your images...
[19:42:30 CET] <Mysoft> so concat the pipe input + video...
[19:42:36 CET] <Mysoft> so then you don't have a temp file....
[19:42:48 CET] <Mysoft> but... that will be troublesome for the end part @_@
[19:42:56 CET] <markvandenborre> :-(
[19:43:35 CET] <Mysoft> BUT
[19:43:44 CET] <Mysoft> if the image is the same on begin end
[19:44:07 CET] <Mysoft> the input can be infinitely repeating the same image
[19:44:21 CET] <markvandenborre> a -loop 1, yes
[19:44:32 CET] <Mysoft> and then you can just double the contact
[19:44:47 CET] <markvandenborre> not sure I uderstand you well here
[19:45:30 CET] <Mysoft> but there's another problem... the normal concat protocol
[19:45:42 CET] <Mysoft> requires a .txt file with the inputs to concat...
[19:46:11 CET] <Mysoft> so i think that avoiding temp files causes more problems than it solves hehe
[19:46:18 CET] <markvandenborre> seems so, yes
[19:46:35 CET] <Mysoft> but i  guess on a batch file
[19:47:04 CET] <markvandenborre> oh well, I can probably get more improvement suggestions by building the solution with temp files first
[19:47:20 CET] <markvandenborre> then going from there
[19:47:50 CET] <Mysoft> you probabily could get the .txt from pipe...
[19:49:39 CET] <markvandenborre> Mysoft: https://gist.github.com/markvdb/b23cb2f4cc2a158e15f65ad1e390a776
[19:49:44 CET] <markvandenborre> is what we used in the past
[19:49:59 CET] <markvandenborre> it made total sense at the time
[19:50:15 CET] <markvandenborre> but now, video input arrives as picture in picture with a background already
[19:50:32 CET] <markvandenborre> so we don't have to reencode any of the essentials anymore
[19:51:17 CET] <markvandenborre> (the script was the simple standalone test version, but you get the idea I guess)
[19:51:32 CET] <markvandenborre> now we don't do an overlay with the title anymore
[19:51:33 CET] <Mysoft> have you read this?
[19:51:33 CET] <Mysoft> https://trac.ffmpeg.org/wiki/Concatenate
[19:51:37 CET] <markvandenborre> yes
[19:51:42 CET] <Mysoft> at the end i think there's your magic
[19:51:49 CET] Action: markvandenborre reads again
[19:52:47 CET] <markvandenborre> ah, I stopped reading before at "With any vaguely-modern version of ffmpeg, the following script is made redundant by the advent the concat filter"
[19:54:17 CET] <Billys> hello?
[19:55:09 CET] <markvandenborre> Billys: hello
[19:55:19 CET] <Billys> can i ask for help?
[19:55:36 CET] <markvandenborre> Billys: better immediately ask what you would like to ask
[19:55:42 CET] <Billys> http://www.tiikoni.com/tis/view/?id=414b420
[19:56:44 CET] <Billys> when i play a straming video i think on mpeg4 i get this "lines"
[19:56:45 CET] <Mysoft> interlaced merging...
[19:57:05 CET] <Billys> ok so i search for "interlaced merging" ?
[19:57:42 CET] <Mysoft> that would work... as that is what happens there
[19:57:53 CET] <Mysoft> the video is interlaced (odd lines in one frame... even lines in another)
[19:58:17 CET] <Billys> there is a parameter to fix this?
[19:58:53 CET] <Mysoft> you said streaming... so the source of the video is a file?
[19:59:22 CET] <Mysoft> or it's a capture device?
[19:59:48 CET] <Billys> basicly is from tv tuner right now
[19:59:56 CET] <Mysoft> ok makes sense
[19:59:59 CET] <Billys> with other player the video its ok
[20:00:19 CET] <Mysoft> it may have auto-detected the interlace and applied the de-interlace filter
[20:00:25 CET] <Mysoft> (the player)
[20:01:05 CET] <Mysoft> so if you want to stream without the interlace then you need to reencode the vidoe while streaming with a deinterlace filter
[20:01:12 CET] <Mysoft> so that the video will become progressive then :)
[20:01:25 CET] <Billys> i see..
[20:01:57 CET] <Billys> so there is a command to enable that filter on ffmplay?
[20:02:14 CET] <Mysoft> yes
[20:02:34 CET] <Mysoft> probabily just
[20:02:37 CET] <Mysoft> ffplay -deinterlace
[20:02:39 CET] <Mysoft> will do the trick
[20:07:28 CET] <Billys> im soooo noob.... i cant set the parameter correctly :P
[20:10:44 CET] <Billys> anyway im pretty sure that i will make it work.
[20:10:49 CET] <Billys> Thank You!!
[20:11:12 CET] <markvandenborre> Mysoft: thank you for helping me too
[20:14:54 CET] <pgorley> hi all, what are the dependencies for libavfilter? i keep getting undefined references
[20:15:43 CET] <pgorley> i'm passing -lavfilter to the linker
[20:22:08 CET] <Mysoft> oh... dang -deinterlace is just for ffmpeg... so ffplay requires something like "-vf yadif"
[20:22:11 CET] <Mysoft> but it's too late @_@
[20:25:06 CET] <klaxa> pgorley: probably also needs other libs? pastebin your errors
[20:27:00 CET] <pgorley> klaxa: it's just a bunch of undefined reference to libavfilter functions
[20:28:11 CET] <klaxa> well show us what gcc/ld tells you
[20:34:07 CET] <pgorley> klaxa: http://pastebin.com/CBRWPG86
[20:35:50 CET] <c_14> Can you make V=1 or something?
[20:36:01 CET] <c_14> something to show the actual g++/ld calls
[20:37:07 CET] <pgorley> http://pastebin.com/wFdWXqE8
[20:38:48 CET] <c_14> Are the libav*.so files in your default LDPATH ?
[20:39:46 CET] <pgorley> i'd assume so, since it finds all the other libav*.so files
[20:40:47 CET] <pgorley> i can confirm that libavfilter.so is located at the same place as libavcodec.so and the others that i use
[20:41:22 CET] <c_14> Which is?
[20:41:24 CET] <pgorley> in /usr/lib/x86_64-linux-gnu
[20:44:57 CET] <klaxa> try adding -L/usr/lib/x86_64-linux-gnu?
[20:45:08 CET] <klaxa> to the command that fails
[20:46:51 CET] <klaxa> maybe your configure command also has something like --extra-ld-flags or you can set it via environment variables
[20:46:57 CET] <pgorley> will do
[20:47:02 CET] <klaxa> or add it in the makefile maybe
[20:47:10 CET] <klaxa> as long as it does the right thing :P
[20:52:47 CET] <pgorley> doesn't quite work, tried the env var and hardcoding it in my makefile
[20:53:05 CET] <c_14> What does "quite" mean? Does it do something different?
[20:53:20 CET] <pgorley> oh, sorry, no it doesn't
[20:54:04 CET] <pgorley> might have something to do with the order of dependencies in the makefile
[20:54:17 CET] <pgorley> does libavfilter depend on another libav*?
[20:56:28 CET] <iive> it definitely depends on avutil
[20:56:41 CET] <iive> it might also depend on avcodec
[20:57:27 CET] <pgorley> which are both placed before avfilter in the makefile :S
[21:07:56 CET] <CFS-MP3> I have a weird problem. Doing this:
[21:07:58 CET] <CFS-MP3> mpeg -i Westworld.S01E01.1080p.HDTV.6CH.x265.HEVC-PSA.mkv
[21:07:58 CET] <CFS-MP3> -filter_complex "fps=12,scale=416:-2:flags=lanczos,showinfo" -map v
[21:07:58 CET] <CFS-MP3> -codec:v libx264  -f segment -segment_time 6 -segment_list_type m3u8
[21:07:58 CET] <CFS-MP3> -segment_list wwseg.m3u8  wwseg%05d.mp4
[21:08:07 CET] <CFS-MP3> Generates many unplayable mp4
[21:08:20 CET] <CFS-MP3> however, if I generate .ts files isntead, and then convert those to .mp4, the result plays fine
[21:08:22 CET] <CFS-MP3> any idea?
[21:08:57 CET] <CFS-MP3> The conversion ts -> mp4 is as simple as ffmpeg -i ${filename} -acodec copy -vcodec copy "$(basename $filename .ts).mp4"
[21:28:46 CET] <persina> why does -loop take significantly longer than using -concat?
[22:42:26 CET] <markvandenborre> persina: because -loop is more resource intensive?
[22:42:45 CET] <markvandenborre> in some cases, concat can just copy instead of having to generate single frames?
[23:02:51 CET] Action: markvandenborre is having trouble concatenating using -c copy
[23:03:48 CET] <markvandenborre> I'm getting "[mp4 @ 0x56383ab4aa60] Non-monotonous DTS in output stream 0:0; previous: 197241, current: 157187; changing to 197242. This may result in incorrect timestamps in the output file.
[23:03:52 CET] <markvandenborre> "
[23:04:13 CET] <markvandenborre> a lot of warnings similar to that
[23:05:12 CET] <Mysoft> heh i remember getting a flood of this message too
[23:05:14 CET] <Mysoft> at some point
[23:05:25 CET] <Mysoft> but i dont remember the cause/outcome
[23:06:13 CET] <markvandenborre> and what makes me feel uncomfortable is that I get
[23:06:14 CET] <markvandenborre> [concat @ 0x560cc1820ae0] Unsafe file name '/tmp/prerollsilentaudio.mp4'
[23:06:39 CET] <markvandenborre> when I don't give concat the -safe 0 option
[23:07:18 CET] <markvandenborre> according to ffmpeg.org, the filenames I use should be perfectly safe, like the above
[23:07:26 CET] <markvandenborre> or am I missing something there too?
[23:10:10 CET] <markvandenborre> Mysoft: I see that the ffmpeg generated files have an encoding date set of 1904-01-01
[23:10:52 CET] <markvandenborre> quite sure that has nothing to do with the relative timestamps, but I'm really at a loss about what is happening here
[23:11:12 CET] <markvandenborre> where to investigate further what I am doing wrong
[23:13:12 CET] <markvandenborre> for some reason, I see the frame rate of the resulting file is not 25 FPS like all of the incoming ones,
[23:14:19 CET] <markvandenborre> but according to mediainfo, it's a variable frame rate mode, with frame rate 24.975 FPS, minimum frame rate 0.711 FPs and maximum frame rate 12800 FPS
[00:00:00 CET] --- Sat Nov 26 2016