[Ffmpeg-devel-irc] ffmpeg.log.20160622

burek burek021 at gmail.com
Thu Jun 23 02:05:01 CEST 2016

[00:03:43 CEST] <llogan> how can we duplicate this issue?
[00:48:20 CEST] <alsi> hi all
[00:49:06 CEST] <alsi> how can i stream memory data to ffmpeg console?
[00:49:14 CEST] <alsi> i am using c#
[00:49:45 CEST] <alsi> i grab rtsp camera frames
[00:49:50 CEST] <alsi> do some processing
[00:50:52 CEST] <alsi> and would like to encode the result into file via console call / pipes from c# app
[02:09:21 CEST] <wallbroken> hi guys
[02:09:38 CEST] <wallbroken> c_14 how to show audio track lenght with ffprobe?
[02:10:58 CEST] <llogan> wallbroken: https://trac.ffmpeg.org/wiki/FFprobeTips#Duration
[02:12:09 CEST] <wallbroken> i'm using: ffprobe.exe -v error -of compact -show_entries stream=duration "file.avi" -select_streams v:0
[02:12:18 CEST] <wallbroken> but it gives me the video lenght
[02:12:27 CEST] <wallbroken> i want to use it with an mp3 file
[02:12:33 CEST] <c_14> -select_streams a:0
[02:13:07 CEST] <wallbroken> ok, long time ago you said to me to use another directive
[02:13:10 CEST] <wallbroken> <c_14> wallbroken: ffprobe -show_format will show you up to 6
[02:13:13 CEST] <wallbroken> -show_format
[02:13:24 CEST] <wallbroken> i used -show_entries
[02:13:32 CEST] <wallbroken> something changes?
[02:13:55 CEST] <c_14> select_streams v:0 selets the first video stream, select_streams a:0 is the first audio stream
[02:15:08 CEST] <wallbroken> yes, but about the other question i asked?
[02:16:28 CEST] <c_14> format will show you the total length of the file while stream will show you the length of the individual streams
[02:17:39 CEST] <wallbroken> so if i have a file with 1 video track and 2 audio track, -show_format will give to me 3 values?
[02:18:09 CEST] <c_14> no, just one. Usually of the longest track
[02:18:52 CEST] <wallbroken> not very clear
[02:19:08 CEST] <wallbroken> i want to only know the lenght of the audio track, or of the video track
[02:19:17 CEST] <wallbroken> depending on what i need at that moment
[02:19:24 CEST] <wallbroken> which directive do i need to use?
[02:20:23 CEST] <c_14> use show_entries stream=duration and use select_stream to select the one that you want
[02:22:24 CEST] <wallbroken> ok thank you
[05:35:17 CEST] <noobsy> hi all, I'm wondering if there's any way I can supply ffmpeg a youtube video download url, extract audio from it, and return the audio stream as a C# Stream object
[06:21:18 CEST] <klaxa> noobsy, dunno about the last part, but you can probably use youtube-dl and some piping to get it into ffmpeg and then get the audio, maybe you can even request audio only from youtube with youtube-dl
[09:46:41 CEST] <`Misiak> Hi. Please, does anyone know, if ffmpeg supports FEC (SMPTE 2022-1) in a RTP stream?
[11:23:16 CEST] <pZombie> i found my code i did some years ago for taking screenshots and passing them via JNI, using a direct byte buffer, into java
[11:23:47 CEST] <pZombie> could this help in any way to add capturing the desktop in windows other than just the slow gdi method?
[11:24:11 CEST] <pZombie> here is the code, in the early stages. I also had gdi working at some point, but unfortunately i seem to have deleted it
[11:24:12 CEST] <pZombie> http://pastebin.com/6fryKKaw
[11:27:06 CEST] <pZombie> the other unfortunate thing is that i do not really remember what i have done there, back then. Only snippets
[11:30:50 CEST] <pZombie> If someone would tell me what kind of dll he needs for ffmpeg, as in what function call and what kind of object to return to ffmpeg with the screenshot data, i could try to work it out
[11:43:08 CEST] <waster> hello, can someone explain me please is it possible to insert timed metadata with ffmpeg in generated hls stream?
[12:16:26 CEST] <f00bar80> I installed ffmpeg , when tried to run i got ffmpeg: error while loading shared libraries: libavdevice.so.57: .. ,  ldd `which ffmpeg` returned lot of libs aren't not found as shown here http://vpaste.net/aZVTE , as well  find / -name libavdevice.so.52 returns nothing found , any idea what's wrong ?
[12:18:21 CEST] <CruX|> f00bar80: easyest way for you is to use static binaries
[12:18:40 CEST] <CruX|> i am using https://sites.google.com/site/linuxencoding/builds
[12:19:12 CEST] <f00bar80> CruX|: https://www.assetbank.co.uk/support/documentation/install/ffmpeg-debian-squeeze/ffmpeg-debian-jessie/ < that's how i compiled/installed ffmpeg
[12:23:20 CEST] <f00bar80> CruX|: should i install libav in my case ?
[12:23:57 CEST] <CruX|> there are both libav and ffmpeg
[14:07:48 CEST] <_Vi> How do I specify to scale filter that video should be scaled down to 50% without explicitly mentioning the size? Something like -vf scale=w=w/2:h=h/2 fails.
[14:14:04 CEST] <furq> _Vi: w=iw/2
[14:17:09 CEST] <_Vi> furq, Thanks.
[14:19:16 CEST] <_Vi> Why doesn't it say something like "Undefined variable 'w'" in the first case? I though that scale filter just doesn't support formulas.
[16:21:00 CEST] <azlux> hi
[16:21:48 CEST] <azlux> i'm new on ffmpeg, and i'm really impress how it's work
[16:21:54 CEST] <azlux> ... works weel
[16:22:53 CEST] <azlux> but i have a question when we manage audio from radio stream. If a start  reading a stream, can i change the filter for the volume without restarting ffmpeg ?
[16:23:24 CEST] <furq> no
[16:23:38 CEST] <furq> you could use something like the dynaudnorm filter
[16:24:42 CEST] <BtbN> The volume filter supports dynamic changes at runtime
[16:24:47 CEST] <BtbN> No idea if the cli tool does though
[16:25:09 CEST] <azlux> humm, no really i use it for a mumble bot, so it's the the global volume
[16:25:24 CEST] <azlux> BtbN: so i need to search for the cli tool
[16:26:59 CEST] <furq> ffmpeg is the cli tool
[16:27:31 CEST] <azlux> :) oups
[16:28:17 CEST] <azlux> i just found one interresting value for the volume filter
[16:28:18 CEST] <azlux> frame
[16:28:18 CEST] <azlux>     evaluate expression for each incoming frame
[16:28:28 CEST] <azlux> aie , bad copy paste .
[16:31:21 CEST] <azlux> only evaluate expression once during the filter initialization, or when the volume command is sent -> so we can "send" the command volume ?
[16:36:45 CEST] <c_14> You can send commands via the ffmpeg client on stdin, but I honestly don't know the syntax
[16:37:01 CEST] <c_14> You have to send a 'c' followed by the command change you want and then a newline
[16:37:24 CEST] <BtbN> Is that documented anywhere? And where does ffmpeg.c do that? I couldn't find it calling process_command
[16:38:06 CEST] <BtbN> ah, it's using avfilter_graph_send_command
[16:38:35 CEST] <c_14> Well, you can press '?' while ffmpeg is running and it'll tell you to use 'c' to send a command. Not actually sure if it's documented anywhere else
[16:39:08 CEST] <BtbN> also looks like it completely stops processing until the command is entered
[16:40:13 CEST] <azlux> indeed, press c , and volume -1 0.2
[16:40:15 CEST] <azlux> for example
[16:41:08 CEST] <azlux> need to look at the stdin, because i've disable it because of my subprocess
[16:42:05 CEST] <azlux> BtbN: i don't know if it's stop when we entrer new command, I can try only the evening
[16:42:28 CEST] <BtbN> well, it goes into a while != \n loop, so i very much assume it does block.
[16:43:02 CEST] <azlux> too bad
[16:44:01 CEST] <azlux> may be we can real the volume from a file to not interfere with the process. Don't know if it's exist
[16:44:07 CEST] <c_14> Well, the output device has a buffer so if you send the command fast enough you shouldn't notice.
[16:44:39 CEST] <azlux> c_14: thank
[16:44:46 CEST] <BtbN> if you send the input in a single write, it should not be a noticable block
[16:44:55 CEST] <azlux> it's a good idea
[16:46:23 CEST] <azlux> so i need to reactivate the stdin and try it
[16:52:21 CEST] <azlux> thank a lot
[16:59:17 CEST] <pgorley> hi all
[17:00:01 CEST] <pgorley> i'm implementing vaapi in my application, and i have a quick question
[17:00:34 CEST] <pgorley> if the hwaccel_retrieve_data callback is not called, does that mean the frame is not fully decoded yet?
[17:03:51 CEST] <jkqxz> ?  The retrieve_data callback is only a thing the ffmpeg program, not lav*.
[17:09:36 CEST] <pgorley> i'm using ffmpeg.c as an example
[17:11:14 CEST] <azlux> wow
[17:12:52 CEST] <azlux> netsplit ?
[17:19:37 CEST] <DHE> or some major routing issue
[17:20:37 CEST] <azlux> maybe
[18:09:09 CEST] <agrathwohl> Are colorspace conversions and scaling processes still single-threaded in FFmpeg?
[18:09:59 CEST] <furq> they are with swscale
[18:10:03 CEST] <furq> i believe zscale is multithreaded
[18:11:09 CEST] <DHE> I believe other processing can be done in the background though. One thread does per-frame conversions but other filters can be done in other threads
[18:13:49 CEST] <AnonBaite> can I post a sample here?
[18:22:15 CEST] <agrathwohl> DHE is that still the case when using the FFmpeg CLI? I cannot find in the documentation how to specify that filters each take their own threads
[18:26:12 CEST] <DHE> a quick search through the libavfilter docs show it does have threading support. actually looks like only sliced (frame fragments) filtering is supported
[18:26:21 CEST] <DHE> so it exists. not sure about the CLI interface specifics
[18:26:54 CEST] <AnonBaite> Here I go:
[18:26:54 CEST] <AnonBaite> https://www.sendspace.com/file/e05gkx
[18:27:28 CEST] <AnonBaite> I tested this file on ffmpeg and it says it has an unsupported codec
[18:29:11 CEST] <ocZio> hi all, reading this, https://trac.ffmpeg.org/wiki/Create%20a%20video%20slideshow%20from%20images is it possible to apply a watermark/overlay at the same time ?
[18:29:22 CEST] <ocZio> or I should make the video then apply the overlay afterwards ? :-)
[18:35:57 CEST] <furq> ocZio: do you mean a watermark and then a separate overlay
[18:38:13 CEST] <ocZio> furq, I am creating an mp4 out of 12 images, this part works fine
[18:38:34 CEST] <ocZio> what I would like to is to add a logo at the bottom left or any position on each frame
[18:38:47 CEST] <ocZio> my images don't have the logo included but I do have the logo image file
[18:39:02 CEST] <ocZio> is it possible to do this at the same time or it will involve to separate ffmpeg operations ?
[18:39:14 CEST] <furq> you can do it at the same time
[18:40:13 CEST] <ocZio> ok, any hints? :P so I can rtfm..
[18:42:39 CEST] <furq> https://ffmpeg.org/ffmpeg-filters.html#overlay-1
[18:45:03 CEST] <ocZio> ok
[18:45:23 CEST] <ocZio> ty, will read this
[18:51:42 CEST] <purftzc> would you call ffserver alive? or is it rather abandoned/dead?
[18:51:59 CEST] <BtbN> It still exists for some reason, but that's pretty much it.
[18:53:41 CEST] <purftzc> thanks
[18:53:55 CEST] <furq> it may not exist for much longer by all accounts
[18:54:53 CEST] <furq> only its spirit will live on in the form of bad stackoverflow answers
[18:55:04 CEST] <BtbN> Someone should just send a patch that removes it.
[18:55:15 CEST] <BtbN> Or disables it by default for a start.
[18:55:34 CEST] <furq> i thought it was disabled by default
[18:55:48 CEST] <BtbN> is it? Might be Gentoo enabling it then.
[18:56:10 CEST] <BtbN> In that case, removing it is.
[18:56:18 CEST] <furq> https://github.com/FFmpeg/FFmpeg/blob/master/configure#L119
[18:56:21 CEST] <furq> maybe not then
[18:57:49 CEST] <remitamine> is there a way to choose the quality that will be played by cvlc for ism manifets.
[18:58:20 CEST] <remitamine> sorry asked in the wrong irc.
[19:10:25 CEST] <ocZio> furq, I am having a slight issue by using that
[19:10:45 CEST] <ocZio> ffmpeg -y -framerate 2 -pattern_type glob -i '*.png' -i globe.png -c:v libx264 -vf "fps=25,format=yuv420p,scale=-2:720,overlay=main_w-overlay_w-10:main_h-overlay_h-10" out.mp4
[19:10:52 CEST] <ocZio> is this the format for the overlay ?
[19:19:01 CEST] <Guest35> hi there ffmpeg folks, I've been trying to figure out how to get avcodec_decode_video2 to report frame errors to me (i see them printed out)
[19:19:13 CEST] <Guest35> so that frames with errors won't get displayed
[19:20:14 CEST] <Guest35> could someone please explain how that's done? Clearly libavcodec is aware of the errors, and the frames are showing them. I want to just skip those frames and not show them. But how can I tell which frames actually had errors in them?
[19:20:32 CEST] <Guest35> I've tried, among other things, av_frame_get_decode_error_flags, but that always returns 0
[19:20:42 CEST] <Guest35> i'm on 3.0.2
[19:41:46 CEST] <soulshock> is there a comparison of the different resizing algorithms somewhere?
[19:41:50 CEST] <soulshock> for downsizing
[19:59:08 CEST] <cowai> according to the idet filter, "-target pal-dv" does not make the output interlaced.
[19:59:26 CEST] <cowai> I am trying to produce a standard pal dv avi file from prores 422
[19:59:47 CEST] <cowai> idet reports the file being mostly progressive.
[20:01:26 CEST] <cowai> ffmpeg -ss 00:15:00.00 -i prores.mov -t 30 -target pal-dv -flags -ilme -top 0 -y kanal10-dv.avi && ffmpeg loglevel info -i kanal10-dv.avi -vf idet -c:v rawvideo -an -t 10 -f mov -y /dev/null
[20:01:26 CEST] <cowai> [Parsed_idet_0 @ 0x2205e20] Repeated Fields: Neither:   251 Top:     0 Bottom:     0
[20:01:26 CEST] <cowai> [Parsed_idet_0 @ 0x2205e20] Single frame detection: TFF:     0 BFF:     0 Progressive:   219 Undetermined:    32
[20:01:26 CEST] <cowai> [Parsed_idet_0 @ 0x2205e20] Multi frame detection: TFF:     0 BFF:     0 Progressive:   246 Undetermined:     5
[20:02:02 CEST] <kepstin> cowai: 'idet' doesn't report whether the video feed is interlaced or not, it reports whether there's visible 'combing' artifacts that would indicate an interlaced (e.g. 50i) source.
[20:02:15 CEST] <cowai> aha
[20:02:58 CEST] <kepstin> cowai: it could very well be that your video is using an interlaced coding, but is actually 2:2 pulldown, so there's no interlacing artifacts visible (this is common for movies, and "film-look" tv shows in PAL)
[20:03:29 CEST] <cowai> the prores source is 720p50
[20:03:41 CEST] <cowai> and I am using ffmpeg 2.8
[20:03:52 CEST] <kepstin> 50 frames per second progressive? are you sure?
[20:03:55 CEST] <cowai> yes
[20:04:29 CEST] <cowai> 100% positive.
[20:04:29 CEST] <kepstin> if so, the "-target pal-dv" setting is probably simply dropping every second frame.
[20:04:50 CEST] <cowai> I dont want that.
[20:05:01 CEST] <cowai> I want it to extract every other line from each frame.
[20:05:09 CEST] <cowai> producing 50i as dv is
[20:05:17 CEST] <kepstin> cowai: then you'll want to use the "interlace" filter in ffmpeg
[20:05:36 CEST] <kepstin> cowai: make sure to read the options to that filter, so you configure the tff/bff correctly.
[20:06:40 CEST] <kepstin> I believe you'll want to use "-vf interlace=scan=bff"
[20:09:58 CEST] <cowai> there is no progressive dv so it seems strange having to deal with this.
[20:10:50 CEST] <cowai> Kepstin: Thanks I will try that.
[20:11:44 CEST] <cowai> -vf interlace=scan=bff, did not change the output of idet.
[20:12:32 CEST] <kepstin> you used that '-vf interlace=scan=bff' when encoding the dv avi file from the prores source, right?
[20:12:38 CEST] <cowai> yes
[20:12:46 CEST] <cowai> I just added it to the command I posted.
[20:13:12 CEST] <kepstin> well, the command you posted was two commands
[20:13:22 CEST] <kepstin> it has to go into the first one :0
[20:13:22 CEST] <cowai> Ah, I meant the first part of it.
[20:13:28 CEST] <cowai> yes of course ;)
[20:14:20 CEST] <cowai> it may very well be that my understanding of the idet filter is wrong and that the file really is interlaced.
[20:14:47 CEST] <cowai> As dv is either top or lower field interlaced, I would assume it actually is.
[20:14:56 CEST] <Guest35> does anyone know how to get avcodec_decode_video2 to report frame decoding errors to the caller? I'd like to avoid presenting frames that have errors.
[20:15:09 CEST] <cowai> But I dont know how to confirm if the file really is interlaced.
[20:15:11 CEST] <kepstin> cowai: idet is looking for "combed" frames, i.e. ones where the top and bottom field are from different times.
[20:15:31 CEST] <kepstin> lack of those doesn't mean the input isn't interlaced - it could just be static (no motion)
[20:15:46 CEST] <cowai> ah, there is lots of motion.
[20:15:53 CEST] <cowai> So it should be different
[20:16:17 CEST] <kepstin> watching the dv file in a desktop video player (e.g. mpv) with the deinterlacer disabled should make it obvious
[20:16:21 CEST] <cowai> is there any other way I could see if frame 1 has a second line or if frame 2 has a first line?
[20:16:37 CEST] <cowai> it does not look interlaced.
[20:17:03 CEST] <cowai> wait.
[20:17:13 CEST] <cowai> hold on, "-vf interlace=scan=bf" does look interlaced.
[20:17:15 CEST] <cowai> Thanks!
[20:21:32 CEST] <cowai> I would call this a bug though, that the target does not produce a valid dv file.
[20:32:54 CEST] <Bray90820> Can someone tell me what these errors mean?
[20:32:54 CEST] <Bray90820> http://pastebin.com/raw/NcDTyXTw
[20:33:50 CEST] <kepstin> cowai: I'm quite sure it does produce a valid dv file, just one where both fields in the interlaced encoding come from a single frame
[20:34:01 CEST] <kepstin> cowai: (which is sometimes called 2:2 pulldown)
[20:35:07 CEST] <cowai> what do you mean both fields?
[20:35:15 CEST] <cowai> you mean the next line in the next frame?
[20:35:48 CEST] <cowai> both line 0 in frame 1 and line 1 in frame 2 is from the same original frame 1 ?
[20:36:50 CEST] <kepstin> hmm? no, I mean content in field 1 and the content in field 2 both come from the same original progressive frame.
[20:37:16 CEST] <kepstin> and then when displayed on a computer monitor, the result will /look/ like 25 frames/s progressive video.
[20:39:44 CEST] <cowai> makes sense.
[20:40:18 CEST] <cowai> Although I though interlaced meant that first frame line 1,3,5,7... and second frame used 2,4,6,8
[20:40:26 CEST] <cowai> or somehing to that effect.
[20:40:47 CEST] <kepstin> nah, in interlaced encoding those are called "fields", not "frames"
[20:41:15 CEST] <kepstin> the first field encodes one set of lines, then the second encodes another set of lines that are offset to fit between the first set
[20:41:23 CEST] <cowai> yes
[20:41:25 CEST] <kepstin> and they alternate back and forth
[20:41:33 CEST] <cowai> yeah that is what I mean
[20:41:38 CEST] <cowai> even/odd alteration
[20:42:17 CEST] <kepstin> most digital video systems store interlaced video by putting all the lines from the first field into the even (or odd - this is the tff vs bff difference) lines of a frame, and then the second field into the other lines.
[20:42:17 CEST] <cowai> so what I gather from all this is that -target pal-dv is tuned for input that is 25fps (progressive) and not for 50p
[20:42:37 CEST] <kepstin> cowai: -target pal-dvd expects input that is 25p or 50i
[20:42:55 CEST] <cowai> pal-dv not pal-dvd
[20:43:06 CEST] <kepstin> it will not work correctly on 50p; it'll effectively put an "fps=25" filter before encoding the video
[20:43:23 CEST] <cowai> in effect loosing half the information
[20:43:40 CEST] <kepstin> yep; but interlacing (e.g. with -vf interlace) loses half the information too
[20:43:43 CEST] <kepstin> just a different half
[20:44:04 CEST] <cowai> yes interlacing is bad :)
[20:44:27 CEST] <kepstin> interlacing loses spatial information (drops half the lines), fps adjustment loses temporal information (drops half the frames)
[20:44:39 CEST] <cowai> yes I know.
[20:45:16 CEST] <cowai> ffmpeg could be a little smarter about handling 50p+ material for the target pal-dv
[20:45:17 CEST] <kepstin> so yeah, if you don't say anything else, ffmpeg (being mostly a frame-based application used a lot with progressive video) will do the framerate adjustment.
[20:45:35 CEST] <kepstin> if you want something else, you have to say what you want
[20:47:19 CEST] <cowai> yes, perfecting interlaced output is kind of not the way forward ;)
[20:48:37 CEST] <cowai> so just to sum everything up, -target pal-dv -vf interlace=scan=tff will take half the lines in first frame of original source and half of the next to produce 50i ?
[20:49:00 CEST] <cowai> for me this would give the best result wouldn't it?
[20:50:03 CEST] <kepstin> assuming you value having smooth motion over having higher resolution, and are not intentionally going for a "film look", yes
[20:51:21 CEST] <kepstin> (and I'm not certain, but I *think* pal-dv should be bff, not tff)
[20:52:00 CEST] <cowai> you are correct. But the playout this file is going to played in, has a SDI outout which is tff
[20:52:06 CEST] <cowai> so they use tff
[20:53:31 CEST] <cowai> I dont actually loose any resolution height wise since this will be played from a SD playout which discards the lines not used in interlaced.
[20:53:43 CEST] <kepstin> in theory, the playout should realize that the dv source is bff and offset the video :/ Putting tff video in DV will just be confusing for whoever comes across the file later :)
[20:54:58 CEST] <kepstin> i suppose making it work in practice is probably more important than theories about video fields :)
[20:55:46 CEST] <kepstin> at least if you get it wrong, it's obvious; stuff will jitter back and forwards horribly.
[20:56:01 CEST] <cowai> yes, its not my playout so I dont care.
[20:56:23 CEST] <cowai> I will give them a sample of both and they can check.
[20:56:43 CEST] <cowai> I just want to make sure they dont just rerender it before playing it.
[20:56:50 CEST] <cowai> that's my biggest concern.
[20:57:51 CEST] <cowai> Kepstin: thanks for your help btw!
[20:58:01 CEST] <kepstin> just be glad you're not dealing with ntsc stuff :)
[20:58:56 CEST] <cowai> I will soon, when we are digitizing hundres of old 1inch tapes.
[20:59:17 CEST] <cowai> That will be a night mare dealing with the framerates.
[21:01:37 CEST] <kepstin> ntsc is fun, because SDI stuff is all 486 line, but most digital storage formats are 480 line, and they're taken from different offsets within the 486 lines :/
[21:02:04 CEST] <cowai> haha
[21:02:12 CEST] <cowai> americans are crazy.
[21:02:37 CEST] <cowai> :)
[21:02:52 CEST] <DHE> famously acronym'd "never twice same colour"
[21:05:01 CEST] <kepstin> it's fun because ntsc DV is bff and broadcast digital video 480 line format (mpeg stuff) is tff; but unlike pal this isn't the order of the fields is simply reversed - it's instead because they grab different lines of the original video :/
[21:05:38 CEST] <kepstin> (and you have to put a bunch of "usually, but not always" in that sentence, too)
[21:05:57 CEST] <cowai> How does this happen really
[21:06:17 CEST] <cowai> Why do not people talk with eachother when making these choices.
[21:26:15 CEST] <kepstin> oh, huh, the pal DV spec acutually has the same issue, it's offset by 1 line from the normal sampling position :/
[21:26:37 CEST] <JEEB> > DV > 4:1:1
[21:26:44 CEST] <JEEB> someone please dot jaypeg
[21:34:44 CEST] <P4Titan> I have a pcm to aac encoder that I wrote up
[21:35:14 CEST] <P4Titan> However, when I input the pcm data, it encodes and produces a file with a reasonable size for aac
[21:35:28 CEST] <P4Titan> but when I run it, it says that it is invalid or something
[21:35:47 CEST] <P4Titan> Would someone be willing to help me out on this one?
[22:14:28 CEST] <Bray90820> Most ffmpeg commands are cross platform right?
[22:16:14 CEST] <durandal_1707> should be
[22:16:17 CEST] <DHE> yes, but they're command-line (short of ffplay)
[22:18:11 CEST] <Bray90820> Yes I know there command line
[22:22:06 CEST] <jnorthrup> is there a succinct link to get acquainted with ffmpeg variable substitution?  i tend to cruise along understanding the static stuff well enough but then substitution is typically exemplified shortly after the simple version.
[22:34:10 CEST] <Expo> anyone have experience with using force-style on srt subtitles?
[23:33:28 CEST] <rmbeer> hello
[23:35:36 CEST] <rmbeer> i found a list of error in "ffmpeg -i 'input file' 'output file'" for first video of https://mega.nz/#F!xZFkADpb!CoYwZWy6k6YHxqtfONUyfQ
[23:40:53 CEST] <rmbeer> the command is this, nothing weird. and output is the infinite count of this text: '[aac @ 0xXXXXX] element type mismatch 3 != 0'
[23:44:23 CEST] <kepstin> rmbeer: it's hard to say, given that you didn't actually follow the instructions you were given, but the problem is probably that your file is corrupt...
[23:48:15 CEST] <rmbeer> this no need a procedure for report the problem. i say all that needed for test and report if have any bug, i unknown if this error only have in my machine or really have any bug. i did the test and getting a endless error. and this is all, i can't do more for give more details.
[23:49:47 CEST] <rmbeer> only remains that other also testing and view if really have a error or not.
[23:50:25 CEST] <rmbeer> thank for you help. :)
[00:00:00 CEST] --- Thu Jun 23 2016

More information about the Ffmpeg-devel-irc mailing list