[Ffmpeg-devel-irc] ffmpeg.log.20140825

burek burek021 at gmail.com
Tue Aug 26 02:05:02 CEST 2014


[01:58] <djdduty> how can I make a timelapse video made with ffmpeg.exe -i %d.png -i song.mp4 -r 20 video.mp4 go longer than the song length?
[01:58] <djdduty> it keeps stopping the compilation at the length of the song rather than doing the entire thing.
[04:26] <Primer> How can one specify multiple input files with ffmpeg? Just multiple -i arguments?
[04:27] <Primer> I was hoping I could use a bash pattern, as I have several files...was hoping to not have to specify them all manually
[04:28] <Primer> seems that doesn't work anyhow
[04:29] <Hello71> !glob
[04:30] Action: Hello71 grumbles
[04:37] <Primer> I presume that's a hint to answer my question?
[04:37] <Primer> but the bot's not working?
[04:39] <Primer> I want to create a timelapse from several video files. I've been able to turn one video into a directory of images, then make a video of those images, but I have about 60 video source files
[04:39] <Primer> I was hoping to just batch that into one command line
[04:47] <Primer> nm, figured it out
[06:46] <benlieb> how do I change the scale of an mp4? I'm trying by -vf scale=WWW:HHH, where the dimensions give would in theory change the aspect ratio, but the aspect ratio doesn't change.
[06:55] <benlieb> anybody here?
[07:03] <cbsrobot-> benlieb: setdar=xxx/yyy
[07:04] <benlieb> every time I try this stuff I think i finally get it, but I don't.
[07:04] <benlieb> cbsrobot-: should I do that in addition to the scale -vf?
[07:05] <benlieb> I'm working with a video is reporting it's format as: JVT/AVC Coding, 592 x 416 AAC, 48000 Hz, Stereo (L R)
[07:05] <benlieb> and size: 720 × 505
[07:05] <benlieb> what is the 592 x 416 if not size?
[07:09] <cbsrobot-> yeah just write: -vf scale=xxx:yyy,setdar=uuu/vvv
[07:29] <benlieb> cbsrobot-: my real question is this: I have a lot of converted DVD videos with the dimensions above, and am trying to make "intro" videos with text from a database to add to the front of them. I'm just having trouble getting the sizes right. h264 is complaining about can't divide by 2...
[07:36] <cbsrobot-> benlieb: lets make it simple:
[07:37] <cbsrobot-> at least I know what you are dealing with
[07:37] <cbsrobot-> 592 x 416 is even, but I wonder where 720 × 505 come from
[07:38] <benlieb> yeah, I don't know...
[07:38] <benlieb> I don't understand what these 2 different dimensions refer to...
[07:39] <benlieb> jesus I figured it out.
[07:40] <benlieb> I'm viewing the video in quicktime, and the size it lists is the size of the video after any mouse resizing...
[07:40] <benlieb>  ug
[07:40] <benlieb> so i had made it bigger a while back... hm
[07:40] <benlieb> Anyway the command I'll be using will be:
[07:40] <benlieb>     `ffmpeg -y -i #{@dir}/intros/#{@lesson.id}.mp4 -i #{@dir}/raw_lessons/#{@lesson.id}.mp4 -filter_complex 'concat=a=1' #{@quality} #{@dir}/mp4_big/#{@lesson.id}.mp4`
[07:41] <benlieb> this command will just concat to vids together.
[07:41] <benlieb> my general question is do the sizes have to be exactly the same and aspect ratio too?
[07:41] <benlieb> I'm still trying to figure out if SAR and DAR are properties or calculations ...
[08:04] <Baked_Cake> is there a room for libfdk_aac
[08:05] <Baked_Cake> or aac in general, i got a question about the vbr
[08:15] <cbsrobot-> benlieb: not sure it does what you want, but without the actual console output I cannot help you - sorry.
[08:54] <benlieb> can someone point me to the filter_complex syntax docs? Can't find it on the website (which is looking good these days btw)
[09:33] <K4T> Hi! I need to convert 23.976fps film to 25fps aka film to PAL. As I know I have to speed up my input about 4% (video and audio). How I can achieve it? I am trying to use filters in ffmpeg: -filter_complex "[0:v]setpts=0.5*PTS[v];[0:a]atempo=2.0[a]" -map "[v]" -map "[a]" -r 25
[09:33] <K4T> But I don not have idea what values should I use
[09:33] <K4T> can someone help me with it?
[09:37] <Repe> hey, I am trying to convert alac -files to mp3 with ffmpeg and LAMEpreserving metadata and copying files from my musicserver to mounted mp3 player at the same command, keeping the folder structure as the same as in server. I dont know much about scripting and I am a bit lost here...Can someone help me with this one, please?
[10:00] <K4T> someone?
[15:27] <K4T>  what difference is between trancoding and transrating?
[15:33] <sacarasc> What is transrating?
[15:33] <sacarasc> Transcoding is changing from one encoding to another.
[15:35] <K4T> I think transrating is when we encode video with the same codec
[15:35] <sacarasc> It seems transrating is keeping most of the stuff the same, but changing the bitrate.
[15:36] <K4T> yeah
[16:33] <John_Loki> hi, how to user ffserver without involving transcoding, meaning just forward the stream which is already H.264
[16:33] <John_Loki> use*
[16:40] <Fjorgynn> aha
[16:41] <Fjorgynn> -c copy?
[16:41] <Fjorgynn> or what
[16:41] <Fjorgynn> John_Loki: I don't know. Just wait
[16:42] <John_Loki> -vcodec copy
[16:43] <John_Loki> has been used but a lots of frames start missing
[16:44] <John_Loki> idea is that the steam which has been mapped on the stream elements should be streamed as it is, no transcoding required
[16:44] <John_Loki> stream*
[16:44] <John_Loki> feed element *
[17:49] <kaotiko> hi
[17:57] <fajung> Hi, I'm trying to convert and mkv to mp4 using ffmpeg like this: http://pastebin.com/J4K7a6uE | I can play the mp4 but in my WD TV Live can't play it, does anybody knows why?
[17:58] <Mavrik> fajung, add the output also to pastebin
[17:58] <Mavrik> so we can see the format of files involved.
[17:58] <Mavrik> also, why are you transcoding =
[17:58] <Mavrik> :)
[17:59] <fajung>   >> ./a.log?
[18:00] <fajung> ok, I'll have to re encore it again, I'll came back as soon as it finished
[18:26] <fajung> Mavrik: here is the log: http://pastebin.com/a9getf8b
[18:40] <Mavrik> fajung, I see no reason why WD TV wouldn't play that
[18:41] <Mavrik> fajung, btw, why don't you just do ffmpeg -i <file>.mkv -codec copy <file>.mp4 ?
[18:41] <Mavrik> it's way faster and you don't lose quality
[18:46] <fajung> I'm doing it to get a smaller size file
[18:52] <sfan5> fajung: maybe WD TV Live only supports H.264 Main or Baseline
[18:53] <fajung> so, I'm new in all this things?
[18:55] <Mavrik> ah
[18:55] <Mavrik> lemme pull up tech specs of wd tv
[18:56] <fajung> I'm going to smash it against the wall
[18:57] <Mavrik> hmm, it seems that it should support the high profile
[18:57] <Mavrik> I see no reason why that wouldn't work :/
[18:57] <fajung> it play all the .HDTV.x264-LOL.mp4 that I have
[18:57] <Mavrik> try v:profile main
[18:59] <fajung> within the -vcodec libx264 ? | -vcodec libx264 v:profile mail ?
[19:06] <fajung> I can't sign in the wdtv forum to ask :'(
[19:18] <giamnt> hi
[19:22] <giamnt> I am encountering problems with the encoding of a file.  [ avconv -y -i video.mp4 -c:a mp3 -c:v mpeg4 -vtag xvid video.avi ] returns me: [ Unknown decoder 'libmp3lame' ]
[19:23] <giamnt> Same thing with [ avconv -y -i video.mp4 -c:a libmp3lame -c:v mpeg4 -vtag xvid video.avi ]
[19:25] <sacarasc> #libav is the avconv channel.
[19:25] <giamnt> ubuntu 12.04 64bit, I have already contacted the channel #ubuntu and already reinstalled all lib, but nothing seems to work.
[19:25] <fajung> try to recompile ffmpeg
[19:26] <sacarasc> giamnt: This is for ffmpeg, not libav/avconv. You're asking the wrong channel.
[19:26] <fajung> https://trac.ffmpeg.org/wiki/CompilationGuide/Ubuntu
[19:26] <giamnt> sacarasc yep same error with [ ffmpeg -y -i video.mp4 -c:a libmp3lame -c:v mpeg4 -vtag xvid video.avi ] = [ Unknown decoder 'libmp3lame' ]
[19:27] <sacarasc> On Ubuntu, usually the ffmpeg executable is just a front end for compatibility with the fork.
[19:27] <fajung> --enable-libmp3lame
[19:28] <giamnt> fajung you mena: ffmpeg --enable-libmp3lame etc etc ?
[19:28] <giamnt> *mean
[19:29] <fajung> try to compile ffmpeg as say in the copilation guide
[19:29] <fajung> I had the same trouble but with aac codecs
[19:30] <giamnt> fajung samething more simple, like edit a conf file?
[19:30] <giamnt> *something
[19:30] <fajung> follow the guide, it will work
[19:33] <giamnt> fajung ok, I'll try that....thanks by now :)
[19:40] <Andy2> Hi guys, i am playing around with -vf delogo and noticed that it removes the logo but changes the whole image color slightly. it gets a gray foggish tint.
[19:40] <Andy2> Is that a known bug?
[19:41] <Aiena> I had made a script a while bakc and it worked perfectly. ffmpeg seems to encode teh video filne but then I get "[mov,mp4,m4a,3gp,3g2,mj2 @ 0x7f28c80008c0] moov atom not found0/0
[19:41] <Aiena> file.mp4: Invalid data found when processing input" on playback.
[19:41] <Aiena> oops sorry that became multiline.
[19:42] <Aiena> Any ideas as to why. I'll pastebin the contents of my script here one moment.
[19:44] <Aiena> https://privatepaste.com/73746286c1 is the script
[19:50] <Aiena> OK found the issue
[19:50] <Aiena> it was badly exiting ffmpeg.
[19:50] <Aiena> script is ok
[19:51] <llogan> why do you upscale?
[19:51] <Aiena> becauswe someone s<id that videos are downscaled on playback in  RGB mode so quality is better
[19:51] <Aiena> said
[19:52] <llogan> i don't understand
[19:53] <Aiena> Yeah that someone gave some explanation which I can no longer remember.
[19:53] <Aiena> I wrote that long back and it always worked it still does only never pressed q to properly terminate ffmpeg
[19:54] <Aiena> hadn't used that script for a while.
[19:55] <Aiena> it was long back when I made that.
[19:56] <llogan> you can use mkv instead of mp4
[20:02] <Mavrik> *grumbl* why does x264 segfault while using crop filter
[20:03] <llogan> Mavrik: is it reproduceable?
[20:04] <Mavrik> yeah, but it's from API use, so I have to rule out my messup first
[20:04] <llogan> The acronym "API" redenders me useless
[20:05] <zerwas> Is there a way to compress a video file and passing the desired file size in the command with ffmpeg?
[20:05] <Mavrik> zerwas, sadly no
[20:05] <Mavrik> you'll have to use ffprobe to get length and then calculate bitrate
[20:06] <zerwas> Mavrik: that's what I suspected. Thank you very much
[20:06] <llogan> zerwas: example https://trac.ffmpeg.org/wiki/Encode/H.264#twopass (although the math might be wrong in the example)
[20:07] <llogan> also https://trac.ffmpeg.org/wiki/FFprobeTips
[20:08] <zerwas> llogan: Thank you! I will try to figure it out
[20:09] <blight_> is there a way for interleaved audio/video packets to know how much the audio stream will be delayed to the video stream? what i mean is i have a file here where first come 25 video frames and then comes the audio for the first frame and i wonder how much delay to expect or how to handle this properly
[20:10] <blight_> i could either cache the video packets before decoding them to get to the audio packets or i could open the file twice, once for reading the video and once for audio i guess... anybody has another idea?
[20:11] <Mavrik> why not just get a large enough buffer?
[20:11] <Mavrik> there's an upper limit on how much video and audio can differ defined by the standard (to conform to devices with limited memory)
[20:12] <Mavrik> depending on format of course.
[20:12] <blight_> atm i just buffer decoded frames in the gpu memory and with full hd or even higher res it can use quite some memory
[20:12] <blight_> and i let the user set the maximum amount of gpu memory to use
[20:13] <Mavrik> well you buffer incoming packets before decoding of course.
[20:13] <Mavrik> why would you do that in gpu memory?
[20:13] <blight_> i dont do that atm ;)
[20:13] <Mavrik> they're not THAT large
[20:13] <blight_> yeah
[20:13] <blight_> but i wonder if there is any reason not to open the file twice because i thought about playing a file backwards which i think would be easier when you can treat audio and video as independent streams
[20:14] <blight_> otherwise keeping the audio correct when seeking backwards through the video while playing backwards is hard i think
[20:15] <Mavrik> well if you actually have a file, no reason :)
[20:15] <Mavrik> so yeah, open it twice if you have an actual seekable file :)
[20:15] <blight_> yeah i always have files
[20:16] <blight_> i think i might try that in the future and in the meantime cache the packets before decoding until i find an audio packet or something
[20:16] <blight_> should be easier to implement
[20:18] <Baked_Cake> its hard to beleive that aac_he can cut a stereo track down to 64K and sound perfectly fine
[20:18] <blight_> 64Kwhat? K bit/second?
[20:18] <Baked_Cake> ya
[20:18] <blight_> i dont like lossy codecs
[20:19] <blight_> i mean lossy audio codecs
[20:19] <Baked_Cake> i guess if i had $1000 speakers id probly be saying teh same thing
[20:19] <Hello71> if only you don't want to think you wasted any
[20:20] <Mavrik> Baked_Cake, yeah, I was rather impressed by HE-AAC quality at low bitrates as wel
[20:20] <Mavrik> very nice for mobile use
[20:20] <blight_> i am not sure if it's imagination or real but when you listen carefully and long enough on a mid-class hifi setup i believe you can hear "artifacts" of say 128kbit mp3
[20:20] <blight_> but even if say 320kbit mp3 sounds totally find i prefer using flac to store stuff because space is so cheap nowadays
[20:20] <Baked_Cake> id agree with u on mp3
[20:21] <Camusensei> hello guys, Having a bit of trouble usinf -vf subtitles here... I'm doing "ffmpeg -i video.mp4 -c:a copy -vf subtitles=subs.srt out.mp4" and the subs are not on the video... what am I doing wrong please? :)
[20:21] <Mavrik> blight_, yep, but sometimes you just don't have 320k of sustained bandwidth :)
[20:21] <blight_> vbr is the worst btw :p
[20:21] <Baked_Cake> i usual encode my mp3 at vbr seetting at about 224K
[20:21] <Mavrik> hmm, we just use AAC now for most use
[20:22] <blight_> probably depends on the exact vbr setting but when it goes down to 96 and you hear the background noise coming and going its no fun ;)
[20:22] <Baked_Cake> id use aac for my music but my car stereo isnt that fancy
[20:22] <Mavrik> blight_, also depends on encoder :)
[20:22] <Baked_Cake> i only use aac for tv shows if i have a flac source
[20:22] <Camusensei> by the way, the errors I have while encoding are
[20:22] <blight_> Mavrik: yeah, another issue with lossy codecs ;)
[20:22] <Mavrik> Baked_Cake, yeah, we do broadcast video, for music I just use whatever Google Play Music does
[20:22] <blight_> Baked_Cake: rockbox?
[20:22] <Camusensei> [Parsed_subtitles_0 @ 0000000005e30ea0] Neither PlayResX nor PlayResY defined. Assuming 384x288         and                [Parsed_subtitles_0 @ 0000000005e30ea0] No usable fontconfig configuration file found, using fallback.
[20:23] <Camusensei> (note: this is a 1080p video)
[20:23] <llogan> Camusensei: windows?
[20:23] <Camusensei> llogan: yes
[20:23] <llogan> some other users had a similar issue i think. i don't know how any of that was resolved since i'm using linux.
[20:24] <Camusensei> llogan: ok I'll do that on linux and come back to you. thank you for the help
[20:24] <llogan> (they were using builds from Zeranoe
[20:24] <Baked_Cake> blight_: i use foobar on my pc, whats rockbox?
[20:25] <blight_> mp3 player firmware
[20:25] <blight_> i thought maybe for your car stereo :p to play aac
[20:25] <Baked_Cake> interesting
[20:26] <Baked_Cake> the idea of being able to flash my stereo's firmware never crossed my mind
[20:26] <Baked_Cake> its probly too old tho
[20:28] <blight_> dunno i think rockbox exists since like 2000
[20:28] <blight_> worth a look ;)
[20:28] <llogan> Camusensei: possibly related thread but i didn't read it much: http://ffmpeg.zeranoe.com/forum/viewtopic.php?f=10&t=318
[20:29] <Baked_Cake> yea from what im reading i dont even have to flash the stereo
[20:29] <Baked_Cake> it looks like this is mostly for mp3 players tho
[20:30] <Baked_Cake> ipods and such
[20:30] <blight_> yeah
[20:43] <Camusensei> llogan: I'm looking into it while it compiles ffmpeg ^^
[20:51] <Baked_Cake> o it looks like i can buy a source input adapter for my ipod that the steroe normally uses for the external disc changer
[20:52] <Baked_Cake> doesnt work if i just plug it into the usb input
[20:52] <Baked_Cake> if that works then my ipod courd do all the decoding
[21:00] <Baked_Cake> my ipod only hold 4gb tho, i should consider getting a new mp3 player
[21:06] <Baked_Cake> damn id have to iether get an ipod or a jvc player
[21:06] <Baked_Cake> what a scam
[21:24] <Aiena> llogan now I remember why UI upscaled so the video res is always a power of 2. ffmpeg used to complain otherwise
[21:49] <JayBlanc> Hi. Quick question, does ffmpeg return an error value if it encounters any kind of decoding/encoding error, even if it was able to continue the encode?
[21:51] <JayBlanc> The potential error values returned by ffmpeg don't appear to be documented, and this makes it hard to use in a script. Does anyone know when/what error codes are returned and what they indicate?
[21:55] <Aiena> hmm I am not sure. But normally when ffmpeg returns and error it highlights the output in red and gtells you waht the problem was.
[21:55] <Aiena> *returns an error
[21:56] <JEEB> JayBlanc, there's a setting to make ffmpeg fail at the first "error" it comes up with, but otherwise I would guess the result is zero unless decoding failed completely in some way at some point
[21:56] <JEEB> although reading ffmpeg.c most probably is the best way to learn :P
[21:56] <JayBlanc> Ah. It... appears to be set by a global variable.
[21:56] <JayBlanc> That's... Not great.
[21:56] <JEEB> "global variable"? AFAIK it was a setting like any other
[21:57] <JayBlanc> I just read the source code.
[21:57] <JEEB> I remember helping a person with this but I've since forgotten the setting :P
[21:58] <JEEB> and yes, if global var in C context then sure, not that I remember the details
[21:58] <JayBlanc> There's a global variable defined called "main_return_code" that gets used as the return value.
[21:58] <JEEB> but it sounds like it's the simplest way
[21:58] <JayBlanc> That's a bad idea.
[21:58] <JEEB> I don't disagree
[21:58] <orib> -xerror
[21:58] <JEEB> I've not looked at ffmpeg.c too much, thankfully
[21:58] <JEEB> mostly lavf and lavc internals
[21:59] <orib> ffmpeg -help full is your friend - it documents "-xerror error" as "exit on error"
[21:59] <orib> I guess it does what you want, though I'm not sure
[21:59] <JayBlanc> It means that some decoders will, and from what I've been seeing with trying to write a wraper script, are setting the main return value to some positive value if they encounter any error even if it's recoverable.
[22:00] <JayBlanc> Orib: Thank you, but that is not related to the question I asked.
[22:00] <JEEB> the decoders themselves are not setting any "global return value" as far as I know
[22:00] <JEEB> they just return a value as far as I know
[22:00] <orib> JayBlanc: Then can you rephrase the question? I don't understand it. Unless you're referring to the library rather than the executable?
[22:01] <JayBlanc> orib: Do you understand what I mean by 'return value' of a unix program?
[22:01] <JEEB> but yes, an error can be "recoverable" or "unrecoverable"
[22:01] <JEEB> and if you tell ffmpeg you want to exit on error it will do so with everything'ish
[22:01] <JEEB> there are some other modes too, though, IIRC
[22:01] <orib> Yes. When ffmpeg succeeds it returns 0 like any good old program
[22:01] <JayBlanc> Ah. Yes, you'd think that.
[22:01] <orib> and -xerror will supposedly make it exit with an error if it sees one
[22:01] <JayBlanc> It does not.
[22:02] <orib> oh ....
[22:02] <orib> It'
[22:02] <JayBlanc> See above: re Global variable that anything can set to make the return value be positive.
[22:02] <JayBlanc> And apparently some things are.
[22:02] <orib> It's been a while since I've used the cli...
[22:02] <orib> Silly me, then.
[22:03] <JayBlanc> Which makes ffmpeg kind of hard to use in wraper scripts when you want to use it to repair corruptable input streams.
[22:03] <JayBlanc> Because some decoders apparently set that global return value for main if they *ever* see any kind of error.
[22:03] <JEEB> they don't specifically set that
[22:03] <JEEB> something else uses it when calling the decode function
[22:04] <JEEB> and then the actual decode function just returns an error value depending on circuimstances
[22:04] <Camusensei> llogan: Nope, doing the same on linux :/
[22:04] <JayBlanc> Well, something seems trigger happy on setting a positive value on that global variable.
[22:05] <orib> JayBlanc: consider grepping the error stream
[22:05] <JayBlanc> How would that help?
[22:05] <orib> Ignore the return code, see if you had an error somewhere along the way
[22:05] <JEEB> JayBlanc, basically whatever is calling av_read_frame and avcodec_decode_video2 is then setting the return code :P
[22:05] <JEEB> the libraries don't touch any "global return code"
[22:05] <JEEB> :P
[22:06] <JEEB> aanyways, you will probably want to look into the code path of the demuxer or decoder you're planning on using for further information
[22:06] <JEEB> it could hint at the option you are interested in
[22:07] <JayBlanc> Something is causing ffmpeg.c's main to return a positive value, and ffmpeg.c returns that global variable as it's return value.
[22:07] <orib> JayBlanc: by "error stream" I meant the stderr (/dev/fd/2, /dev/stderr, 2>, whatever you call it)
[22:07] <JEEB> yes, and I'm saying it's not the decoders or demuxers since they don't touch that
[22:07] <JEEB> it's whatever is calling those :P
[22:07] <JEEB> (and encoders and muxers)
[22:07] <JEEB> also, #ffmpeg-devel could possibly be more interesting when discussing such specifics
[22:08] <JayBlanc> orib: Is there a unique common indicator I can search for in the error stream that indicates an unrecoverable error?
[22:08] <JEEB> esp. since no-one of the more knowledge'able side seems to be active right now here
[22:08] <JEEB> (I can't exactly know all of the parts of the code base)
[22:08] <JayBlanc> Well, return values are really supposed to be documented on the user side of things.
[22:09] <JEEB> if they are the error values as-is they're documented in libavutil or so I would guess
[22:09] <orib> JayBlanc: Not that I know of - I would guess the answer is "no" if you need to support all codecs, but "probably yes" if you only need to support one or two, though you'll have to look for them in the source code.
[22:09] <JEEB> but yes, documentation could always be better
[22:10] <JEEB> I'm just saying that you will most probably poke someone more knowledge'able of the command line specifics on -devel, since it's mostly end users here who really have no idea how ffmpeg works internally :P
[22:10] <llogan> Camusensei: and please provide subtitles file
[22:10] <JEEB> I know how lavc/lavf work to a point
[22:10] <JEEB> but I haven't worked on the cli side of things
[22:10] <JEEB> :P
[22:10] <Camusensei> llogan: doing that
[22:11] <JayBlanc> At the moment the short term solution seems to be to make another wrapper script that just serves to spoof a return value.
[22:14] <JEEB> if I were less tired I'd actually look into it again as I did for this one other guy, but if you have no interest I'm glad I didn't :P
[22:14] <JayBlanc> And since I can't actually be sure that I'd know what to look for in fatal error messages, there doesn't seem to be any point to grepping the stderr output.
[22:14] <JEEB> also seriously, poke -devel since as I said there seem to be no developers active here :P
[22:15] <JEEB> if you are actually goddamn interested in how shit works
[22:15] <Camusensei> llogan: http://sprunge.us/MReH
[22:15] <Camusensei> subs uploading...
[22:20] <Camusensei> llogan: http://c.thican.net/subs.srt subs here
[22:33] <llogan> Camusensei: i'm confused. subtitles filter requires libass, but I don't see --enable-libass in your configure. i expect to see "No such filter: 'subtitles'" in your console output
[22:34] <Camusensei> llogan: yeah, I don't know why it's not in there. but is was compiled with it, I'm sure of that. http://prntscr.com/4gf1of
[22:37] <llogan> Camusensei: your srt works for me. i don't know what is going on with your build. I'll blame gentoo.
[22:38] <Camusensei> llogan: well, zeranoe on windows does the same, so it's maybe not gentoo's fault
[22:38] <llogan> that might be a separate issue
[22:38] <llogan> but i don't know.
[22:38] <Camusensei> llogan: do you have --enable-libass displayed when you run your ffmpeg?
[22:38] <llogan> yes
[22:38] <llogan> but you can get away with using ass filter instead
[22:39] <Camusensei> llogan: I'll dig around that then.
[22:39] <llogan> ffmpeg -i input.srt output.ass; ffmpeg ... -vf ass=output.ass
[22:39] <Camusensei> ah, --enable-libass is displayed in my zeranoe version
[22:39] <llogan> wait...i might be wrong.
[22:40] <llogan> ass just doesn't require lavc and lavf (I was thinking of libass instead)
[22:41] <llogan> and, obviously i suppose, ass does need libass as shown in configure.
[23:21] <benlieb> is there any way with ffplay to get an idea of "where" you are in the video while playing?
[23:25] <llogan> benlieb: the last line in the console output shows the time elapsed in seconds
[23:25] <benlieb> llogan: yes but that gives me no indication of where I am compared total time of the video
[23:26] <benlieb> i've read the docs and I can't find anything.
[23:26] <benlieb> seems like a huge oversight
[23:26] <llogan> you could use mplayer, vlc, mpv, etc
[23:27] <llogan> or a fancy filtergraph
[23:27] <llogan> and drawtext
[23:27] <benlieb> llogan: I have other players, but I'm looking to take "thumbnails" of videos and I need accuracy to the ss.xxx
[23:28] <benlieb> so I need the decimal time, which no other player gives me that I know of
[23:28] <benlieb> just a bit like flying blind
[23:29] <benlieb> it would have been trivial to include the percentage in text on that last line
[23:30] <benlieb> there may be some combination of options that would get this...
[23:31] <llogan> ffplay -vf "drawtext=fontfile=VeraMono.ttf: timecode='00\:00\:00\:00': r=24000/1001: x=(w-tw)/2: y=h-(2*lh): fontcolor=white: box=1: boxcolor=0x00000000 at 1" input
[23:32] <llogan> possibly not accurate
[23:32] <benlieb> yeah, doesn't work but i get the idea
[23:46] <ubitux> huh?
[23:46] <ubitux> don't use timecode
[23:46] <ubitux> it's a smpte thing
[23:46] <ubitux> use text=%{pts\:hms} instead
[23:47] <ubitux> and you don't need the rate option then
[23:47] <ubitux> llogan, benlieb ^
[23:48] <ubitux> fontfile is also useless if you have a fontconfig
[23:50] <benlieb> can't get that command to work. think it has to do with font paths
[23:51] <ubitux> --enable-fontconfig allows you not to specify it
[23:52] <ubitux> then just -vf "drawtext=text='TS\:%{pts\:hms} N\:%{n}':boxcolor=white:box=1"
[23:52] <ubitux> should do what you want
[23:52] <ubitux> anyway gtg, gl
[00:00] --- Tue Aug 26 2014


More information about the Ffmpeg-devel-irc mailing list