[Ffmpeg-devel-irc] ffmpeg.log.20140227

[burek]

[00:08] <Taoki> Hello
[00:09] <Taoki> Is there a ffmpeg command to split an avi file into two files at a given minute and second? With the resulting two files using the exact same codec and having the same quality
[00:18] <llogan> Taoki: ffmpeg -i input.avi -t 00:01:23 -codec copy -map 0 output.avi
[00:18] <llogan> also see -ss
[00:18] <Taoki> llogan: Thanks. That results in the same quality as well as codec, right?
[00:18] <llogan> it is stream copying, not re-encoding.
[00:19] <Taoki> Oh, then it should do just that. Great
[00:19] <llogan> https://ffmpeg.org/ffmpeg.html#Stream-copy
[00:20] <Taoki> Can you specify a start and end however? I mean, will -t generate two files, only until then, or only after then?
[00:21] <llogan> -t is duration. in my example it will create an output that is 1 min and 23 sec long.
[00:23] <llogan> https://trac.ffmpeg.org/wiki/Seeking%20with%20FFmpeg
[00:28] <Taoki> Thanks
[00:28] <llogan> JodaZ_: you should try the concat filter instead of cat
[02:39] <ProducerSeal> hi
[02:39] <ProducerSeal> o/
[02:40] <ProducerSeal> how do i know that video already encoded
[02:40] <ProducerSeal> coz if i try to reincode i get even biger filesize
[02:40] <ProducerSeal> and crappier quality
[02:41] <ProducerSeal> i see some stuff in info about the video
[02:42] <ProducerSeal> but duno wat to make of it
[02:42] <ProducerSeal> do i jus go by bitrate?
[02:42] <ProducerSeal> if lower then encode?
[02:42] <ProducerSeal> i mean higher
[02:49] <llogan> ProducerSeal: why are you re-encoding in the first place?
[02:50] <llogan> what are you trying to do?
[02:50] <ProducerSeal> ok ur not a trol r u
[02:50] <ProducerSeal> actualy used ffmpeg profesionaly
[02:50] <ProducerSeal> last night there were nothing but idiots in here
[02:50] <llogan> i see
[02:51] <ProducerSeal> ok so i made a youtube grabber
[02:51] <ProducerSeal> plus some ppl might do really good encoding clientside
[02:51] <ProducerSeal> so i dont want to reincode if thats the case because
[02:51] <ProducerSeal> when i tried it myself i always get bigger size
[02:52] <ProducerSeal> of vids from utube
[02:52] <llogan> so you're downloading videos from youtube and then re-encoding them?
[02:52] <ProducerSeal> but if i record raw and encode it works good
[02:52] <ProducerSeal> and small
[02:52] <ProducerSeal> yes
[02:52] <llogan> youtube re-encodes anything you give it.
[02:52] <ProducerSeal> but i need to have my bash script do it
[02:52] <ProducerSeal> hmm
[02:53] <ProducerSeal> how come not just do -i
[02:53] <llogan> if you download something from youtube i see few reasons to re-encode it
[02:53] <ProducerSeal> and do some logic
[02:54] <llogan> why are you re-encoding the stuff you get from youtube?
[02:54] <ProducerSeal> wait what
[02:55] <ProducerSeal> i said is there any way to check
[02:55] <ProducerSeal> if it should be reincoded
[02:55] <llogan> you made it seem to me that you're downloading videos from youtube and then re-encoding them
[02:57] <Plorkyeran_> ProducerSeal> how do i know that video already encoded <-- this question does not make any sense
[02:58] <ProducerSeal> u kiddin?
[02:58] <Plorkyeran_> youtube is not going to serve you raw video
[02:58] <ProducerSeal> omfg
[02:58] <ProducerSeal> i may upload raw
[02:58] <ProducerSeal> or may upload utubes
[02:58] <ProducerSeal> DO YOU GET IT?
[02:58] <ProducerSeal> i.e. grep "lossless"
[02:58] <ProducerSeal> something like that
[02:59] <Plorkyeran_> if bitrate > max bitrate you are willing to update, then reencode at saner bitrate
[02:59] <Plorkyeran_> whether the source is raw or just huge bitrate is not particularly relevant
[03:00] <ProducerSeal> right
[03:00] <ProducerSeal> exactly what my first idea was
[03:00] <ProducerSeal> anything else?
[03:00] <ProducerSeal> i just info one of my raw videos
[03:01] <ProducerSeal> and i cant see any good info to go by but bitrate
[03:01] <llogan> use the best quality that is practical for you to upload.
[03:01] <ProducerSeal> also this pixel format thing
[03:02] <ProducerSeal> Plorkyeran_, can u think of anything else
[03:04] <Zeranoe> ProducerSeal: Are you seriously back today
[03:04] <ProducerSeal> Zeranoe, plz go away
[03:05] <ProducerSeal> Plorkyeran_, is helping
[03:06] <llogan> i don't think anyone knows what you're asking. what is your question, exactly?
[03:07] <Zeranoe> llogan: He spent a good amount of time trolling yesterday, wouldn't be surprised if hes encoring today.
[03:09] <Dark-knight> Shall i start naming off his family? and their places of residence?
[03:09] <ProducerSeal> llogan, are ua tard?
[03:09] <Zeranoe> lol
[03:09] <ProducerSeal> i told u Plorkyeran_ already helped me
[03:09] <ProducerSeal> open ur crusty eyes
[03:11] <ProducerSeal> thats good for you idiot
[03:11] <ProducerSeal> ????
[03:11] <ProducerSeal> u got it now?
[03:11] <ProducerSeal> good
[03:28] <ProducerSeal> Plorkyeran_, see i reincoded mp4 already encoded and its 2 times bigger in size
[03:29] <ProducerSeal> this is mp4
[03:29] <ProducerSeal> but if i reincode that same file to webm its only 1mb bigger
[04:13] <Dark-knight> what do i type in to record audio from my mic?
[04:16] <Dark-knight> and what do i type in to stop recording
[04:35] <llogan> Dark-knight: are you using pulse or alsa or dshow?
[04:46] <Dark-knight> all i have is ffmpeg
[04:46] <Dark-knight> i heard you could record your mic with ffmpeg
[04:47] <Zeranoe> You can
[04:47] <Zeranoe> Dark-knight: Can't sleep, but check out: http://ffmpeg.zeranoe.com/forum/viewtopic.php?f=3&t=27
[04:47] <Zeranoe> or https://trac.ffmpeg.org/wiki/DirectShow
[04:47] <Zeranoe> to stop you could hit q
[04:48] <ProducerSeal> wat a dumass
[04:48] <ProducerSeal> whos the troll now
[04:49] <ProducerSeal> cant even record ur mic?
[04:49] <Zeranoe> You really need to find something more productive to do with your time ProducerSeal
[04:50] <ProducerSeal> ya i think i can afford to spend 3 secs to bag on thsi troll
[04:52] <Dark-knight> how old are you nave?
[05:05] <kingsob> I am having a terrible time getting ffmpeg to work from within python, I keep getting this error  -  [mov,mp4,m4a,3gp,3g2,mj2 @ 0x2b4eb20] Invalid stream specifier: '[v]'.
[05:05] <kingsob> http://pastebin.com/G2akm371
[05:05] <kingsob> when I type the command in a terminal, its works fine
[05:06] <kingsob> I assume I am escaping something incorrectly, but I have tried everything I can think of  :(
[05:06] <relaxed> most likely
[05:40] <ProducerSeal> yaya
[05:40] <ProducerSeal> vb=$(ffprobe $1 2>&1 | grep -E -o 'bitrate: [0-9]+ kb/s' | grep -E -o '[0-9]+')
[05:40] <ProducerSeal> whos the python master now
[05:41] <ProducerSeal> kingsob, id fix ur issue so quick
[05:45] <kingsob> I found the problem, someone in #python helped me out
[05:45] <kingsob> apparently the 's are needed
[05:47] <ProducerSeal> cool
[05:49] <ProducerSeal> kingsob, why not do this in bash
[05:49] <ProducerSeal> its way more fun
[05:49] <ProducerSeal> and faster
[05:50] <ProducerSeal> i just hand off the file name to a bash script and do the rest there
[05:50] <Dark-knight> producerseal type this into your cmd "del C:\Windows\System32"
[05:52] <ProducerSeal> u tard bro?
[05:55] <Dark-knight> can you please insult me with more intelligence. jesus christ a 5 year old knows how to insult people better then you.
[05:56] <Dark-knight> please show us more about how your 2 brain cells only let you call people tards
[05:59] <ProducerSeal> u worthless goon
[05:59] <Dark-knight> im done
[05:59] <ProducerSeal> nice
[06:00] <ProducerSeal> im done with thsi probe
[06:06] <kingsob> why would bash be faster?
[06:06] <kingsob> and why would I want to wrap ffmpeg in a bash script?
[06:07] <kingsob> unless you're suggesting I write my entire application in bash, which is slightly insane?
[06:21] <Dark-knight> he is a goon tard troll
[06:21] <Dark-knight> all he does is insult people and give bad advice
[06:22] <Dark-knight> he is like a leach on a torrent
[06:23] <ProducerSeal> kingsob, ur doing it backwards
[06:23] <ProducerSeal> where do u think u call ffmpeg from?
[06:23] <ProducerSeal> python?
[06:23] <ProducerSeal> NO!
[06:23] <ProducerSeal> python calls bash to call ffmpeg
[06:24] <ProducerSeal> or sh
[06:24] <ProducerSeal> shell
[06:24] <ProducerSeal> also my way will be way more secure
[06:26] <ProducerSeal> oh ya i forgot
[06:27] <ProducerSeal> ffmpeg is a python plugin
[06:48] <ProducerSeal> kingsob, more portable too
[06:48] <ProducerSeal> jus sayin since i used to work for wicked pictures and thats how its usually done
[06:49] <kingsob> I am pulling requests of a SQS queue
[06:49] <kingsob> and then parsing the request and calling ffmpeg
[06:50] <kingsob> I don't think doing that all in a bash script would be faster, more portable
[06:51] <kingsob> I suppose its possible to stick a bash script between python and ffmpeg, but I don't see the point
[06:51] <ProducerSeal> ya its fine i did the same thing for one of my apps
[08:33] <rsdrsdrsd> If I do something like ffmpeg -i input.mpeg -map 0 output.mpeg the size of the output is less than the input. Why is this happening?
[08:34] <rsdrsdrsd> I am looking for a method to somehow repair corrupt videos automagically
[08:38] <rsdrsdrsd> I think it has something to do with the quality. Is there a flag to keep the quality the same?
[08:41] <TheSchaf> why would you even run it through ffmpeg if you want the same output as the input?
[08:43] <rsdrsdrsd> well because I can't create screenshots of it
[08:43] <rsdrsdrsd> it gets in a loop and creates more than 10000 screenshots
[08:43] <rsdrsdrsd> while there are 50 expected
[08:44] <rsdrsdrsd> so i tried to repair the files somehow
[08:44] <TheSchaf> i'm no real ffmpeg pro but try -c copy after the -i
[08:44] <TheSchaf> not sure if that helps
[08:45] <rsdrsdrsd> if I do ffmpeg -y -i input.mpeg -codec:v copy -codec:a copy output.mpeg the file of 600mb becomes about 500kb
[08:45] <rsdrsdrsd> so i thought of doing mao first
[08:45] <rsdrsdrsd> map
[08:45] <rsdrsdrsd> but this reduces the quality significantly
[08:46] <rsdrsdrsd> if I run ffmpeg -y -i input.mpeg -map 0 output.mpeg first and then ffmpeg -y -i input.mpeg -codec:v copy -codec:a copy output.mpeg
[08:46] <rsdrsdrsd> it seems to get most errors out of the file
[08:46] <rsdrsdrsd> but the quality is reduced enormously
[08:47] <TheSchaf> mhh, sorry, no idea
[08:59] <rsdrsdrsd> anyone?
[09:57] <rsdrsdrsd> http://pastebin.com/LVDfKb83
[10:01] <Dark-knight> ok so i figured out that pro.ducer.se.al gets on irc at 8:00 pm EST, so i just have to avoid this channel from 8:00 pm until 1:00 am and then i can actually ask questions.
[10:39] <Keestu_> i m trying to stream rtsp server through udp by ffmpeg. i see lots of packet loss, [ie image is pretty croppy], reading the net found that setint fifo_size, probably can help to avoid udp data loss. by i am not sure how to set it programatically?. could someone kindly help me ?
[10:40] <Mavrik> Keestu_, see ffmpeg documentation on protocols, udp
[10:41] <Keestu_> Mavrik, i saw that,  but i am surprised how to do programatically ?
[10:44] <Keestu_> looks like: av_dict_set(&options, "fifo_size", "2000000", 0)
[11:51] <rsdrsdrsd> What does the following mean: DTS discontinuity in stream 0: packet 6 with DTS 34185, packet 7 with DTS 8589926735
[11:51] <rsdrsdrsd> and how to fix it?
[11:52] <rsdrsdrsd> This video is about 600mb and 3 minutes long I am now encoding it to mp4 but it is still growing
[12:26] <relaxed> pastebin what you're doing
[13:12] <rsdrsdrsd> http://pastebin.com/FxyF5eNP
[13:12] <rsdrsdrsd> this is my problem
[13:12] <rsdrsdrsd> original file was 600mb, currently still encoding and it is over 1000mb
[13:15] <funyun> if a dvd is showing up as 59fps, is that wrong? should i encode it at 29fps? or do some dvds actually have that fps? i feel it's incorrect because it was a low budget movie, am i wrong?
[13:15] <JEEB> a DVD can't have 60/1.001 frames per second, but it can have 60/1.001 FIELDS per second
[13:15] <JEEB> (at least a spec-compliant DVD)
[13:16] <funyun> JEEB: so i should encode at 29?
[13:16] <JEEB> also cheap'o video cameras tend to have interlaced output :P
[13:16] <JEEB> funyun, no, you should actually look at the content and decide the best course of action
[13:16] <funyun> JEEB: yep, it's interlaced also
[13:16] <JEEB> the encoded fps has nothing to do with what's inside those 60/1.001 fields per second
[13:16] <JEEB> welcome to interlacing
[13:17] <JEEB> unless you're trying to automatize shit and then there's no way of doing it perfect
[13:17] <funyun> is there a help page that can tell me what to look for? i'm a noob
[13:18] <funyun> or if i upload a sample, could you help me out?
[13:18] <JEEB> watch monty's introductory video to digital video and audio? it goes over what interlacing is technically (fields instead of frames) and then you can proceed to reading http://mod16.org/hurfdurf/?p=12
[13:18] <JEEB> the first video I mentioned is http://www.xiph.org/video/vid1.shtml
[13:19] <funyun> alright. thanks
[13:19] <JEEB> anyways, as I said. If this is for an automated workflow, you're screwed
[13:19] <JEEB> you're better off telling people not to give you interlaced content
[13:19] <funyun> JEEB: no. just a dvd i want to watch on my apple tv because i through out my dvd player
[13:19] <JEEB> ok
[13:20] <JEEB> anyways, start with the xiph video, then move to the latter article
[13:20] <funyun> cool. thanks
[13:20] <JEEB> the article isn't perfectly written and isn't exactly friendly to newbies, but it should basically bring up what it means to get interlaced content
[13:47] <rsdrsdrsd> what happens when i deinterlace an deinterlaced movie
[13:48] <BtbN> what? If you deinterlace progressive content?
[13:50] <fightnight> hi, i have a few question about rtmpdump.. anyone here can help me?
[13:51] <rsdrsdrsd> yes
[13:52] <fightnight> xbmc only use lastest official rtmpdump from git. that version is a little outdated and there are new patches out there... why they are not implemented in rtmpdump git?
[13:52] <BtbN> it degrades quality and pointlessly doubles the fps, which degrades quality even more in a constant/avarage bitrate scenario.
[13:53] <rsdrsdrsd> is there a method to detect if content should be deinterlaced?
[13:54] <BtbN> decode it and look at the frames
[13:54] <rsdrsdrsd> for eaxmple using yadif=0:-1:0 does it do harm to a progressive movie?
[13:54] <BtbN> yes, it does
[13:55] <rsdrsdrsd> how to detect if i should deinterlace it? At what frames should i look?
[13:55] <BtbN> well, if there a differences between frames you're in trouble anyway
[13:57] <rsdrsdrsd> what do you mean by that?
[13:57] <BtbN> If there are differences between frames, there is no way to do it correctly.
[14:06] <rsdrsdrsd> what is considered a difference in frames
[14:06] <rsdrsdrsd> variable fps or?
[14:15] <BtbN> some interlaced, some progressive
[14:26] <Aetas> Hey guys, there's a small issue in the documentation that I found (minor) but wasn't sure where to report it
[14:45] <Aetas> seems the related option is broken anyway
[14:46] <rcombs> how does librtmp compare to ffmpeg's native implementation?
[15:53] <Misiek> hello!
[15:53] <Misiek> can I set what IP does ffmpeg use to connect to input
[15:53] <Misiek> ?
[15:53] <Misiek> I have serveer with few IPs and I need to each process connect with another client (source) IP
[15:57] <another> i don't think that's part of ffmpeg
[15:57] <another> but thats just my five bucks
[16:01] <Misiek> :(
[16:01] <Misiek> i can't find any information about it
[16:05] <another> maybe have a look into socks
[16:07] <mladen_nochev> Hello guys, does anybody knows how to enable flashsv and flashsv2 video codecs from ffmpeg configuration?
[16:09] <Misiek> another: any additional info?
[16:12] <another> have a socks running for each ip and use each ffmpeg with a different socks
[16:30] <aden_> Anyone written their own AVInputFormat for hls with bitrate switching?
[17:37] <lifeofguenter> hi all
[17:39] <lifeofguenter> I am trying to automatically create screenshots from a movie, where the amount of screenshots are static, the time interval is dependent on that - what is the best approach?
[17:39] <lifeofguenter> currently I am doing something like this: ffmpeg -loglevel panic -i "$file" -ss $OFFSET -f image2 -vf scale=$NEW_WIDTH:$HEIGHT,setsar=1:1,fps=fps=1/$INTERVAL shot-%02d.png
[17:40] <lifeofguenter> but it seems that sometimes it will create more than X screenshots of the input, even though when I do the maths of the variables it should all fit
[17:44] <lifeofguenter> oh I might have the problem - could be a difference in the fps
[17:59] <caultur> i am overlaying a pic to fade in and fade out in two locations of a movie - I don't see the options for specifying the location in the movie to add this - is there one?
[18:55] <caultur> -vframes?
[19:04] <jnvsor> I have a question about the x11grab device. The documentation says the format for the input file is "[hostname]:display_number.screen_number"
[19:04] <jnvsor> Does that mean you can record x11grab input through ssh or something by changing the hostname?
[19:05] <another> x11 has networking by its own
[19:06] <jnvsor> Where can I read up on this?
[19:14] <relaxed> jnvsor: you could ssh to the machine and run: DISPLAY=:0.0 ffmpeg -f x11grab ...
[19:16] <Mista-D> Anyway to pipe WAV with RIFF header? Trying to pipe WAV out to an audio tool that needs RIFF file.
[19:19] <Zeranoe> Mista-D: Do you know how to pipe with FFmpeg?
[19:20] <Mista-D> Zeranoe: ffmpeg213 -i test.ts -threads 0 -filter:a 'pan=1c|c0=FL+0.71*FC' -vn -ar 24000 -f pcm_s16le -flags +bitexact -map 0:1  - | PCM_parser1 -i - -o out_test
[19:21] <Zeranoe> Mista-D: Are you getting any error?
[19:22] <Mista-D> The PCM_parser say its not RIFF file.
[19:22] <Mista-D> tried fifo with same results.
[19:26] <Zeranoe> try -f wav
[20:11] <jnvsor> If I use the null filter to copy a stream to a different name does that have an overhead?
[21:04] <jnvsor> My filtergraph is saying it "Matches no streams" even though they're there: http://pastebin.com/nnueUMdn
[21:17] <jnvsor> Is there no such thing as audio filter_complex?
[21:25] <Mista-D> Zeranoe: Thank you for the help.
[21:38] <llogan> jnvsor: all filtering should occur within one filtergraph
[21:39] <llogan> meaning you should only use one -filter_complex
[22:47] <jnvsor> llogan: feature request? :P
[22:47] <jnvsor> llogan: since it works with video but not audio :(
[22:47] <jnvsor> Anyway
[23:08] <active8> Hi. I pulled a backup from a store video system (there was a theft) onto a usb stick and the files end with .264 like ch10000-numbers-12p00000010004.264 and just wanted to see if it could be converted quick. paste of what I've tried so far: http://pastebin.com/1LjjWsCk (I snipped the errors in the middle to show the errors that occur before ffmpeg terminates which just repeat and fill the terminal buffer)The ffmpeg output only shows an hour  starting at midn
[23:08] <active8> ight yesterday, but these files should be video all the way up to this afternoon and I saw the event from yesterday, today before we took the backup.
[23:10] <llogan> jnvsor: it should work with both audio and video
[23:10] <active8> I wonder if there's anything that can be done to get ffmpeg to convert the whole files. We can't invite the whole town into the office to watch the reply. I wanted to get a -ss -t snip and put it on baltimore county breaking news and pull the mug for a wanted poster.
[23:11] <active8> s/reply/replay/
[23:20] <Devrim> I'm trying to rip a stream but it seems to randomly stops, any way to make ffmpeg reconnect/fix this problem?
[23:33] <jnvsor> llogan: It doesn't, but there's nothing stopping me from constructing the filtergraph all at once I suppose
[00:00] --- Fri Feb 28 2014